Understanding Intercooling, Charge-Coolers, Heat Exchangers and Circulation Pumps
 

I’ve been thinking about intercoolers, done some research on specs, and noticed an interesting trend.

Heat exchangers in one form or another are a pretty mature technology, and car radiators, air-con condensers, oil coolers and heater rads have settled into an established pattern. Medium European cars tend to have horizontal-flow alloy rads with plastic end-tanks, and the cores measure around 600 x 400 x 26, 32 or 40mm thick, depending on engine size (ie: around 6 to 9 litres volume). Heater rads are about ¼ the size of the engine rads, and condensers are almost invariably 16mm thick these days.

However, intercoolers have evolved a lot over the last few years. Diesel turbos with air-air intercoolers are very common, and there’s a lot of information available from on-line sellers of after-market intercoolers. There’s not so much information for current models, or large engines, or charge coolers, but I was still able to do a useful survey of typical intercooler specifications. There’s a clear trend for IC’s to get bigger. They started off quite small, oil cooler size, but have now reached much the same size as the engine radiator.

Modern turbo-charged engines will typically have a 30mm thick intercooler and 30mm radiator, with a 16mm condenser sandwiched in between, and they will all have much the same width and height. Here’s a summary of some common stock coolers, with some truck & tuning units thrown for comparison. The core dimensions are in mm and the volume in cc, which I think gives a reasonable measure for comparison:

CXRacing HE 002: 24x8x2.5” = 610x203x64 = 7,863
Frozen Boost HE 101: 24x7x3.5” = 610x178x89 = 9,663
ZZ Performance Stealth HE: 660x349x44 = 10,215
THS Up-rated VAG 2.0 TFSI IC: 610x415x57 = 14,430
Silicone Intakes IC Type 2: 610x300x101 = 18,469

1991 VW Golf 1.9 TDi: 175x145x100 = 2,538
1995 VW Sharan 1.9TDi: 500x70x88 = 3,080
1998 Mercedes S320CDI: 530x115x65 = 3,962
2001 BMW E46 318d: 540x130x50 = 3,510
2002 Mercedes S600 TT: 580x265x21 = 3,228
2004 Mercedes S65 TT: 580x465x21 = 5,700 (estimate)
2004 Renault Laguna DCI: 530x390x32 = 6,614
2004 Toyota Avensis 2.0D: 642x418x26 = 6,977
2005 Nissan Pathfinder 2.5 TDI: 445x188x65 = 5,438
2005 Seat Alhambra 2.0 TDI: 633x412x25 = 6,520
2005 Opel/Saab All 2.0 turbo: 647x400x29 = 7,505
2008 VW/Seat All 2.0TDI: 617x404x32 = 8,687
2001 All Volvo Turbos: 688x421x30 = 8,689
1993 Volvo FH D12 Truck 340hp: 885x752x63 = 41,928

Granted, a water cooler is more efficient than an air cooler, as the water tubes can be much thinner than air tubes. I think there are two benefits – more cooling fin area for a given volume, plus more ambient air flow due to the reduced obstruction. Therefore I guess that a charge cooler HE needs to be about half the size of an air-air cooler for the same efficiency. That sounds about right, as the charge cooler and the heat exchanger are effectively “two halves” of the intercooler – the water circuit just providing the thermal interface between the two.

IIRC, the V12TT charge coolers have a combined volume of around 500 cu-in, or about 8 litres, which I think is quite reasonable. Having looked at a lot of after-market charge cooler kits (CXRacing etc) it seems that the charge cooler and HE volumes should ideally be similar. That seems to be consistent with the way air-air coolers are designed, with about the same volume dedicated to the internal and external cooling volumes.

Another very rough rule of thumb seems to be that the CC and HE volumes should each equal the engine capacity itself – and double it for high-boost engines (double-digit PSI – that sort of level). That’s not a rule that’s written down anywhere, it’s just my observation. I think its fair, given that the more recent examples above have typically four times the engine capacity – equivalent to twice the capacity for water cooling.

Whichever way you look at it, all this illustrates how relatively small the S600/CL600/SL600 heat exchanger is. It’s only 3.2 litres, and really ought to be twice that – similar to the charge coolers (and most cars engine radiators). And there’s an argument for saying it should be four times greater. The dimensions above came from radiator distributor web-sites, and they allow easy comparison of radiator and intercooler specs. For modern cars and trucks (which have used intercooled turbos for a long time, and are well-developed) it’s now normal for the intercooler to be the same size as the radiator – and even slightly bigger in some cases. The S600 radiator is 641x469x40mm or 12 litres, which sounds like a good target. The bottom line seems to be – the bigger the better.

Nick

I'm pulling together various posts and charts on pumps and heat exchangers into one thread, and here are a few posts copied from elsewhere. Don't worry if you already saw this, I've got somenthing useful coming up....

I suspect the benefit of a high flow pump is the reduced delay to START cooling the charge coolers - when the pump is first switched on due to high IAT, that is. There's quite a lot of hot water sitting in that system - especially thinking about the pipes that run past the turbos - and it must take time to get that round a full circuit.

The argument that high flow rates result in higher charge temps doesn't work on many different levels. Its probably best consider a cooling circuit interms of how it carries heat energy around, rather than how it behaves from a temperature perspective.

Firstly, if the water doesn't stay in the HE long enough to cool down, then the converse must also apply - the water won't stay in the charge coolers long enough to heat up.

Although the water will pick up less heat in each pass through the IC when it flows fast, it will complete more circuits in a given time, so it will complete more passes.

So the question is whether the faster flow is more than enough to offset the shorter heat transfer time. Heat exhanger behaviour is complex, there's nothing simple or linear about how they work. Heat transfer is governed by many variables, but the simplest one is that heat flow is proportional to temperature difference. Therefore if the coolant is cold, the heat transfer willbe fast.

If the coolant circulates slowly, so that it has time to absorb a lot of heat, its temperature will tend to approach that of the air its cooling. As it does so, the temp differential falls, and rate of heat flow also falls. Therefore the highest heat transfer happens when the coolant is coldest.

That's another way of saying that heat transfer increases as flow rate increases, which is one of the other guiding principles of heat exhangers. That's not linear though, as heat transfer depends on other factors - like the surfaces of the tubes. Rough sufaces and high flow speeds will encourage turbulent flow, which is more effective at heat transfer. There's more convection transfer when there's turbulence, so you don't just rely on the conduction of the coolant.

Turbulent flow is difficult to predict, and depends on the fluids, the surfaces and the Reynold's number. Although I did study aerodynamics and fluid mechanics when I was at British Aerospace, I don't pretend to fully understand it all. However, if there's one thing that DOES encourage turbulence, its high fluid speed. Maybe the Bosch pump is fast enough to cause turbulence, but a faster pump MAY encourage it more, and it certainly won't make it worse.

Water has a high heat capacity, and it could be that the stock pump is fast enough to ensure that the coolant temperature rise through the IC's is low enough during each pass to ensure that the capacity of the coolant itself isn't the bottleneck in the overall cooling system. The limitation may well lie elsewhere, such as with the thermal resistance of the HE to ambient air, or with the IC to the intake air. In that case, the cooing fins may be held at the coolant temperature quite effectively, but there might simply be not enough of them to conduct the heat to what is a fairly insulating medium. But I'm starting to speculate there.

The end result is generally that higher flow rates increase heat transfer, and the END RESULT is that the medium being cooled will be cooler.

Of course it could be that the real benefit is that cooling simply STARTS faster for the first reason I gave. Given that this is effectively a thermostatically-controlled system that only runs on demand, that could be the overriding advantage.

Nick

This is quite an interesting pump - or rather an interesting pump controller.

http://www.daviescraig.com.au/Contro...0-details.aspx

When we talk about Bosch, Johnson and Meziere pumps, we talk in terms of swapping pumps out and using the same electrical interface.

I suspect that one of the biggest issues with the stock pump is the way its controlled - pump turns on at 47oC IAT; pump turns off at 35oC IAT. Maybe it works better than I give it credit for, but it does seem a rather crude algorithm.

Running at full speed all the time is probably too wasteful, but I think it needs better control, that allows the coolant to be kept cooler all the time. Some sort of variable speed control is probably what's needed, and that's what the Davies Craig controller appears to do. There's a variable duty cycle - 25% at low temperature, 50% at medium, and continuous at high temperature.

Since the controller is targetted at engine cooling systems, there isn't the flexibility to support low-temperature cooling systems like ours. The target temperatures can be set to 75, 80, 85, 90 or 95oC, which is too high for us. However, at the lowest setting, the pump would be run at least 10sec on/30sec off all the time, increasing to 10 on/10 off for any temp over 55oC, and continuous over 70oC.

That low duty cycle would probably be ideal to stop the coolant in the pipes heating up during cruising, when there's no boost. But I'm not sure its enough to keep the temp below 55oC when the engine IS on boost.

Perhaps there's a simple way to add a resistor in series with the thermistor, so the controller thinks the coolant temp is higher than it is.

And thinking out loud for a moment, I'm sure there must be a temp sensor somewhere that the stock system already uses. That should make the installation a bit easier. I guess its only an air sensor though.

I expect the Davies Craig controller is intended to be used with their own electric water pumps, but just to prove a theory, it would be interesting to try the controller with the stock Mercedes pump, and see what that does to IAT and subjective performance. I'd like to use one of their pumps as well, but I'd like to understand the individual contributions from the controller, pump, HE, etc.

People are always drawing comparisons with water circuits when they try to explain elementary electronics, so a backwards comparison might work as well. The hydraulic pump is like the dynamo or battery, the pipes are the wires, and the IC & HE are the resistors. The pump generates a certain pressure, and this is analogous to voltage.

The whole circuit has an overall resistance to flow, which restricts the current, so to speak. If there was just one IC, the resistance would be higher, as all the flow has to go through one resistor, rather than being split between two parallel units, and so on.

Unless we make changes to the IC or HE, that resistance is fixed, so we can only change the pump. The pump will only achieve its maximum voltage/pressure when it’s delivering no current. That's like putting a voltmeter across a battery - you get the max voltage, but there's no load. As soon as current starts to flow, the voltage drops, and when you take it to the extreme with nothing connected to the pump, you will achieve the maximum flow, but there will no pressure. Similarly, when you crank an engine, the battery voltage plummets. These extreme conditions are just like open circuit and short circuit on a battery - you either get maximum pressure or maximum current, but not both at the same time.

I believe pumps typically operate in a middling condition, where both the pressure and flow are somewhere near half the respective max values. So if a pump can achieve 6 psi open circuit and 10 gpm open circuit, I'd expect it to achieve something like 4 psi at 5 gpm in a typical installation. That's what all those pressure/flow charts are all about - they show how much pressure loss you get when coolant starts to flow - or alternatively how much the flow drops when you add resistance to the circuit – it’s the same thing. Pressure times flow is power, and pumps deliver most power to the fluid when they work in the middle region. If the flow or the pressure is close to the maximum value, the pump is working against too little or too much resistance respectively, and the pump won't be doing as much useful work as it could.

The faster you push the coolant round, the more pressure you need (but its not linear, it’s more like a square function). So, if we fit a bigger pump with a higher flow rate, that MAY increase the flow around the system, but ONLY if the pump can generate a higher pressure at that higher flow rate. What we're looking for is not just a pump with a high max pressure, or with more max flow, but one where the pressure/flow curve moves up and to the right, in the area where the system actually works.

Alternatively, we could fit a larger HE, hopefully one with more pipes, and hence less flow restriction. That in itself will increase the overall flow rate slightly, because the total resistance of the cooling system is reduced. The resistance will reduce, the flow rate will increase, and the pressure will reduce. The operating condition of the pump will find a new equilibrium point, and will move to the right slightly. That's a good thing, because we gain from the reduced thermal resistance of the larger HE, plus the higher flow rate round the cooling circuit.

So we do need to consider both pressure and flow rate, but they are related.

Many electric pumps are designed for engine cooling, so they often have 1.25" outlets. The Meziere and Davies Craig pumps have the option of various diameter outlets, including 19mm, which should be ideal for our cars.

I think there's no doubt the Davies Craig would be a good pump, but there's a few things that point me towards the Meziere WP136S or Johnson CM90.

Because this is a modification, I think its best to tread in someone else's steps - in other words, use a proven route to achieving what you want. Let other people find out the pitfalls for different potential solutions where you can. Meziere and Johnson aren't totally without their issues, but they're widely used and fairly well understood.

My impression is that the Davies Craig 80 and 115 pumps are intended as replacements for engine pumps. They have high-flow characteristics: shallow pressure/flow curves and large ports. I think the charge cooler circuit isn't like that - its higher resistance and lower flow. There are two, small heat exchangers, and the pipes are small. The EWP80/115 aren't auxilliary pumps, whereas Bosch, Meziere and Johnson are intended as aux pumps. They make other models for engine cooling.

Davies Craig do make a couple of specific auxilliary pumps, the EBP (EBP15) and EBP25 (which looks suspiciously familiar) but they doen't flow very much. However, they do have 3/4" connections

The EWP80 & 115 have large inlets and outlets, which are suited to engine cooling. All the auxilliary pumps have 3/4" connections, which makes life easier for us.

What's really interesting for me about Davies Craig is the pump controller, which gives a us a better option than the Mercedes ECU. I think that might be really worthwhile, and not necessarily with Davies Craig pumps. It could be considered as an independent option to the pump itself - something to add after upgrading the pump, or even to be used with the stock pump, as an effective alternative to changing the pump.

Davies Craig pumps seem to be well-distributed around the World, but I think I would go for a WP136S or CM90, there's no great problem with importing from the US. I've done that with lots of electronics and car parts with no issues.

I've been trying to put together a comparison of all the popular circulation pumps, but the manufacturers don't use the same data for comparisons. Sometimes we just get a open outlet flow figure, and sometimes its flow against a particualr head of pressure, whcih is more meaningful (but its never the same pressure). Some pumps are designed for engine cooling, which has low restriction. Others are designed for auxilliary circulation like charge cooling, which has less flow, smaller heat exchangers and long, small pipes. These pumps will have high pressure, low flow charcteristics, but its difficult to figure which these are.

To add to the confusion, there are lots of different units in use. Flow can be given in litres/min, litres/hour, gallons/min or gallons/hour (and that's before US vs imperial gallons even gets in the way). Similarly, pressure can be given in PSI, kPa, bar, etc. As well as providing incomplete information, some manufacturers even post the wrong info in their data sheets (one of them can't convert kPA to PSI correctly). So how anyone possibly know whether a Johnson is better than a Bosch when there are so many variables?

I think its important to understand how these pumps are being used - what part of their operating envelope is used in our charge cooler systems? That needs some understanding of the flow resistance of the system, which tells us how much presure is required to achieve a certain flow rate. Unfortunately its not linear (like electrical resistance). Its probably best represented as a second-order polynomial function, which is typical of fluid drag. Car aerodynamic drag increases with the square of the speed, and I think coolant circuits are similar. Davies Craig published a resistance curve for a typical 6-cylinder engine in some of the EWP115 literature, so I've assumed its representative, and modelled it as a simple square-law function with appropriate scaling. I called this Engine Cooling Resistance.

Charge cooling probably has similar characteristics, but with higher resistance, for the reasons given above. I read some very useful posts on MBW that said some typical pumps achieve around 5-6gpm installed flow, and this gives a big clue as to where the flow/pressure curve sits for charge cooler systems. I modelled another square-law curve that fitted a sensible characteristic, and called this Charge Cooling Resistance. Where the pump curve crosses the resistance curve, that's how much flow you'll get in a particular installation. Of course its only a guess, but its based on the best information I can find, and there's not much of it. If anyone can help fill in the missing pieces with actual installed flow or installed pressure, I'd be really grateful.

I've recorded all the information I can find on a spreadsheet, and plotted some charts, which I'll post tomorrow (takes lots of file conversions). I consolidated everything into kPa pressure vs litres/min flow (sorry Americans!), but I'll try to plot some dual scales. 1 bar = 14.7psi = 101kPa

The first conclusion from all this is that unrestricted flow data gives littel indication of installed performance. Charge coolers are relatively low flow/high pressure circuits, and a pump's ability to generate a high pressure into a completely blocked outlet is probably a better indication of installed performance (but we never actually get that information from anyone). Back soon

Nick
Attachment 409982

My goodness this was hard work.....

The pdf below is MUCH easier to read.

Attachment 409983

Great write ups ! ! !

I'm looking to add CXRacing HE to my car but I can not figure out where to find the space for it!

What I've done in the chart is to plot all the data points on the spreadsheet. The data comes from manufacturer websites and data sheets, and very little of it is directly comparable. Fortunately, circulation pumps tend to have simple characteristics, so I plotted what points I had available and ran polynomial interpolations to give a good fit. Most curves were simply second-order, but a few are slightly better with third-order. The graph shows the pumps' data points, but the curves are all interpolations

I also plotted two square-law cooling circuit resistance curves. These are characteristics of the load that the pump sees, and the point where the pump and resistance curves cross gives you the installed flow. The engine cooling resistance comes from Davies Craig data sheets, and is probably reasonable. The charge cooler resistance is based on what installed flow and pressure data I could find on-line, mostly on MBWorld. Its a guess, but I think its representative.

Of course, the charge cooler resistance is much higher than the engine, so an electric pump that's suitable for engine cooling is unlikely to be good for charge cooling, and vice-versa. I think the M275 charge cooler operates in the 20lpm / 30kPa region, so that's the installed pressure and flow that candidate pumps have to meet simultaneously. Its no good being able to dump 20l/min into a bucket, for example....

Most pumps have good data of some sort, but I have to admit I've come up short on Meziere WP136S data. The manufacturer only gives the open outlet flow, so I've guessed that its performance will be similar to the Johnson CM90 with the 20mm outlet. So there's a big health warning there, and if anyone knows any better, please say so.

Aside from that, I think we now have a good way to make fair comparisons between different pumps, and understand what's likely to work in a charge cooler circuit. Some of teh conclusions are:

1) Since a few of the pumps have optional porit sizes (CM10, CM90, EWP80) its interesting to see how that affects performance. Obviously the larger outlets flow more into low resistance, but once they're in a charge cooler system, outlet size makes little difference.

2) The difference between the old and new Bosch pumps is clear, they're both good with high-resistance loads, and you can see how much better the 0 392 022 010 is.

3) Davies Craig have the widest range, from the tiny EBP15 (actually a Bosch pump) up to the EWP115, which is top dog. Although that's designed for engine cooling, it still seems to have the highest static pressure of all, and it's still not that big or expensive.

4) I think the most interesting result though, is where the CM30 fits in. Although its small, cheap, light, and has low power consumption, there really aren't any performance reasons to chose it over even the old Bosch 0 392 022 002, let alone the Meziere.

5) The other thing that occurs to me is that the resistance curve is rather steep, so a bigger pump will only give a small improvement in flow over stock.

We may get more flow improvement from reduced cooling system resistance, than from a bigger pump. Its difficult to do anything about the charge coolers themselves (though they're a good size) so the scope for improvement lies with the heat exchanger. I'm thinking in terms of using a largerHE, or a second HE in parallel with the stock one. As well as the reduced thermal resistance, that would almost certainly improve the total system flow as well.

Indeed, the current drawn by each pump over their operating flow range (where given by the manufacturer) often increases with reducing resistance, indicating that the pump is drawing and delivering more power. If you imagine the resistance curve moving slightly to the right, towards the engine resistance curve, the crossing point will move down to a point with lower pressure and higher flow. High pressure in itself isn't what we need for cooling - its the mass flow rate, and a low resistance circuit gives us more flow for free.

There, that's what I've been meaning to get to the bottom of for a long time. Hope you find it useful.

Nick

good info thanks for sharing

There are two other things I wanted to understand, and I'm not really there yet.

The first is just how much the water circuit itself restricts the performance of the charge cooling system. It could be that charge coolers are like water cooled PCs. The gaming boys really want to keep their over-clocked CPU temps down, just we want to keep our over-boosted IATs down. Water has such high heat-carrying capacity, that a bigger pump or large bore pipes make very little difference to the processor temperature. The PC cooling bottleneck remains at the interface to the air - the heat exchanger.

Well, we know that water has high heat capacity - more than 4 times higher than air, by mass. It also has high heat conductivity - about 20 times higher. But those figures don't take into account that water has higher density - nearly 1000 times higher. Combine them together, and water's cooling capacity, volume for volume, is nearly 100,000 times greater than air.

When you look inside radiators and heat exchangers, the water passages naturally have much smaller x-sectional area than the air passages. I couldn't find much information about this, but I guess the water passages are about ten times smaller than the air. Even when you consider the extra area of the cooling fins on the air side, I'm sure that water has far more capacity. So could it be that a bigger pump on the V12TT makes little difference, aside from getting the coolant circulating faster, once it's switched on.

I think its clear that a better pump will circulate the water faster - though perhaps not very much faster. But I don't know whether it already goes fast enough?

Nick

The second thing I'd like to understand is the heat exchanger's influence on the cooling circuit resistance. I think typical charge cooler circuits are relatively low-flow/high resistance circuits, but I'm not sure where the resistance comes from. My guess is its the heat exchanger, and this is why. From air-air intercooler design, it seems that the optimum design is where the compressed air and ambient air channels have similar total volume within the intercooler. This suggests to me that an HE should have a similar volume to the charge cooler (the IC and the HE being two halves of a virtual intercooler - just separated by a circulating water interface).

In the case of the V12TT, the HE is rather smaller. Its also a different shape. If you consider the water flow through those narrow-section tubes, its a pretty slim HE, with relatively few tubes, and rather long ones at that. I think water flow through the charge coolers is rather easier; they're short, fat and deep by comparison,and they're connected in parallel. So conceptually, I imagine the two IC's as being stacked one on top of the other, air flowing one way, and water flowing orthogonally to that. If you stack them together their combined shape is close to being a cube.

When I think of the resistance to water flow, I think of the parallel with electricity. If you want to make a high resistance wire, you make it long and thin. If you want to make it low resistance, you reduce the length, and increase the width and height. In less hand-waving terms, resistance = (resistivity x length) / (cross-section area). Does that analogy hold true with water flowing through a heat exchanger?

Well, when you look inside the IC's, they seem have much the same configuration as a radiator. The water channels still have around 10% of the size of the air channels, and I think that's a bit like saying the IC and the HE have similar "water resistivity". There you might be able to get a measure of the relative water flow resistance by comparing the dimensions, which from memory are as follows:

IC: LxWxH = 20x20x17.5cm (height is two units stacked)
HE: LxWxH = 58x2.1x26.5cm

So what is the relative water resistance (arbitrarily assuming for comparison purposes that "water resistivity" for both is unity)?

IC: L/(WxH) = 20/(20x17.5) = 0.057 water resistance units
HE: L/(WxH) = 58/(2.1x26.5) = 1.04 water resistance units
Total cooling resistance = 1.10 water resistance units

So just going by their shape and dimensions, the HE would have around 18 times as much water flow resistance as the charge coolers, with much of that simply being down to the smaller number of water tubes running across the exchanger. Of course it won't be as simple as that, as some of the total resistance will come from the inlets and outlets to the tubes, but even its half that difference in reality, the HE may still have an order of magnitude more flow resistance than the IC's. And I guess that would probably dominate the behaviour of the cooling circuit - if we want to increase flow, reducing resistance is probably a good way to do it, and the HE is probably the best place to start.

Remember I said in the first post that the HE ought to be at least as large as the IC? Well, consider the big CX Racing HE, which has similar volme to the IC's. What would the water resistance of that be?

CXRacing HE 002: 24x8x2.5” = 61x20.3x6.4cm = 7,863cc
Water resistance = 61/(6.4x20.3) = 0.469 water resistance units
Total cooling resistance = 0.526 water resistance units

The CX HE has more than twice the core volume, and less than half the resistance, of the stock cooler. And the combined IC + HE resistance is also more than halved. It stands to reason, as larger HE's usually gain volume from height and width (their cross-section area, in fact). In other words, since the length is usually fixed around 24 inches, the core colume and the flow resistance will both improve in direct proportion to the cross-sectional area. To take an extreme example, if we used an S600 engine radiator (12 litres) as the HE, that would might have four times the volume and one quarter the resistance of the stock HE (and still not as low as the IC's).

I think that illustrates that the HE is the major bottle-neck in charge-cooling flow (as well as in the overall heat capacity) so there are double gains to be had in improving the HE. Flow would be significantly greater even with the stock pump, and with the IC resistance curve moving right towards to the engine resistance curve, there's a lot of scope for a bigger pump.

Nick

Thanks. What did you have in mind? Are you looking to keep the stock cooler and add another? Have you tried fitting anything, or just taken some dimensions? Have you taken the bumper off? That's not difficult once you've got the wheel arch liners off (and first disconnected the parking sensors loom!) For a big car, there's not a lot of space in there. Is your car the W220 or W221?

I think CXR make three types, the largest being 24 x 8 x 2.5" core. Similar size to the Frozen Boost cooler. I'd like to know if they fit, but I'm not sure. I did have my front bumper off a couple of months ago, but I was takling ABC problems, and didn't heve time to experiment with HE's. I'm not sure if the best place is under the bumper bar or behind it. Mounting beneath it, there seems little room to fit in the engine oil cooler. Going behind, its squeezed by the air con condenser. Whever you go, you gotta consider physical interference with inlet & outlet, headlamps, and don't forget the bonnet catch!

The thing that I really shouldn't be dreaming about is fitting a second radiator in there somewhere. Why mess around with small HE's when there are so many proper ones about? I like the idea of a full height HE, so ALL the air goes through the HE, AC and rad. That's how most modern turbos are configured. They don't have to be thick to have a large core volume, and don't forget that many rads have tranny coolers built-in. Why not run the ABC coolant through that? The steering oil goes through the AC. That would all make for a tidy front end. Of course I'm having difficulty finding a decent size rad that would fit in there.....

Nick

I'm thinking of replacing the factory HE with the CXR unit (24x8x2.5") and leaving the factory pump in place but I'm going to wire a switch from switched aux. power to run the pump whenever. There is just not enough room for those 2 HE i think and besides, the stock one is restrictive anyway. I agree with your point on increasing the flow won't allow the HE enough time to cool the liquid. With the more efficient HE and reduced resistance, the flow rate will increased slightly anyway almost the same effect as having higher flow pump. With having the pump run whenever the engine is running will probably help keep the IAT low and ready for any surprise runs.

Cheers

Excellent Info there :-)

I've been doing a bit more research, and I have to correct myself there. The S65 has a full-size HE in front of the condenser, but that's the exception to the rule.

Recent cars that were designed around turbo engines almost invariably have the condensor at the front, and the intercooler sandwiched in between. I guess the condenser doesn't like the heat from the intake cooler. Heat exchangers are usually arranged hot-to-hot and cold-to-cold, so if the rad runs hotter than the HE runs hotter than the AC, then that's the logical sequence. It seems to have become the rule these days, apart from some Fords, BMWs and Mercs that still have deep, low-profile HE's mounted low down and at the front, where they don't obscure the condenser.

Which opens up some interesting options... If its possible to move the S600 condenser forwards a few cm (might not be feasible) then there are some advantages in sandwiching a large HE in the middle. Although the condenser sits snuggly between the rad's end tanks, there are four AC & PAS pipes that are squeezed in front of the rad, and they make it difficult to put a large HE at the front. If the condenser can go at the front, that helps fitting the HE.

All the current units - Rad, AC & HE - are all cross-flow, but a down-flow HE with tanks at top & bottom, with the AC nested in-between, might be a good packaging solution. I'm not sure what that solution might be; packaging is definitely a problem. The HE outlets must still be able to get past the condenser, and that rules out several options, including all the purpose designed HE's that I've ever seen. Its not a problem that new cars have.

Nick

I'm being a bit stupid there, and have to correct that as well. The large CXR HE has a filler cap on the top, so you can't mount it under the bumper bar, like you can with other HE's. The filler would foul the bumper of course.

Realistically, there probably wouldn't be room beneath the bumper and above the engine oil cooler anyway.

I think it would have to behind the bumper beam, and if it does't fit there, I'm not sure what would - it would have to be something smaller.

CXR do a smaller HE that measures 21 x 6 x 2", and that doesn't have a filler, so it might fit under the bumper. It is rather smaller though - 4130cc vs 7866 cc, and its a dual pass HE with inlet and outlet at the same end (not what we want).

Nick

<FONT style="BACKGROUND-COLOR: #f5f5ff">
<SPAN style="FONT-FAMILY: Arial; COLOR: black; FONT-SIZE: 9pt">Late Model Racecraft 2012+ ZL1 Camaro Upgraded Heat Exchanger
Core: 616 x 413 x 50mm = 12,720 cc
Overall: 667 x 420 x 57mm
Cross-flow, Large HE, but length & thickness may be too high

Sorry about that last post, I tried to update it several times but couldn't edit it for some reason.

Here's a quick list of various heat exchangers with comments, plus comparisons with the stock S600 units:

Many of the obvious heat exchanger options are either relatively small or expensive or both. Since many modern turbo-charged cars have larger intercoolers than radiators, I thought I'd aim a bit higher with the V12TT and see if there are any engine rads that might work as an HE.

I'm looking for something with similar L x W x D as the S600 rad, but there aren't very many of those. There are plenty of smaller rads, but I wouldn't want the HE tanks to block the flow through to the AC condenser, so I want to fill the space as much as possible. The inlet and outlet have to clear the bumper, headlights, oil cooler, AC & oil pipes, distronic radar and bonnet catch, and the header tanks can't take up too much room.

On the wish list is a transmission oil cooler, for cooling the ABC oil. That way, I'd have the tranny oil running through the engine rad, the steering oil running through the AC condenser, and the ABC oil running through the HE. I really like the sound of that.

The other "ideal" solution is for the HE to be sandwiched between the radiator and the condenser, just like they do on most new turbo cars. That difficult because the HE would have to be "nested" within the radiator tanks, just like the stock condenser is. The HE hoses would also have to come forwards past the condensor, but it would give a very integrated, stock, appearance.

So here's my short-list, and I'm afraid there aren't any ideal solutions. There's advantages and disadvantages with each of them. I think the idea is worth pursuing though, as a radiator could make a very large and very cheap HE.

I found this AMAZING website by a fellow Brit called Elliot Mansfield, who has been doing remarkable things to his Land-Rover based Dakar racer. Its not so much the performance he's achieved, or how he's presented the website that struck me, but how he went about everything. He's an IT consultant with a soft spot for off-roaders, and has an unmatched mix of mechanical and software skills and enthusiasm. He restored the car, replaced the engine and turbo-charged it, fabricating all his own mounts and manifolds. He then built his own fuel injection system, and developed instrumentation systems to monitor what everything was doing.

The interesting thing for us is that he decided to develop his own intake cooling, and built several systems. He found out the hard way what works, what doesn't, and why. So he looked at different IC's, HE's and pumps, and measured the performance of each. He instrumented water and air temperatures and pressures at every point, and used those to diagnose what was going on & what needed to be changed. Basically, he found out what made charge cooling tick. I won't steal his thunder by trying to summarise the output of years of work, but will just strongly recommend that everyone here read this web-site.

http://www.mez.co.uk Home - Twin Turbo - Intercooling - Discussion

It'll take some time to read, but not as long as it took to do, and its an absolute gold-mine.

Nick

Up, Down or Cross-Flow? And where should the ports be?
 

I never paid much attention to where the coolant should go in and where it should go out.

Radiators used to have top and bottom tanks, and the coolant flow is usually downwards. This fits in with the thermostat outlet being situated above the water pump inlet, so it simplifies the plumbing. These days, most radiators are cross-flow (though most Japanese cars are still down-flow). Does it matter which direction the water flows? I wouldn't have thought so. Since the heat capacity of water is high, its temperature doesn't vary that much around the system. Its a low-flow / low temperature system, and other considerations may be more important. The V12TT engine bay is tight, so packaging matters.

Attachment 409984

Circulation pumps are usually connected to the outlet of the heat exchanger, which makes sense as this is the point of lowest dynamic pressure and temperature. Sure enough, the CL/S600 HE (above) is connected to the pump inlet, which means coolant flows from right to left. This is an upside-down drawing, the only one I could find. The inlet (right) is at the bottom, and the outlet (left) is 3/4 of the way up. This means that cooling water flows upwards, unlike cross-flow engine radiators, but I doubt this matters much, as the real temperature gradient is horizontal, across the matrix, not vertical.

How the system copes with air is probably more important. Water cooling systems are often arranged with the outlet of each component at the top, which helps to avoid air-locks. Engine radiators don't do this, but they usually have small bleed pipes to the expansion tank. The V12TT doesn't have an expansion tank, so it's difficult to get rid of all the air in the system. Since the HE outlet isn't at the top, I'd say it was impossible, and this may expalin why bleeding is so difficult.

I believe the HE is only filled and bled in situ, both at the factory and in the workshop, when the anti-freeze has to be changed. This makes me think that most HE's out there are part-filled with water - perhaps three-quarters full, up to the level of the outlet, with air filling the top part. That means its effective area is just 580x200mm, not 580x265mm (and the bottom of that area is kind-of-shielded by the engine oil cooler).

Upgrading the HE may be an opportunity to improve the air bleeding, either by putting the outlet at the top, or by using a filler cap, or by incorporating a bleed pipe and/or an expansion tank.

Does anyone know where the outlet is on the S/SL/CL65? If its in the same place as the 600, that means even more of the HE will be filled with air, so I hope the outlet is high up.

Nick

In another thread elsewhere, someone asked if a particular electric cooling pump would be suitable for the IC system, and I replied that its best to tread in someone else footsteps – someone who’s actually done what you’re thinking of, and let them make the mistakes first. Using a radiator for the HE comes into that category, and I don’t think anyone’s actually done it before. I’m convinced there must be a radiator out there that’s suitable, but I’ve been struggling to find something that will actually fit, but is also a worthwhile upgrade. I’ve done lots of research based on what I think is important, but since I haven’t actually tried it yet, I don’t know if I’ve thought of everything.

A gap in my reasoning just occurred to me. Finding space in the final, installed, position is one thing, but getting it in there is another. How would I actually fit it - from above or below, in front or behind? The inlet & outlet stick out a couple of inches, and one of them will probably have to pass the bumper bar, but there’s not much space. For this to be a good solution, I want to avoid removing the bumper or the radiator/headlamp support, but the engine rad may have to come out. So that’s two fills of anti-freeze, then – not the end of the World.

Knowing how long it takes me to do anything, the “do nothing” approach had a lot of appeal, but the better weather has encouraged me, and I’ve decided to buy a few candidate rads and try to fit them to my V12TT. So which ones go on the short-list? The ideal best candidates tend to be cross-flow rads from large cars, medium 4x4s or small vans and trucks, made in Europe during the ‘90s. Turbo-diesels with both auto gearbox and air conditioning are best. So which of the above candidates should I experiment with? Any of them? Nope – none. All have drawbacks of some sort or another, and by extending my search I came up with some better candidates. This is a very diverse list, and includes a couple of “home grown” solutions:

LAND ROVER Defender TD4 Radiator
Core: 554 x 438 = 2426 cm2 x 44 = 10,670 cc Overall: 655 x 456 x 66
Cross flow, manual, inlet close to bumper – 16cm up – may sit low
http://www.ebay.co.uk/itm/281071674398?ssPageName=STRK:MEWAX:IT&_trksid=p398 4.m1423.l2649

LONDON TAXI Radiator
Core: 537 x 432 = 2320 cm2 x 46 = 10,670 cc Overall: 570 x 520 x 60? + 30mm filler
Down flow, oil cooler, centre filler cap, small area, foul oil cooler? Too tall?
http://www.ebay.co.uk/itm/271097745803?ssPageName=STRK:MEWAX:IT&_trksid=p398 4.m1423.l2649

Radiator for MERCEDES 200D-300D (W124) - 62698A 1990 –
Core: 532 x 488 = 2596 cm2 x 40 = 10,390 cc Overall: 625 x 513 x 55
Cross-flow, oil cooler, narrow core, fits between Rad headers – sandwich candidate
http://www.ebay.co.uk/itm/310454376395?ssPageName=STRK:MEWAX:IT&_trksid=p398 4.m1423.l2649

RADIATOR - MERCEDES 190 (W201) 1990-1993 62721A
Core: 574 x 449 = 2577 cm2 x 40 = 10,310 cc Overall: 680 x 476 x 60
Cross-flow, oil cooler, too wide
http://www.ebay.co.uk/itm/310439288336?ssPageName=STRK:MEWAX:IT&_trksid=p398 4.m1423.l2649

Radiator for BMW X3 2.5-3.0-2.0D- 60803A
Core: 580 x 499mm = 2894 cm2 x 32 = 9,260 cc Overall: 652 x 521 x 58
Cross-flow, huge area, outlet may foul bumper, upside-down? No oil cooler
http://www.ebay.co.uk/itm/171018613226?ssPageName=STRK:MEWAX:IT&_trksid=p398 4.m1423.l2649

BMW E36 3-SERIES Manual Radiator
Core: 580 x 450 = 2610 cm2 x 32 = 8352 cc Overall: 652 x 470 x 51
Alloy: 580 x 450 = 2610 cm2 x 50 = 13,050 cc
Cross-flow, outlet may foul bumper, ideal width, fit upside-down?
http://www.ebay.co.uk/itm/230958208606?ssPageName=STRK:MEWAX:IT&_trksid=p398 4.m1423.l2649

So I’ve started buy a few trial rads – Land Rover, BMW X3 and Mercedes W124 250D, and see which will fit. Fingers crossed.

The W124 rad is even a candidate for the sandwich configuration, where the HE is sandwiched between the condenser and the rad. The oil cooler connections are tricky, as they foul the steering cooler and AC condenser connections, but if the rad goes upside-down, they all clear, and there's even room for the HE inlet above the top of the condenser. That would leave the AC condenser right at the front, very close indeed to both the bumper and the Distronic radar, but with nothing at all in front of it. Most current production cars with turbo engines have a configuration just like this.

Nick

I know this isn't capturing everyone's attention, but it will later, bear with me a little longer.

The first thing is to add the Pierburg electric water pumps to the discussion. These are probably best know as "The BMW pump", but its actually a range of pumps for different applications. They range from the 15W WCP circulation pump (for heating and cooling when the engine is off) to a 1100W monster to replace the biggest mechanical pumps. An electric pump is still a bit of a novelty, but following BMW's lead, I think they're likely to become commonplace. There are probably four of interest:

CWA 50 . Circulation pump, 50W , 6000 rpm, 24 lpm @ 0.60 bar
CWA 100 Circulation pump, 100W, 7200 rpm, 30 lpm @ 0.85 bar

CWA 200 Coolant pump, 200W, 4500 rpm, 120 lpm @ 0.45 bar
CWA 400 Coolant pump, 400W, 10000 rpm, 150 lpm @ 0.80 bar

The CWA 50 is used as a charge cooler pump on the recent BMW V8TT engine, and the CWA 200 is used as the coolant pump on a wide range of recent BM's. The circulation pumps have relatively low flow and high pressure, while the coolant pumps are obviously high flow. Like all good pumps, they're specified by their flow rate under pressure, so we know what their installed performance is going to be. Contrast that with the Meziere WP136S, which simply claims 20 gpm / 76 lpm open pipe. Like the Meziere, the big Johnson and Davies Craig pumps also claim high open pipe flow, but when I put the Pierburgs onto my Flow Characteristics chart, an interesting picture emerges.

For the purposes of charge cooling, these pumps rock. If you look at the Charge Cooler Resistance curve, which gives an indication of the pressure/flow characteristics for a typically constricted charge cooler system, you can see which pumps will give good installed performance. The CWA 50 beats all-comers so far, and the 100 is better still, even achieving one bar pressure differential at low flow rates. For charge cooler systems, these are two meaty pumps. I believe the CWA 100 is what Renntech use for their new IC pump, and it seems to be exactly what's needed. Shame about the silly price.

The CWA 200 & 400 coolant pumps are off the chart, flowing over 100 lpm, and I didn't even try to plot them. They're huge, but you can see they meet different requirements. Where the 50 & 100 (like the Bosch pumps) fit the Charge Cooler Resistance curve pretty well, the 200 & 400 are closer to the Engine Resistance characteristic. In fact, the Pierburg stats tend to back up what I've supposed are the installed characteristics for IC and engine cooling, with a big difference in the pressure/flow curves. If anything, I think the differences should be even greater, with an even steeper IC curve and a shallow coolant curve.

The other thing that's interesting about the Pierburg pumps is that they facilitate electronic variable speed control. They can be slowed down almost to a stop, for fast warm-up and to save electrical power. In theory, they can control water flow well enough to avoid the need for a thermostat. That might not avoid hot spots in the engine during warm-up, so I don't believe anybody has actually implemented that yet, but it sounds ideal for a charge cooler.

There's lots of interesting discussion, brochures and information about the Pierburg pumps here:

http://north.america.kspg-ag.com/ind...id=2195&lang=2
http://www.pierburgspa.it/pdfdoc/ksp...olant_pump.pdf
http://www.kspg.com/fileadmin/media/...lmittelp_e.pdf
https://mbworld.org/forums/w211-amg/...now-about.html
https://mbworld.org/forums/w211-amg/...-bmw-pump.html
https://picasaweb.google.com/renntec...50086765057074

And here's my updated spreadsheet - note that the scales have been expanded since last time, and I've added imperial flow and pressure scales. I've been meaning to do that for weeks; its frustrating when manufactuers quote specs in so many different units; its difficult to make fair comparisons otherwise, so here they are.

Attachment 409986

Nick

Nice Nick,
what do you think about this report?
https://mbworld.org/forums/w211-amg/...-vs-wp136.html

Great Write-Up Nick,

thanks for sharing. :y

How much heat does a charge cooler have to remove?

When you specify a system to do a job, you normally establish a set of requirements. However, I don’t really have a feel for what the charge cooler is doing, only that I want it to be better. Well, I thought it might be a good idea to understand how much heat it’s trying to dissipate. Figuring out the load on the engine cooling system isn’t too difficult. It’s well known that it dissipates much the same power as the engine produces, so that’s got to be up to about 400kW. That’s dissipated with a 12 litre radiator that runs at about 90oC – around 60oC over ambient. Cooling systems are mature and well-proven, so we know they work.

The charge cooler is a bit more difficult. We normally expect the output of the IC to be 35 - 46oC (pump control temps), but how hot would the intake air be if it wasn’t cooled? Well, when you compress air, it gets hotter, and the pressure increases by more than the volume decreases. If its compressed quickly and doesn’t have time to cool, then it’s an adiabatic compression, for which the equation is:

P1*V1exp1.4 = P2*V2 exp1.4 (where 1.4 is the ratio of constant pressure / constant volume specific heat capacity for air)

This shows that if the intake air is compressed to 2 bar at some temperature, then the volume of the air becomes 61% of the original volume. Knowing the compressed pressure and volume, we now use the general gas equation to get the compressed air temperature:

P1*V1/T1 = P2*V2/T2

Assuming the ambient temp is 300K/27C/80F, the compressed temp becomes 366K/93C/199F. So one bar of boost increases the charge temp by typically 66C just because of the increase in pressure, and that’s an intrinsic change – it doesn’t depend on engine size, speed or power. Having said that, it will probably be higher, as hot turbos tend to heat the intake air even when it’s not being compressed. And if the boost is more than one bar, both the temperature and the mass flow will be higher, so there’s a double whammy for tuned engines.

So how much air does the V12TT use per second? Allowing for a realistic volumetric efficiency, 5000 rpm times 5 litres equals about 200 litres/sec. The density of 2 bar air is about 2g/litre at those temps, so that’s 0.4kg/sec of air flow at full power. The heat capacity of air is 1.0 kJ/kg/C, so assuming the charge cooler needs to drop the temp at least 50C, that means a cooling dissipation of 0.4kJ/s/C x 50C which equals 20kW.

That’s a useful figure to know, and shows that the IC isn’t putting much extra load on the engine rad by heating the incoming air too much. It’s about 20 times less than the engine cooling system – but of course there’s less temp difference available to do it. There’s another interesting comparison to be drawn with engine cooling. The Pierburg documents I linked above suggest that a 200W electric pump is sufficient for a 200kW/270bhp engine, so by the same logic, an IC system should only need one of the small Johnson or DaviesCraig pumps. An 80W Meziere or Pierburg pump sounds overkill - except that IC systems are definitely low-temperature, high-resistance, systems.

Nick

Now, what about the IC coolant? Looking at the pump characteristics chart, it seems the normal operating point is around 5gpm/20lpm at 30kPa. That’s just a third of a kg/sec, and even the best pumps only get that up to ½ kg/sec. So the coolant mass flow rate happens to be very similar to the air mass flow rate – though we do need the reduction in air temp to be greater than the increase in water temp. The heat capacity of water is 4.18kJ/K/kg, so the coolant has a capacity of about 1.4kW/K – perhaps 2kW/Kfor a big pump. At full power, that means a temp increase of 20/1.4 or 14C, or around a quarter of the air temp reduction, as you’d expect from the flow rates and heat capacities. I’m not certain, but from a few articles I’ve read I think that’s pretty similar to the temp drop across the engine radiator. Most people expect to see 20F across any radiator, whether it’s on a go-cart or a nuclear power station.

My first impression is that the stock IC system was designed against the requirements of the stock engine, at best.

Second, I hadn’t considered that tuning would increase the cooling requirements exponentially – due to the collective increases in intake air pressure, temperature, density and flow rate.

Third, I’m also surprised that the heat carrying capacity of the water isn’t higher – if it was comfortably high enough, the temp delta might be a few degrees, but 14C sounds kind of marginal for a low temp system.

Fourth, in my earlier reasoning I hadn’t considered how high the intake air mass flow rate - and hence the cooling requirements - would be. If the coolant rises 14C in one cycle, that doesn’t give it much margin or storage capacity – just a few seconds at WOT and it’s all heated up.

Finally, none of this considers the heating contribution from the turbos themselves – aside from the compressing of the air, that is. I don’t know what that would be, but it will only make the IC system’s job harder.

Nick

Here's a rather scrappy scratchpad of links and ideas, with a bit more evidence to support the requirements and implementation of charge coolers.

http://f10.m5post.com/forums/showthread.php?t=588858

Interview with, Jürgen Poggel, Head of Engine Development at BMW M Power GmbH
The charge air exiting the compressor measures as much as 130 °C.
The turbo-chargers are mounted in the middle of the V, with the inlet manifolds where the exhaust would be. Bizzare.
The M5 has a cross-over exhaust manifold where each turbo takes exhaust from both cylinder banks.
The temperatures of the exhaust manifolds and turbochargers can be as high as 1000 °C and the oil hits 200 oC.

Attachment 409987





Low temperature circuit for cooling the charge air and the Digital Motor Electronics:

A Coolant radiator
B Additional coolant radiator
C Electric coolant pump for cylinder bank 1
D Charge air cooler for cylinder bank 1
E Expansion tank
F DME cylinder bank 1
G DME cylinder bank 2
H Charge air cooler for cylinder bank 2
I Electric coolant pump for cylinder bank 2
J Additional coolant radiator

The charge cooler has two CWA 50 pumps and THREE heat exhangers.
Two of them measure 25 x 25 cm, and the other looks like 60 x 25 cm.

http://www.realoem.com/bmw/showparts.do?model=LX91&mospid=55181&btnr=17_0617& hg=17&fg=30

Attachment 409988

http://www.michelecaroli.com/autopar...PORSCHE&tpr=S3

Porsche 911 Turbo air-air coolers measure 2 x 325 x 220 x 80 mm = 11.4 litres, equivalent to 5.7l charge cooler plus 5.7l HE.
These are the largest stock IC's I've found. The Panamera Turbo and Audi Q7 IC's are slightly smaller - about 10l.

http://www.autospeed.com/cms/A_110200/article.html

The Corvette ZR1's LS9 Supercharged Small Block
Maximum boost pressure is 10.5 psi (0.72 bar).
Because the pressurized air is hotter than naturally aspirated air, the LS9 employs a liquid-to-air charge cooling system to reduce inlet air temperature after it exits the supercharger – reducing the inlet air temperature by up to 78 degrees C (140 F).


http://www.are.com.au/Big%20HP/Craig...sAWComorig.htm

http://www.are.com.au/Big%20HP/Craig...owww6pages.pdf

Extreme turbo-charging: 1895 whp from 6997cc @ 8000 rpm @ 40psi boost
Air intake mass flow: 2 kg/s
IAT: 230 oC
Charge cooling requirement: 409 kW

http://stratifiedauto.com/wordpress/...azdaspeed6.jpg

Two dyno runs on a Mazda 6 turbo:
230 bhp @ 88 oF ambient
262 bhp @ 66 oF ambient

12 oC air temp reduction gives 13 % power increase.
Using PV = nRT, the temp increase gives a 6.5% reduction in air density.
The remaing 6% power loss is probably down to ignition retard or boost reduction.

By way of getting a feel for what charge coolers do, and what makes them tick, I've been expanding on some of my earlier sums. I used some simple assumptions and thermodynamics equations to estimate how much thermal energy and power was involved, but that was only a baseline. Upgraded intercooling is often wanted for re-mapped engines, but I said before that re-mapping makes things exponentially harder for the IC. But how much harder? I ran the sums for a range of absolute MAP starting with ambient.

P,V,T,D are absolute pressure, volume, temperature, density & 1 = ambient, 2 = compressed

P1. . . . . . V1 . . . . . T2/T1 . . . T1 . . . . T1 . . . . D1
1.0 bar . . 1.0000 . . 1.0000 . . 300K. . . 27C . . . 1.000

P2. . . . . . V2 . . . . . T2/T1 . . . T2 . . . . T2 . . . . D2
1.5 bar . . 0.7485 . . 1.1228 . . 337K. . . 64C . . . 1.336
1.7 bar . . 0.6788 . . 1.1676 . . 350K. . . 77C . . . 1.473
2.0 bar . . 0.6095 . . 1.2190 . . 366K. . . 93C . . . 1.640
2.5 bar . . 0.5197 . . 1.2993 . . 390K. . . 117C . . 1.924
3.0 bar . . 0.4562 . . 1.3687 . . 410K. . . 137C . . 2.192
3.7 bar . . 0.3913 . . 1.4555 . . 436K. . . 164C . . 2.556

V2 = V1*(P2/P1)exp-0.71428 (negative reciprocal gamma)
T2/T1 = (P2/P1)*(V2/V1)
D2 = D1*(V1/V2)

These figures show what happens to compressed intake air if you don't cool it down. The density column, which is normalized to ambient, is particularly interesting as it tells us how much extra air mass flow you get for the extra boost pressure. If you could keep the IAT down to ambient, D2 would be the same as P2, but it illustrates the loss of charge density and presumably torque caused by the IAT increase.

Consider the case of a tuned turbo with the boost turned up from 1.0 bar to 1.5 bar. Relative to stock, the mass flow goes up by 1.924/1.640 or 17%, but the air temp goes up 90/66 or 36%. Therefore the total extra thermal load power on the charge cooler goes up by 1.17 x 1.36 or 60% over stock.

That 17% extra air mass flow probably translates quite directly into a 17% increase in torque, from say 600 to 700 lb/ft, which sounds about right. Turbo tuning doesn't increase power as much as torque due to turbo throttling and volumetric efficiency, so real world peak power probably goes up 10% from say 540 to 590 bhp.

Therefore a 10% power increase costs you a 60% increase in charge cooler thermal loading. That's what I meant by exponetial increase. So for a typical tuned car, the IC needs to remove 20kW x 1.6 = 32kW heat power. Divide that by 1.4kW/C stock cooling capacity gives 23C, or 41F, which is the coolant temp increase in one pass through the IC. In practice it's impossible because it exceeds the temperature budget, and the coolant temp and the IAT have to rise. It's about twice as high as it should be for a high temperature cooling system, and much too high for a low temp system, where the coolant is close to ambient temp.

That's why I believe the stock IC, for tuned cars in particular, needs to be heavily upgraded. Of course most car manufacturers have figured this out now, such that the current Golf 1.6 Diesel appears to have greater cooling capacity than the S600TT. Truck makers have known this for a long time. By a process of engineering evolution, that's simply what's needed to make turbos work.

Nick

As I see it, the turbos put energy into the intake charge in three ways:

1. Heating the air by compressing it (as above)
2. Heating the air because the turbos are hot (don't understand yet)
3. Increasing the pressure of the air

I understand (1) through the posts above. Its a necesary by-product of forced induction, and it also happens with natural aspiration during the compression stroke. The heat energy put into the air needs to be removed by the IC.

The figures in yesterdays post don't quite match up - the IAT is higher in practice than (1) predicts, so I assume that (2) is a significant effect. It must be a function of turbo heating under load, and I don't really have a handle on it. Maybe empirical evidence is more relevent that analysis - ie: just go measure it.

The useful part of what turbos do is (3). That's what we need them to do, and when we try to get rid of (1) and (2) its just throwing away waste. however, I didn't have a feel what the useful power was, so here it is for interest. I calculated this for circulation pumps with a representative example:

• Power = force x displacement / time = pressure x area x velocity = pressue x flow rate
• Water Power = 40 kPa x 30 l/min = 40 kPa x 0.5 l/sec = 20 W

Intake air flow runs at not dissimilar pressure and mass flow, but the volume is three orders of magnitude higher:

• Air Power = 100 kPa x 200 l/sec = 20 kW (or 30kW @ 1.5 bar)

That's up to 40 hp, so its quite a lot of power, and partly tells us why a supercharger needs a strong belt. The pumping power and the heat power are similar, so it tells us that half the work done by the supercharger or turbocharger just goes into heating the air, and needs to be thrown away. Only about half of the mechanical power actually goes into doing useful work and compressing the air. This doesn't have much bearing any anything to do with turbos, but I thought it was interesting.

Nick

I'll have something really worthwhile to post about in a few days, but for now I just had an idea to put down on paper.

At the beginning of the thread I supposed that the cooler and the HE in a charge cooling system ought to follow in the tracks of well-developed air-air coolers - they should have a similar volume. When you look at the insides of an air-air cooler, the total combined volume of the cooler occupied by the air that's being cooled is generally similar to that occupied by the air that's doing the cooling. That sounded like a reasonable starting point to me.

Then today I wondered if there was another, technologically mature precedent for the optimum design of heat transfer systems. Something where heat is taken from inside the car and disspiated outside the car? Air conditioning of course! There's an evaporator "inside" the car that absorbs the heat, and a condenser "outside" that dissipates it. There's a heat pump involved of course, but I think the principle is similar.

I already said in the first post that AC condensers had matured and converged on a consistent design between systems. In most cars, the condenser has a similar area to the engine radiator - around 600 x 450mm - and is almost invariably 16mm thick. The volume of the core is typically between 3 - 4 litres.

But what about the evaporator in the HVAC system? If my principle has some credibility, then the evaporator will also have a similar volume. So I did a survey of a large number of evaporators today, and found that they're similar in size to the heater matrix, but rather thicker. We're talking about half the width and height of the condensor - around 300 x 200mm - but much thicker - about four times as think in fact, about 65mm. The Audio A6 is an exception that proves the rule, having a huge 6 litre core, but otherwise they're similar. At first I thought the W220 had an unusually small evaporator for any car, let alone a luxury car, but then I realised I was looking at the seat cooling HE, which only measures 2.2 litres. The cabin HE is normal size.

They don't vary much from car to car, and it seems that everyone has figured out the best way to do air conditioning efficiently these days. Its mature technology and well-developed. And it just happens to have the same solution as my conjecture, where the cooling HE has about the same volume as the cooled HE. Granted, they're a different shape, just as they are with charge cooling, but I think the volume of the core is key.

I never thought that before, but it is another reasoned argument why the HE should be as big as the cooler, rather than a fraction of its size.

Don't worry if this is getting boring, the best is yet to come....

Nick

Nick,

What do you think of the Killer Chiller system. They've been around for a while. They're made for Ford Mustang and Lighting, but I'm sure it can be adapted to our cars somehow. I think someone has it installed on an E55.

http://www.killerchiller.com/system_2_page.htm

https://mbworld.org/forums/w211-amg/...gree-iats.html

I think a few people have installed Killer Chillers in supercharged Mercs, but I’m not sure what the practical benefit is, and have some questions.

What’s the installation – use with a heat exchanger, or with a trunk tank?
What’s the application – street or strip?
What’s the objective – lower temps than passive cooling, or lower than ambient?
Continuous or intermittent use?
Stock or tuned?

I doubt that a Killer Chiller has much capacity compared with a conventional system, I think a charge cooler needs to dissipate more heat power than the AC system can cope with. We’re probably talking about a few kW vs a few tens of kW. It’s only likely to be effective for occasional strip use where you want to get the IATs below ambient. You’d probably have to use a cold water tank to "store" the cooling capacity. That heat has to go somewhere, and with a Killer Chiller it ends up being dissipated by a 16mm thick condenser – I’d rathe rit was dissipated by a 40mm thick HE. Also, how do you know that the engine power gain from a Killer Chiller isn’t offset by the extra load on the AC compressor? Or do you run the AC for a while to cool the trunk tank, then turn it off for a run?

There are some weak points in the stock IC system, and I’m trying to find simple, efficient and effective upgrades that keep the car as a daily driver, and as close to stock as possible. I want it fast – yes, but also comfortable, quiet, reliable, practical and reasonably economical too. If I justwanted fast, I’d remove all the seats and trim. I’ve done that before, and great fun it was too, but I’ve moved on since then.

If somebody wanted to add a Killer Chiller to a stock system and keep the HE, then the mods I’m talking about here would be complimentary to that, and work well together. A better pump would probably help with the added coolant flow resistance from the KC. If you want to use a trunk tank, I think that’s a completely different path to take.

Nick

Nick

So based on your research so far, what pump would you recommend with an upgraded heat exchanger? Is the Johnson CM30 enough to move around the coolant, or do we need a pump with a higher flow rate.
Thanks.

No, I think the CM30 is quite unsuitable, even for stock systems. There's no reason to chose that over either of the Bosch pumps.

I'm leaning towards using one of the Pierburg pumps that are used by BMW, and I've just bought one.

nick

Nick

I have an extra CM30 lying around. What would happen to the flow rate if I was to add this extra pump before the heat exchanger and have it turn on with the other pump simultaneously. I would speculate the flow and pressure to increase twofold. Any thoughts or comments.
Thanks.

Good question, I hadn't thought about that.

If you have two similar pumps, you will definitely get more flow, though not twice as much, as that corresponds to maybe eight times more pumping power. Power equals pressure times flow, and pressure is proportional to flow squared.

If you combine two different pumps, its a bit more unpredictable. If the bigger pump gives more flow as installed than the small pump does open outlet, then the small pump will actually slow the bigger pump.

Look at the pump comparison chart in my April 21 post, and compare the Bosch and Johnson pumps. The installed performance of the Bosch pump is where the red curve crosses the Charge Cooler Resistance curve (about 22l/min @ 37kPa for the -010 pump). That point corresponds to a flow rate LOWER than where the CM30 curve crosses the horizontal axis (30l/min) - therefore the Johnson won't be slowing down the Bosch, and it should ADD to the circulation flow rate.

How much it will add is difficult to predict. I've never done this before. I think you could add the Johnsons' flow rate to the Bosch flow rate and generate a combined pressure/flow curve, and see where that curve crosses the Charge Cooler Resistance curve. I think that would be the new flow rate. The Bosch flows quite a lot more than the Johnson, and the resistance curve is quite steep, so the flow gain won't be that great. I'd estimate the flow would go up from 22 to 24 l/min.

Nick

Nick

If you double the size of the front mount heat exchanger and double the size of the air to water intercoolers would you think a double pump setup would be more logical?
In theory you are increasing the cooling capacity two fold and therefore would need extra push in volume. I do believe that a double pump setup would move more liquid and put less of a stress on just a single pump.
Curious as to what you think.

Yes, I think that would make a lot of sense. If you want to double the flow, you need to double the capacity of everything, not just the pump.

You can always use brute force and ignorance, but increasing the pressure in itself doesn't help - its flow that counts. And if you want to increase flow, I think the best way is both the pump and the HE.

I doubt that the charge coolers themselves cause much flow restriction, as they have lots of tubes, but the more the merrier!

Only thing with uprating the charge coolers is what to do with the air filters. If you lose the cold air intakes, you undo a lot of the work with the improved cooling. I'm not sure how hot is the air in the engine compartment, but all of it has come through the radiators, so I bet its pretty warm.

Nick

OK, this is all very interesting, but where's this thread heading?

We're not there yet, but this is the next piece of the jig-saw:

I first posted this in another forum, but it seemed to go unnoticed. I think that many charge-cooled turbo cars use continuous-running circulation pumps - Lotus Carlton, Jaguar XJR & XKR etc, and the blown Mercs seem to me to be at a disadvantage with their modest, thermostatic pumps. When looking for other pump options, like many others I found the Davies Craig pumps, especially the EWP 80 and 115. What was particularly interesting about Davies Craig was the pump controller. This implemented thermostatic control not only of the big Davies Craig water pumps, but any EWP in principle, and allowed you to modulate the pump output and adjust the target temperature in a fairly crude way. The BMW circulation and cooling pumps discussed in the earlier post are more effective and controllable, but the only way to control them is with a BMW ECU. That is (as they say on Top Gear) until now....

Last year a small German company called Tecomotive developed an electric pump controller (the tinyCWA) specifically for the Pierburg circulation and cooling pumps used by BMW: the CWA 50, 100, 200 & 400. With a suitable temperature sender, this enables thermostatic control of the BMW pumps, so that any car can benefit from their efficiency and flexibility. Like the Davies Craig controller, you can adjust the target coolant temperature, but it has much more flexible controls. Here's the web page and here's the users manual. The pump can set to run continuously at variable speeds, it can be run manually, and the controller can display the coolant temp or the pump speed. This seems to be intended for engine cooling, but I believe it can be used with the 50W and 100W circulation pumps as well.

Sometimes I think our cars are unfathomably complicated, yet sometimes I realise you only need to understand a little bit about how they work to realise that there's no black art or mystery to them.

I had assumed that the charge coolers themselves were simple cross-flow heat exchangers. The air flows inwards towards the center of the engine, and the coolant flows cross-wise. But it only takes a simple inspection to realise that its not like that. The charge coolers are obviously dual-pass coolers, since the water inlet and outlet are on the same side - at the rear of the coolers. Look closely at the coolers and you can see diagonal weld-lines at the front, that correspond to a shallow triangle-shaped header tank to turn the coolant round.

They're not like dual pass radiators, where each pass is parallel to the air flow - two coolant streams cross two different air streams. The V12TT coolers are series dual pass, where a single air stream flows across the two coolant streams. More than that, they're counter-flow heat exchangers, where the air and the water (kind of) flow in opposite directions. The MB implementation is like two coolers in series - a high temp cooler and a low-temp cooler. The hot air flows into the hot cooler first, then into the cold cooler. The cold water, by contrast, flows into the cold cooler first, then into the hot cooler.

This is how the best heat exchangers work, and it allows the air coming out of the cooler to be at a lower temperature than the water coming out of the cooler, which wouldn't be possible in a single cross-flow configuration. There are three main types of heat exchanger - parallel flow, cross flow and counter-flow - and these are a hybrid cross/counter flow configuration. Although this means there's more air and coolant resistance, both the air and coolant paths are still pretty short, and I think these coolers have an advantage over other coolers - and definitely over all air-air intercoolers.

Nice.

Thanks for bearing with me; this is where it starts to get interesting.

I contacted Tecomotive, the manufacturer of the tinyCWA Pierburg cooling pump controller. Like the Davies Craig pump controller, theirs is designed for engine cooling, so the temp range is appropriate for that application - 75 to 100 deg C. So I asked Tobias Mucke at Tecomotive if they could modifiy their controller to work over a temp range more appropriate for a low temp charge cooler system. And guess what - they said yes!

The changes are implemented in firmware, and are quick and easy to implement. There are sixteen programmed modes of operation (O to F), and Tecomotive offered to set any program I wanted. The stock controller has normal and pulsed modes, which are applicable to thermostat controlled cooling systems, and are deleted for charge cooling. You can see the stock Tecomotive user manual here:
http://www.tecomotive.com/download/manual_tinyCWA.pdf
After lots of questions and emails, we arrived at the following charge cooler program:

Pos. Mode TargetTemperature Description

0 Testmode - Pump Off / Fan Off
1 Testmode - Pump to min. rev. / Fan Off
2 Testmode - Pump to 50% / Fan Off
3 Testmode - Pump to 100% / Fan On
4 Normalmode 0°C The controlled temperature range is plus minus 5°C / 9°F from the currenttarget temperature.
5 Normalmode 5°C
6 Normalmode 10°C
7 Normalmode 15°C
8 Normalmode 20°C
9 Normalmode 25°C
A Normal mode 30°C
B Normal mode 35°C
C Normal mode 40°C
D Normal mode 45°C
E Normalmode 50°C
F Normalmode 55°C

In addition to these operating and test modes (which are great for installation and bleeding), there are lots of advanced modes to make the pump do whatever you want. You can set the pump to work under thermostatic control, or at fixed speeds, or permanantly full on or full off. I think the best mode is with a fixed minimum speed, so the pump is spinning slowly all the time (and you can set the minimum speed to whatever you want).

The knob on the controller sets the mode - essentially the target temperature - and the LED display gives an indication of the actual temperature, or the actual pump speed - whichever you chose to display. The displayed temp range is also customized to match the charge cooler target temp range. Its really a very clever little box; so many tunes to play. Its a new derivative called the Charge Cooler Special, and there's a revised user manual that I could send to anyone interested.

And I've got the first one!

Nick :)

Nick,

In your hypothesis, have you put into any consideration that the Killer Chiller system has a slight advantage with the help of R-134a freon with a boiling point of about -15 *F vs. water at +212 *F at atmospheric pressure?

I will try to answer some of your questions as best as I can:

What’s the installation – use with a heat exchanger, or with a trunk tank? -You can use it with or without heat exchanger. For better result, use it with HE but have the coolant flow into the HE first then Killer Chiller then intercoolers. KC can chill the coolant to below ambient temps and you wouldn't want it to get reheated by the HE. If you're not planning to use HE, I would highly recommend a trunk tank. This way you can cool down your coolant and have it in reserve while driving normally and when needed for those special situation, you'll be ready.

What’s the application – street or strip? -Either or.

What’s the objective – lower temps than passive cooling, or lower than ambient? -To get temps lower than ambient which is not possible with a straight up HE set up no matter how big the HE you can put on the car.

Continuous or intermittent use? -If use with HE, can be continuous.

Stock or tuned? -Either or.

I doubt that a Killer Chiller has much capacity compared with a conventional system, I think a charge cooler needs to dissipate more heat power than the AC system can cope with. We’re probably talking about a few kW vs a few tens of kW. It’s only likely to be effective for occasional strip use where you want to get the IATs below ambient. You’d probably have to use a cold water tank to "store" the cooling capacity. That heat has to go somewhere, and with a Killer Chiller it ends up being dissipated by a 16mm thick condenser – I’d rathe rit was dissipated by a 40mm thick HE. Also, how do you know that the engine power gain from a Killer Chiller isn’t offset by the extra load on the AC compressor? Or do you run the AC for a while to cool the trunk tank, then turn it off for a run? -The Killer Chiller doesn't need more capacity than conventional system because it's using Freon with -15 *F boiling point compared to 212*F for water. If you're using the KC system with HE, you can run your car hard all day long. At the very least, you'll be on your HE. Like I said earlier, for best result, use with a trunk storage tank and existing HE. Running the KC is not going to put extra resistance on the AC compressor that it's not already doing. It's not like an alternator where to put out more juice, it would require more power to run it. The AC compressor is just going for a ride where it's resistance is relatively constant whether you have your temps set at 65 or 75 with outside temp 95 degrees. The only tid bit is the AC compressor may cycle a little longer when you're asking more from the AC system.

I agree that Killer Chiller is the only way to get inlet temps below ambient, but I don't understand the significance of the relative boiling points? The whole point of using (liquid) water as a coolant is down to its high boiling point and high specific heat capacity. If water vapour was used for cooling it would be pretty useless.

There are two interesting questions though. Air conditioning is a heat pump, and it takes heat from a relatively low temp source (car interior) and pumps it into a relatively high temp sink (exterior). A Killer Chiller does much the same thing, but is rather different to a conventional charge cooler, as the refrigerant temp in the condenser is both higher than the sink and higher than the source. That high temp is desirable, as it gives a high temp differential to the ambient air. That allows the heat exchanger to pass a lot of heat to the ambient air, without having to be very large. I don't know what temp the condenser runs at - it would be interesting to know - but I assume its generally higher than charge cooler coolant.

When you upgrade a charge cooler HE, you want to achieve a lower IAT with the same ambient temp, and the bigger the HE, the smaller the temp differential you can achieve. That's what I'm trying to do. Killer Chiller doesn't suffer from that limitation as the refrigerant is always hot, so I guess that comparing HE sizes doesn't tell you very much.

However, I still think an AC system is limited by the capacity of the pump. The heat is lost to ambient air largely through the latent heat of atomization of the refrigerant - the heat that the refrigerant gives up when it condenses in the condenser, which is a fixed figure for a given mass flow rate. So there's only so much heat that an AC system can absorb and dissipate, and that's the cooling capacity of the system. Its not an easy figure to find, but I did do some research before my earlier reply. From what I can find, most domestic and automotive AC systems have a capacity of maybe 5 kW, something like that. Our charge cooler systems have to dissipate at least 30kW. You can get 30kW AC systems, but they're the ones that are used on full-size buses. They're big, heavy and expensive, and use a lot of power. I just don't see how a Killer Chiller can dump that amount of heat continuously. What it CAN do is "store" the cool in the trunk tank water, and give you sub-ambient IAT's for intermitent use. Given how long we can actually use WOT on our cars, that's pretty useful.

Nick

The only reason why I mentioned the boiling point of R-134a is because that's what it makes so effective in reducing the low temp circuit coolant. I'm not suggesting replacing the liquid water coolant with Freon, but merely just saying the killer chiller is a great aid in reducing the IAT temp. KC works by having an AC evaporator in a tub of the intercooler coolant. It works so well that the IAT can get below or near ambient temp. The AC can remove a great deal of heat due to the use of Freon. This gives it a great advantage over straight up HE set up. I think this is the route that I will go with since it's a lot more compact than trying to add or replace the existing HE with a larger unit. I would plumb the KC unit so that I'm picking the Freon after the interior's evaporator so that I don't compromise the cooling of the cabin. This way I would be utilizing the Freon going back to the compressor which would still be cold but not as before the main evaporator. I will just have the KC compliment the existing HE. Another nice feature of the KC is it can reduce IAT even while the vehicle is not moving due to the AC system working. On a conventional HE set up, I'm just helplessly watching my IAT creeping up while stuck in traffic.

Our cars already have AC system so it's not like we're bolting on additional compressor or extra alternator where it would draw additional horsepower from the engine. If anything, it might cause the existing compressor to cycle longer. If you have a 3-4 gallon tank in the trunk, it can probably store enough cooled coolant that is good for several minutes of WOT runs. If you plan to do a sustained WOT run then you still have your existing HE. If I were to go to a 1/4 mile track, I would probably only need it for about 11-12 seconds. Even on a 5 miles high speed run, you're probably not going to be WOT for more than 3 minutes.

I get what Killer Chiller does, and I understand that its a great solution for some people, but I think there are some misapprehensions about how it does it. Although heat pumps can be pretty efficient at moving heat, you still have to put energy in to compress the gas, so it condenses. That does take power - the pump takes several kW from the crank pulley, and in return it removes several kW from the intake air. You don't get something for nothing. Its not like an IC circulation pump that takes 100W, and removes tens of kW from the intake.
I don't think that would work, as the freon that comes out of the evaporator will have evaporated. Therefore it can't evaporate again, and remove any more heat, in the Killer Chiller. I read a KC installation manual a while ago, and I recall that you have to connect the KC & AC in parallel.
Now that's just bragging!

Nick

Killer Chiller doesn't take any more power from the crank anymore than if you were to run your air conditioner on a warm day. The pressure in your AC system will still be about the same as before the install. It's not like you're going to double the AC system's pressure hypothetically where it would put extra load on the compressor. I did just read the installation manual, they recommend to have the Freon to flow through KC first then the cabin's evaporator. They do not provide parts to run it in parallel. Once the Freon converts from liquid to gas (-15* F) at the orifice tube then KC then evaporator it does not completely loses its cooling ability as it makes its way back to the compressor which is why you will notice your low pressure side line to the compressor will be cold to touch after the AC have been running for a while. If that wasn't the case then you would not have air conditioning in the cabin working after the KC install. The price that you'll pay is may be it will take a little longer time to cool the cabin on a really hot day. Although they measured the air temperature coming out of the dash after the install and it was in the lower 40* F range on a 80 degree day which would be good enough for me.

If you're still concern losing a few hp running the AC compressor while you need maximum power then I would recommend a good size trunk storage tank. This can store the chilled coolant while cruising around and when you need maximum power then just turn off your AC system. It'll be sort of like an energy capacitor. Don't forget, you'll still at least have your factory heat exchanger as a back up.

Links to installation instruction. (Not for specific Mercedes).
http://www.killerchiller.com/tech_help.htm

I think I see where you're coming from. (For a given engine speed) the KC won't take any any more power through the AC pump than the AC alone can take itself. There's only so much pumping that a pump can do. However, our cars have variable displacement compressors, so although the max pumping capacity is fixed, it infinitely varies almost down to zero, depending on the AC demand. So I'd expect the effect of the KC was to put more load on the pump. Not sure how it would work with the pump displacement being controlled by the AC, but I guess it would be OK.

You're right about the KC HE being in series with the AC evaporator - my bad. Its in series with the charge cooler HE as well. I was surprised about that, but I guess it works.

I'm not saying that KC doesn't do what it claims to do. There's plenty of anecdotal evidence, drag strip results and dyno runs to show it does work. However, my point remains that you don't get something for nothing. Cooling the IC coolant takes power, we're talking kW rather than tens of Watts for a circulation pump, and that power has to come from somewhere - the crankshaft. Granted, the AC pump can't take more power than the max AC-only load, but with KC you'll be running at higher load for more time.

The consequence of this will be more fuel consumption, but if you're chasing an extra 50bhp, that probably won't be of any concern whatsoever. Neither will the fact that KC only keeps the temp down to 50-55F for short duty cycles. I only have my foot to the floor for a small percentage of the time, and KC would probably be ideal.

So I'm not trying to persuade anyone not to use KC - its probably a great solution for a lot of people. Please fit it to your car and tell us how you get on. I'd expect you to be very pleased with what it does.

I'm not going to fit it though, and that's not what this thread is about. I'd like to draw a line under this and move onto my next tedious chapter of charge cooler improvements.

Nick

So here's a few pictures of what I'm trying to put together for my car

Here are three engine radiators - two Mercedes and one BMW, plus two BMW engine coolant pumps and the Tecomotive electric pump controller.

Attachment 409989

Here's a close-up of the pump controller - its only about two inches across. This is the first pump controller with customized firmware for charge cooling systems. The knob in the middle sets one of sixteen operating modes - essentially the target coolant temperature, plus a few test modes. There's a row of seven LEDs that can display either the coolant temp or the pump speed, depending on the advanced settings. In this picture its showing the ambient temp of my garage, which is about 20degC.

Attachment 409990

And here's the BMW radiator alongside the stock heat exchanger. All the radiators I short-listed for trial installation are about the same width as the stock HE, but about twice the thickness and twice the height. The old Bosch pump is sat alongside the Pierburg BMW pumps. The Tecomotive controller works with any Pierburg electric water pump - there are several variants. They're used on most current petrol engines, except the turbo-charged sixes and eights.

Attachment 409991

One of the advantages of the full-height HE is that all of the air going into the engine compartment has to go through the HE. In the stock configuration, the HE (and the ABC cooler) present the incoming air with a high-resistance, alternative flow paths to going through the condenser and the radiator alone. I had presumed that given the choice, the air will take the low-resistance path, and most of it will go round the HE instead. I think I got some proof of that - see what's behind the ABC cooler:

Attachment 409992

The condenser is quite choked up behind the cooler; its pretty obvious there's not much airflow there. Although the cooler is only 16mm thick, its still enough to persuade the air to go around it. I obviously cleaned it up with a vacuum and a brush.

This is one of the reasons why I've been keen to avoid the popular deep, shallow, coolers that many people use. These are typically around 200mm high and 50-75mm thick. I think that amounts to a lot of air resistance, and although the cores have large volume, I expect they won't get much air flowing through them. I think this picture is some proof of what had only been a conjecture till now

Obviously, I'm now working on getting all this gear fitted to my car.

Nick

Nice hardware!

Let us know how they work out!

And here's how some of the hardware looks when installed, starting with the Mercedes W124 radiator:

Attachment 409993

This is a thick rad with a 40mm core, same as the S600 engine rad. Its a tight fit, but does have one great advantage - I can use the transmission oil cooler as the ABC cooler, and get rid of that horrible little stock cooler. That's essential actually, as there's not enough clearance between the rad and the Distronic radar for the ABC hoses.

The real problem, though, is that the top outlet sits so far forwards that there's no room for an elbow pipe when the hood's closed. I even cut the outlet short and tried to fit a 35mm plumbing elbow, but there's not quite enough space (and it would be difficult to remove afterwards). At best, it would mean cutting away the radiator grille and mounting it on spacers.

Attachment 409994

Here's a BMW X3 rad, with a 32mm core. Its has particularly slim end-tanks, and has the largest core frontal area that will fit - larger than the W220 condenser, and taller than the W220 rad, in fact. It needs several mods to fit - the condenser and radiator mounting brackets need to be trimmed due to the size of the rad, and I had to make some space for the inlet & outlet. This is taller than the W220 engine rad, and I was very pleased to fit it in with the hood and under-tray in position. I had to bend up the flange on the radiator top rail, while on the lower mounting brackets, I drilled new holes for the under-tray mounting clips - otherwise the screws would go into the X3 end tanks.

It seems that the X3 radiator is the largest one that will fit, but it does take some work. The BMW E46 3-series rad is all-but identical - but 50mm shorter. That would allow it to fit with few modifications - just for the inlet & outlet, so for most people I think that would be the ideal solution.

Nick

The other appealing option for me is the Mercedes W201 2.0 diesel/auto/air-con. It's almost identical to the W124 rad, and also has the integral oil cooler on the ABC side. Its a bit wider and a bit shorter, and has almost identical dimensions to the BMW E46 rad, but thicker. Even though its "only a C-class", its a big rad, larger than the X3, and similar to a mid-range 7-series or E-class.

Using the oil cooler for the ABC would mean cutting the metal pipe from the ABC thermostat output and fitting a compression hosetail (its only a low-pressure return hose). The connection from the cooler to the reservoir is just a flexible hose. That would be so neat, but I think I'm going to stick with the X3 rad.



Nick

OK, here’s an open question for everyone. I’m happy that I’ve arrived at a potentially ideal charge cooler configuration,with the BMW radiator and pump, and the Tecomotive controller. The controller measures the charge coolant temp using an in-line sensor, compares that with the target temperature, and modulates the pump speed accordingly. Most in-line temp sensor adapters are used on ENGINE coolant pipes, but I eventually found one for ¾” pipes. But the question is – where should it go? There are essentially four options, and each has a good reason to be used:

1) Charge cooler inlet – may give the best correlation with IAT
2) Charge cooler outlet – indicates how much work the cooling system has to do, and fastest response to increased load
3) Heat exchanger inlet – probably the worst case for high coolant temp
4) Heat exchanger outlet – shows whether the cooling system is actually doing its job

Although I really want to know what’s going on, I’m not inclined to fit a sensor at each position, and play tunes with them. In the name of science I might fit a cheap aftermarket in-line temp sensor with a digital dash display, and see how that squares with the controller’s display.

Conventional wisdom is that the thermostat and temp sensor should go at the hottest part of the system, but I’m not sure that will tell us what we really want to know. If the engine is working hard, we KNOW that the turbo outlet temperature is high. But I think we want to know whether the charge cooling system is doing its job properly, and that means keeping the coolant temp down at the charge cooler inlets.

One of the unknown quantities is what happens to the coolant temp as it passes alongside the exhaust manifolds. Maybe it actually heats up as it travels to and from the charge coolers? That’s why I think 3) is the worst case. My answer may be that the controllers job is to control the pump, so it needs to know when to speed up, and that depends on the HIGHEST temperature in the system.

What do you think?

Nick

Nick,
I disconnected the hose that returns the coolant from the intercooler on the drivers side and ran a dedicated 3/4" heater hose all the way to heat exchanger. I passed it right along the outer engine bay next to the condenser hose securing it with tie straps.
Why would anyone design a steel tube right over the exhaust manifolds and pass the coolant there superheating it further....?? I know .....production reasons. Not every Merc would have received this monster motor, so why plumb lines on the chassis that are not going to be used with every model. Makes sense...I guess. So they built it right on the engine itself.
How about this...when the coolant tries to reach ambient temperature in the heat exchanger it gets pumped right next to the passenger's side of the engine. In a steel pipe that is scorching hot bolted on the head and the exhaust side???!!! What were they thinking.
Point is, reroute all intercooler hoses away from the engine. We do not need the coolant heating up before it enters the intercoolers. What's the purpose of cooling it in the first place ... It defeats the purpose.

I'm just thinking out loud here. From what I understand, the function of the thermostat is to allow the engine to quickly get up to operating temperature and to maintain that temperature set by the thermostat for better fuel efficiency and lower emission which doesn't always mean the most power output and not to mention to have decent heat in the cabin for when it's cold out. But with the intercooler fluid temp, I don't think you would really have to follow that rule unless you'll be spending a lot of time below 0* F.

With that said, wouldn't you want the intercooler fluid to be cool as long as possible and to have it take forever to heat up? If the intercooler pump is wired to run continuously while the engine is running and also have an override switch for when waiting to run the next round at the track or just whenever, would that delay the fluid from getting hot especially with your new oversize heat exchanger? Isn't that the goal or am I missing something?

+1The car's ecu turns the pump on when the IAT reaches a certain temp. Just move more coolant (bigger he) and faster (better pump).

I think that makes a lot of sense (unless those pipes are there to somehow help protect the coilpacks from the heat of the exhaust).

Any chance of posting up some pictures of what you've done? I knew some people had already done this, so I've been thinking about doing it. I've bought lots of 3/4" tubing, but finding a way past the bulkhead seemed a bit tricky. The hood struts seem to take up the ideal space for pipes. Re-routing the pipes would also give me more flexibility to fit my new charge cooler header tank, and to locate the temp sensor in any position. There's no room next to the charge coolers themselves.

Nick

Continuous pump running is another quick win. I can understand the reasons for thermostatic pump control, so we'd have to accept that the pump is going to fail more frequently if it’s always on.

Having said that, the BMW pump is intended to run all the time, so it’s had to be engineered to be more reliable. It doesn’t have any rotating shaft seals, only static ones, and the commutator is electronic, not mechanical, so there are two problem areas avoided right away. TheTecomatic controller can run the Pierburg pump at slow speed continuously, regardless of the coolant temp. That would help to stop the pipes heating up when the IAT stays below the 116F threshold.

The more I think about pipe re-routing the more I like it. I wonder if it’s possible to route them in the wheel arches, behind the plastic liners? Its seems wasteful to take the 1½”connections on the BMW radiator and pump, and reduce them down to ¾” in order to use the stock pipes, so this sounds like the obvious opportunity to use 1”ID hose for many of the runs:

One of the things that I haven’t done is compare the volume of the stock HE and the big engine radiators. I’ll try that tonight.

Nick

Once you remove the fender liners everything will start to make more sense.

You can come up right next to the battery tray make a hole and enter the IC from the passenger's side. On the IC exit, make a hole and go down where the master cylinder is and route to the front HE.

The difference in temps is going to be worth this extra step. This is one of the best mods out there.

Nick,

just a little thought.

A few years ago I also changed the factory HE on my former pullied SLK32 AMG to the massive Needswings HE. It was excellent in terms of reducing IATs. It was very BIG and THIN and was the first cooler in row where the air moved through. I also had CM90 Pump, bigger reservoir, special fluid in the system, heat insulation etc.

Don't know exactly anymore how much, but it definetely also affected the engine cooling, because some flow into the engine cooling radiator was blocked from the big Heatexchanger in the front.

Keep an eye on the engine cooling temperatures under sustained hard driving. I would not say it was dramatic but the engine coolant went up a little bit faster and higher under hard driving on the Autobahn.

B

"Keep an eye on the engine cooling temperatures under sustained hard driving. I would not say it was dramatic but the engine coolant went up a little bit faster and higher under hard driving on the Autobahn."


The majority of us will never see the Autobahn with our cars, so top speed for extended periods of time is out of the question. We like to have fun on the street and blow the doors off anyone that comes to play. A few seconds of rush to put a smile on your face. Doing 200 mph is not realistic and could be fatal...not for us. I doubt engine temps will be a concern in this case.

Gee, that's a big HE! How big is it? Let me guess - 24" x 18" plus 2" wide end tanks? Is it this one?

http://www.needswings.com/NeedsWings...r--srt6he.aspx

I expect that would be a beautiful fit for the V12TT if only the end-tanks were slimmer. As it is they would block the engine air intakes, but having the inlet & outlet at the ends is a great advantage.

I'm quite mindful of the effect of an upgraded HE on engine cooling. I've read a lot of radiator specs lately, and one of the common patterns is that when air-con is fitted, the engine cooling radiator is often deeper or thicker, in order to make up for the loss of cooling. An HE would probably have the same effect - though remember that the 6 litre TT has a lot more power AND a full-height HE compared with the 5.5 TT, but the engine radiator is the same.

Having said that, these days its normal for ordinary turbocharged cars to have a full-size HE sandwiched between the AC condenser and the engine rad. A typical configuration is: 16mm AC, 32mm HE & 32mm Rad. The latter is usually the same size for N/A and turbo cars, so it could be that a large HE simply isn't heating-up the incoming air very much. Of course, it will ideally be working at a relatively low temperature, and won't put too much thermal load on the engine rad.

Engine cooling has to be an issue of some sort though. Since it ain't that hot in the UK, I'm confident that I won't get too many problems, and I'm definitely going ahead with my upgraded HE. But part of the trial is to show that it does work without impacting engine cooling or AC operation etc.

Well, that's what I'm hoping.

Nick

Yes Nick,

this was exactly this Needswings-Cooler. It was a quite tight fit. :)

If I remember correctly, in the 65 the Heatexchanger is also in the Front and in the 600 it is sandwiched between other coolers.

If I would have to reduce IAT-Temperatures with the M275 I would also keep an eye on insulating the Intercoolers themselves with aluminium foil etc., because of the near position to the hot engine.

I wish you good luck with your Project! :y

B

Having rationalized the HE/radiator options down to a coupe of units, I'm loathe to widen the net again, but I did come across these options that have some specific and worthwhile advantages:

http://www.ebay.co.uk/itm/3606717446...84.m1423.l2649

BMW rads are a good choice to avoid having the outlet foul the radiator grille when the hood is shut. The E46 rad is a safe bet and easy to fit, but the shear capacity isn't that great compared with the W124 or W201 rads. This is an uprated version with a 42mm core and 10.1 litres capacity.

http://www.ebay.co.uk/itm/3904777525...84.m1423.l2649

Neither does the E46 rad have an oil cooler. BMW have long used an oil/water heat exchanger to cool the transmission fluid on automatics. However, 20 years ago many of their 6-cylinder automatics used this radiator - its very similar to the E46 rad, but it has an integral oil cooler. That should allow me to get rid of the ABC oil cooler, which is very tempting.

http://www.ebay.co.uk/itm/4004709720...84.m1423.l2649

Attachment 409995

This is a particularly interesting purpose made HE. It has 19mm inlet & outlet, which are mounted on the side - like a proper HE - so easy to hook-up. Its similar to the stock cooler, but twice as thick. Best of all, its being sold directly by the manufacturer, not a distributor, and they're happy to do a customized version. I asked them if they could make it 20" tall instead of 12", and they said yes - for 25% more. I'm sure they can make a single-pass version, but I might ask if they can do the attachment brackets as well.

Some of their alloy "race" rads actually have integrated oil coolers. That's unusual for this sort of product, so it makes me wonder if they could integrate an ABC cooler into this as well?

Nick

The Winner Racing HE is looking more attractive.

Their stock HE has core measurements of 21.5" x 12" x 1 5/8" and overall 25" x 12.5" x 2"

Having measured up carefully tonight, I think this would be a perfect fit at 19" height (maybe 20")

I have a better feel for all the brackets, pipes and mountings now, and this probably wouldn't need any metalwork modifications - just a few cut-outs to the plastic mouldings.

Basically, Winner say they can make an HE to my drawing, so that includes single-pass, the pipe positions and even the brackets. Considering this would be full-height HE, it would make sense to add two top mountings to pick-up on the condenser mountings. That way the HE would be fully mounted to the radiator, just as the stock HE is, and would control the location and spacing properly.

Best of all, they say they can incorporate an oil cooler, so I can do away with the ABC cooler. Fantastic.

Anybody else interested in the biggest bolt-on HE that you can fit into an S, SL or CL?

Nick

I think I have a better handle on what will fit in the W215/220. All versions essentially use the same radiator and condenser, though some of the other coolers vary. ABC has quite an impact on space and installability - there are lots of pipes in the way, plus all the brackets that support them, but all charge-cooled cars have ABC. I spent some time at the weekend looking at how a large HE would be installed, what has to be removed, what has to be fitted before the HE, and what has to go after the HE. Its all very well having the space, but a big HE is no-go if you can't re-fit the radiator mounts or what.

It seems to me that the largest HE that can fit is 25" x 19" x 2". Simple as that. That includes the tanks, but not the inlet & outlet. If you're prepared to do some modifications, you can go 21" tall. And as long as the car is reasonably high, it can go in without having to remove much else. There's no need to disconnect the engine radiator, condenser, engine oil cooler or most of the brackets, and even the front wheels and the headlamps can probably stay put.

But you do have to remove the lower radiator brackets and therefore the bumper. There's not enough width between the brackets to fit anything from below. If the HE is a radiator, the engine rad has to move back, which means removing the fan and the tranny oil line. All the complications are entirely down to the inlet & outlet - what space is available, how to fit it, and how to connect it.

A purpose made HE is much easier, as the inlet & outlet are at the side. That makes it easier to install and connect, but its more expensive.

That's about it, really, and I wish I knew all that beforehand. I've been concentrating on the W215/220, but I did look at the W216/221 for another MBW member, and I'm afraid that's rather different. A custom W220 HE will not fit the W221, and unlike the W220, the W221 S65/CL65 has a different configuration all of its own. This post comes from first-hand experience with my own car, and I'd only be able to advise on the newer cars when I can get my hands one, maybe in a few year's time.....

Nick

I’m slowly getting there, but I’m still finding reasons why nobody has done this before.....

First, I read that the impeller housings on the Pierburg pumps can be removed and rotated so the outlet points upwards (as they always should, to help avoid air-locks). Well, it can’t. Not on the two different pumps that I have here. Part of the outlet port is formed by the main pump housing, not just the removable inlet, which only fits one way. So to get the outlet pointing up I had to turn the whole pump round so it faces backwards. That would make it difficult to integrate with a stock HE, but it’s no problem with my X3 radiator, as the hose takes a different route. Perhaps the smaller CWA-50 and 100 pumps can rotate their plastic housings, but the much more common, all-alloy, CWA-200 can’t.

Looking inside was interesting though. My impression of the Pierburg pumps is that they’re properly developed for automotive prime-time, unlike some circulation pumps. They don’t just achieve high pressure and flow figures, they meet all the other real-life requirements as well – they’re compact, light, reliable, efficient, flexible and have low power consumption. When you look inside you see how they do some of this. The castings and machining are beautifully done – rigid, light-weight, interlocking shapes with no excess material anywhere, and the O-ring mating surfaces shine like mirrors. Of course they’re more expensive than belt-driven pumps, but for our charge-cooling application, with the programmable controller, I think they’re a gift.

The Pierburg pump is too big to fit the stock pump mounting plate, though that’s no show-stopper. A Pierburg will need itsown custom mounting, and the easiest way is to remove the stock mounting. My car has done 156k miles, and the mountings in the radiator are seized, so it’s going to have to stay where it is (unless I replace the radiator).

With the Pierburg in its reversed orientation, the electrical connector is low-down at the front, and it fouls the radiator mounting bracket. I have to admit I never thought of that, but the two problems are solved with a new mounting plate sitting on the stock plate, that lifts the pump higher up in what is actually quite a large space. I'll post some pictures tonight.

Attachment 409996

Nick

Hi Nick,

during a Mercedes-G build during the recent 5 years we had made some modifications and testing to the planned setup as well. first of all it was all about repositioning the whole cooler-package to the rear of the car. due to the engine, m113k, we planned to install a complete cooler-package from the CL55, C215. during our first season, we drove this thing with the package located the rear wall of the cabin with a very deep and tight fit just in front of the race tank. all the plumbing was done by aluminium. the ic-pump was exchanged to the cwa-100 as used in 2000 by amg in its C30 CDI AMG and now again in the sls.

the thermostat was modified to remove the outer valve by ginding it out, leaving the inner valve in place to allow closing of the inner circuit.

for the regulation we developed a special controller, which controls the rpm of the cooler-fan. as it has 850W, it was necessary not to have it toggle between off an on, as this thing sucks about 75amps. it was possible to use the ecu for controlling the fan, because we made some modifications to the air-conditioning side, going away from the pwm-controlled ac-compressor.
the controller not only controls the fan, it has inputs for other temp-sensors displayed on a 20x4 vfd-display as well for switches to have the fan on hold for water-xings, but the controller would take care, if this switch would have been forgotten in race-stress, switching the fan on again at a specifig engine temp etc.. and it had inputs for the air-cond system, to ramp the fan to some rpms when needed by the a/c-system etc..

diameter of the water-lines was approx 50mm for the engine-cooling and 30mm for the ic-circuit to avoid a specific water-reservoir and to reduce "friction" in the circuit itself.

the filling of the system is done by a vacuum-system, allowing the ic-circuit to be filled up to the max, because we had a dedicated reservoir for it incorporated as well.

the system works, but it shows that it was unable to have the needed cooling-power for longer deep-sand sections because the fan was the only thing forcing air through the coolers and because of the bad positioning .

so, besides some more modifications to stretch the whole car, we decided to change whole setup.

the tank was moved to the back of the drivers cabin, allowing to cooler to sit more exposed and higher. and we went to a much larger radiator, from a mercedes truck, Econic. with this thing, we got approx 4 times the capacitiy from the original mercedes g-radiator. the connectors had 70mm ID opposite to the 34mm of the original cooler :-)

the routing of the cooler-lines was changed significantly. i will post some pics later on.

during all these modifications we had the idea to install CWA200 or even the more powerful Continental SCPII. but doing this we would have installed a new possible point of failure and had to remove the mechnical pump. so, decision was to see if the mechanical pump can do the job. the cwa100 was left in place for the ic-circuit, still controlled by the ecu.

all temp-regulation is done with our fan-controller. we had installed the "normal" thermostat again, with opening temp 86°, but we got in trouble with long full-power runs. switching back to our modified thermostat, all was ok.

during our first races we had absolute no temperature-problems any more.

some pics to follow.

best regards
carsten

ive attached some pics shot during the build.

some more pics.

Hi Carsten, that's an interesting account of an uncompromising installation for a presumably much more demanding application. What's a little surprising about road cars like the SL600 or CL55K is that the engine cooling rad is no bigger than the lesser models. I think the reason is that the engine only works hard for a short time, when the car is travelling fast and natural airflow is high. Rather unlike a truck or off-roader, in fact. Presumably a road car can get away with a modest IC HE for the same reasons.

So it sounds like you needed a very good charge cooler, and the obscured position wasn't suitable. Can I ask what monitoring you were able to perform in use? Intake air temperature? Charge cooler water temp? Charge cooler flow rate? Were you unable to have a radiator in the conventiional position - or did you have one there as well?

Regards, Nick

For the SL65 The CL600 condenser with Meziere Pump works OK...
 

I have an 06 SL65 (R230) and I first installed the Renntech designed front condenser system.

Four years ago I went to buy the system from Renntech at $4,200 and I noticed an MB sticker with a part number on their condenser. I thought I should check the MB part number before dropping $4,200 (uninstalled) on a small box of parts and I discovered that they were just using a standard MB heater core found in numerous MBs. Their system had that condenser with a few custom brackets and small rubber hoses. I ordered the condenser from my local MB Dealer for $168 and ordered the brackets and hoses from Renntech @ $200 thus saving almost $3,900 for the exact same parts (sorry Renntech). The MB/Renntech condenser is about 4 inches tall 22 inches long and 2 inches deep.

When I removed the front bumper I was surprised to see that on my SL65 from the factory was the exact same condenser and that Renntech was just adding a second condenser. That system worked better then stock but still heat soaked quickly. I already had the late upgraded Bosh 010 pump so I kept that.

About a year ago I installed the larger CL600 front condenser (pictured herein) that fits very tightly but required almost no mods, along with a BMW Meziere pump from Speedriven that they modify for the application. The system is extremely hard to bleed and takes hours of squeezing the hoses and letting the pump run but once bled does make a huge difference over the previous system. It still heat soaks but takes much longer and recovers much faster.

What I find is that while initially bleeding the system using clear plastic tubing I did notice that very tiny micro bubbles persisted. I have a feeling that either there are micro bubble attached inside that dislodge in the system OR more likely the impeller in the Meziere pump is cavitating and causing the micro bubbles. About every three months I have to re-bleed the system and do find large air bubbles that escape and allow 10-12OZ/300-500 ML more water into the system... I also still see those very tiny micro bubbles floating in the system even after hours of the pump running both with the water hot and ice cold.

Filling and Bleeding
 

Interesting stuff, and quite topical as it brings me onto the next post. I spent a lot of time filling, bleeding and testing, and because I did that at the same time, I learned a lot. Basically, the stock system sucks, it’s a nightmare. It’s critical to bleed all the air out of the system, but it’s quite impossible to do so – the pump doesn’t prime itself, there’s a big air-lock in the HE, the bleeding points have valves, and the system is filled from the top.

The water-antifreeze mixture is more viscous than water alone, and it bubbles easily. When any air gets into the pump, it turns the liquid into foam, chokes the pump and stops everything from circulating. The bubbles are very slow to clear. So slow in fact, that it takes less time to drain the system and modify it with a bottom filling port and re-fill it properly - than it does to bleed the system.

Because I (attempted) to measure the flow with several different combinations of pump and HE, I spent a lot of time emptying, filling and bleeding. After a few days of this I gave up and added a T-piece to the lowest point in the system at the HE inlet pipe, and connected a small header tank with a long hose. Filling the system this way takes minutes rather than hours. You have to take care not to overflow the stock filler, but having a header tank on a long hose makes everything very easy. The system can be slightly pressurized without having to switch on the pump, and you can bleed the IC inlet valves without generating more bubbles than you’re removing.

If I’ve learned anything from this protracted episode, is that you must remove ALL the air from the IC system, and you shouldn’t even THINK about trying to fill and bleed it without adding a bottom fill port.

Nick

gawd... so much reading, can anyone summarize?

.

Think what it must have been like to write

For anyone wondering what happened to my charge cooler project, its all installed and I've been driving around a lot over the last few weeks, but I'm struggling to find time to write it all up. I've got lots of procedures and pictures, and I'll post it up next weekend. I'll cut to the chase though - here are the quick answers:

Does it work? YES
Are there any downsides? NO.

See ya soon, Nick

Here's a concise procedure for fitting the BMW radiator and pump:

Drain the IC system and remove the pump and HE.
Buy a new BMW X3 radiator and cut off all the brackets and ribs.
Shorten the inlet & outlet to leave a single swage, and trim swage back to 1mm high.
Cut back the HE top stiffener to clear the safety catch.
Cut slots into the HE top stiffener to mount the ABC cooler (like AC condenser).
Flatten the ABC cooler bracket and bolt through the HE.
Bolt the ABC cooler outlet pipe bracket through the HE.
Cut and bend the top rail to clear the HE header tanks and ABC cooler
Lower the ABC cooler outlet pipe bracket by ½” and route through the headlight bracket.
Move the safety catch ½” forwards so the tang clears the HE.
Move the ABC thermostat, pipes and brackets ½” to the left (tricky but critical).
Bend or shim the engine oil cooler ½” forwards to clear the HE.
Move the forward fasteners on the forward under-tray back ½”.
Move the RHS horn up ½” to clear the larger pump.
Cut away the LHS radiator bracket to clear the HE inlet.
I cut the RHS radiator bracket back, but this probably isn’t necessary.
Cut away the RHS headlight bracket to clear the HE outlet (though not this much).
The Pierburg CWA 200 pump faces backwards so the outlet faces upwards.
Cut and drill a new pump mounting plate to go directly on the rad.
The CWA 200 pump needs its own, larger mounting plate.
The CWA 50 & 100 pumps can probably use the stock mount & orientation.
Raise the Pierburg pump so the electrical connector clears the radiator bracket.
Connect the pump and HE to the stock plumbing using appropriate reducing elbows.
Trim the radiator intake ducting to clear the additional and relocated plumbing.
Fit IC thermostat to RHS IC outlet pipe.
Wire pump controller to battery, pump, ignition and temp sensor.

I'm going to post lots of pictures of what I did, but rather than spending a lot of time on photo hosting, I'm going to upload directly, one picture per post, so I can describe it OK.

Here's the first. In order to mount the ABC cooler and pipe brackets to the new HE, I made some new brackets on the reverse side, to which I glued some bolts that go through holes pushed through the HE matrix.

I cut a section out of the right hand side headlight mounting bracket to clear the HE water outlet. Sorry, I can't give any dimensions - you'd just have to go by this picture, though I did cut out about 1cm more than I needed to on each side.

I moved the elaborate bracket the supports the ABC thermostat and various pipe backets to the left (as seen by a seated driver). I did this by drilling new holes in the bracket mountings. This is one of the key operations, and has to be done at the point of maximum disassembly. The bracket is VERY fiddly to remove, but it can done if you loosen everything else. The bracket mounts to the impact bar and to the left hand radiator support bracket. The latter is removed in this picture for clarity. When fitted, its damn near invisible. In hindsight it would have been better if I had moved the bracket ONE INCH to the left.

I mounted the Pierburg water pump to a metal bracket, which was in turn attached to the stock bracket. If you can remove the stock bracket (mine was corroded) do that, and bolt the new bracket directly to the radiator instead. In my case I glued some bolts to the stock bracket and to the new bracket, so that I could bolt each part to a set of studs. This makes assy much easier.

Filling and bleeding the IC system is both critical and difficult. Part of my solution (which I recommend to EVERY SINGLE PERSON with a V12TT engine) is to connect a four foot long narrow rubber tube to a T-piece at the bottom point of the system at the HE inlet. This was then run upwards past the ABC valves and left hand headlight, and alongside the AC pipes towards the suspension strut top and the soft bulkhead. This is capped and left fitted permanently, but is pulled out and topped with a funnel or small header tank to fill the system. Seen here in situ, with the headlight removed, it runs kind of diagonally, from top left to bottom right.

The HE and pump are connected to 3/4" pipes via 35mm to 28mm 90 degree reducing elbows (typically made from multi-layer silicone rubber). I ran a tube from the HE outlet under the Left hand head light to the inlet of the pump, at the rear. The outlet is connected with another new length of 3/4 tubing through a new route that goes behind the headlight frame and under the engine air inlet. It clears the radiator fan OK, but gets in the way when removing tha fan, so it would have been a good idea to use the stock pipe instead, even though it has a tight turn in it.

I fitted a temperature sensor hose adapter to the coolant outlet on the right hand side charge cooler. These are difficult to find - they need to fit the Tecomotive temp sensor thread AND fit 19mm pipes. Make sure you have this in your hands before you start any work (if you use the Pierburg / Tecomotive solution like me). I would rather have fitted this downstream of the Y-junction that joins the outlets of BOTH charge coolers, but the only space is in front of the engine. That means the temp sensor response would be slow, due to the length of plumbing between the IC and the sensor. This way the response time is quick.

This was the real money shot for me. Can you tell what this is? I was concerned all along that I simply wouldn't be able to close the hood without the safety catch hitting the larger HE. I modified the safety catch by moving it 1/2" forwards, by drilling new holes for the mounting studs. I also had to put the catch in a vice and bend the catch itself forwards, and away from the HE. The latch works like stock, and it doesn't get too near the HE matrix. Hurrah!

This picture is upright and taken from the left hand side, with the headlights removed but with the hood closed and latched. The HE is on the right and the radiator grille is on the left. It shows the latch - the shaddowy thing in the centre, behind the Distronic radar and bracket. This clearance is fine, even if the hood is slammed down.

The only slight downside of the whole installation is that the small plastic T-handle that sticks out to release the safety catch is pushed forwards by the HE. The catch hinge touches the HE matrix, and needs some protection, and the handle itself doesn't retract all the way behind the grille - its just flush. You could shorten the handle easily enough, but it looks OK. Very happy about this.

The biggest task for me was to find the right radiator, and you wouldn't believe how many I looked at. I'm happy that the 580x499mm BMW X3 radiator is the right one - itys the biggest one I could fit, and its tight. Here are some pictures to illustrate the perfect fit.

What the heck, they have to be shrunk and uploaded sometime - lets host the pictures like before.

The suspension oil cooler (used on all BiTurbo cars) is now bolted to the HE instead of the AC condenser, and is a close fit everywhere. The mountings are essentially the same as stock, but the mountings are reduced height and the cooler is just a few mm away from the HE matrix:

Attachment 409997

The cooler inlet/outlet pipe crimps are very close to the HE matrix - I can just get a piece of paper in between:

Attachment 409998

And the associated hoses are also very close but not touching:

Attachment 409999

This is a sideways view of the top of the cooler and HE. I bent the top mounting brackets slightly, and slotted them into holes that I drilled and sawed into the top of the HE radiator:

Attachment 410000

And this is the clearance between the cooler connector and the back of the Distronic radar. This is one of the reasons that I stuck to a 32mm core radiator for the new HE.

Attachment 410001

This is a poor picture of the HE outlet, with the modified headlight bracket in place. This is the downside of the BMW radiator, but the common alternative is to have the inlet or outlet right at the top, and then it fouls the radiator grille when the hood is closed.

Attachment 410002

This is a general view of the bottom left-hand corner by the HE inlet and engine oil cooler. There are lots of potential fouls down there, but you can see that the oil cooler bracket clears the HE header tank.

Attachment 410003

Same area, but a closer view of the right-angle reducing elbow on the HE inlet, showing how it clears the modified LHS radiator mounting bracket. That, and the RHS headlight bracket, are the major modifications.

Attachment 410004

This is the bottom of the HE from underneath, showing the clearance with the engine oil cooler ducting. That rubber flange goes under the bottom of the HE, and mates up to the forward under-tray, showing how the HE runs full depth in the space available.

Attachment 410005

This is the RHS engine cooler bracket, seen from underneath, showing the clearance with the HE matrix.

Attachment 410006

Finally, this shows that the relocated forward fixture for the forward under-tray clears all the heat exchangers. This is seen from underneath, looking sideways and upwards, with the under-tray removed but everything else in position. The bottom of the HE is top left, the AC condenser top centre, and the engine rad top right. The angled bottom of the RHS rad bracket is at the bottom - its pupose is to support the forward under-tray, but I glued and screwed some 6mm (1/4inch) rubber pads in place to support the new HE. You can just see the old fixing hole slightly to the left of the spring clip. The two mating holes on the under-tray itself obviously have to be re-drilled about 1/2 iunch further back. The under tray screws won't pierce any of the HE's no matter how hard they're screwed in.

Attachment 410007

That's pretty much how I did it. Thanks for your patience.

Nick

Hi Nick, i am monitoring all Parameters available on the CAN-Bus of the Car, besides all Conditions of the ECU, ABS etc.. there are IAT and Coolant-Temps displayed on a self-built 4,5"-Touchscreen. For Fail-Safe Reasons there are separate Instruments showing Coolant Temp/Intercooler-Temps for Backup.

ALL Coolers except for Steering-Fluid are replaced to the rear Position as shown on the pictures. Due to mud, debris etc. the effectiveness of the front-cooler is quickly reduced during a muddy race. During a muddy race you can observe a rise in Oil-Temps caused by mud attaching to the oil-pan alone...

BTW, we are filling our cooler-system by use of a vacuum-filling-system. this one is set on the expansion-tank. by means of pressurized air all remaining air in the cooler-system is removed through a venturi-effect. after that, the coolant is sucked into the system, filling every edge of the system.

i saw the pictures of your 'tight' installation. great work. but i saw some sharp edges and really tight fits to the cooler-core. think of expansion due to temperature and vibrations while driving the car. please check this to be shure that there will be no contact/leakage.

best regards
carsten

Thanks Carsten. When I drive my car around, I look at the 4-dial instrument cluster, wondering what what the engine is doing. Couldn't they at least have included a boost guage? Mercedes probably considered that a bit too boy-racer for a luxury car, reasoning that their buyers wouldn't be interested. But by that logic, why fit a rev-counter? Speed and load are the two key attributes of engine, and I feel like I'm driving around half-blind. I'd love IAT and I think Shardul has an after-market multi-function guage to give some of those OBD2 outputs. My pump controller has a switchable coolant temp/pump speed display, which is useful.

Regarding the installation, I was very mindful that manufacturers always use compliant mountings for radiators. I don't need to know the reason why - I just know that they do it because they need to, so I was careful to use compliant mountings and give the HE some breathing room. Some of the ABC pipes are very close, but they do move with the HE.

One of my biggest concerns was AC and engine cooling, but they seem hardly affected, even in the hottest weather. Maybe the electric radiator fan is working a bit harder than it used to in traffic, but then I never paid it a moment's attention beforehand. I'm just chuffed that I managed to get two engine radiators in there without making it look too obviously modified. With some cars the mod is obvious from outside. In hindsight, I suppose would have mounted the HE 5mm lower and 10mm to the left, but that's all I can think of. I'd like to find a way to get a W201 radiator to fit, but that's for another car....

Looking back at my first post, most modern cars with 2.0l turbo engines (either fuel) typically have 600x400x30mm intercoolers in front of the engine rads. Mine is 500mm tall, and being an HE it has the advantage of narrow water channels, so there are nearly twice as many rows and the incoming air flow through the rad is less restricted. Its a relatively cheap, light, HE as well. It could almost be considered as a sacrificial protector for the condenser and rad.

Nick

So...My next move is to schedule a weeks worth of vacation time...that should give me enough time to finish reading this thread!

Lots of info here, so What is the outside temp and what are the iats after driving around for 30min? Are the iats still close to ambient? Heat soak problem addressed and done with?

This project has taken up so much of my time this year that I've neglected many other things. I've finally replaced a few ABC hoses and a suspension strut; my wife wants me to re-tile the conservatory, my son started secondary school yesterday; I'm trying to get a very ambitious new PC working; my neighbour wants me to finish cutting down a huge hedge; I'm trying to put up a new home cinema projection screen; I've just become a scout leader; I pioneered linear power supplies for Meridian AV surround controllers and have two more units to modify, and I've picked up a burdensome new programme at work, so life's a bit busy to do EVERYTHING I want to.

I expected everyone to ask about IAT's, but just getting the new pump & radiator up and running at all has been an enormous effort and achievement, and I haven't even begun to think about quantifying the improvement. I've been so focussed on avoiding the downsides that I haven't really considered the upside. The HE is much bigger than stock; the pump is much more powerful; all the intake air HAS to go through the HE rather than round it; the pump control algorithm is programmable and much better than before; and I've developed several new methods for filling and bleeding the IC system, which I've found to be absolutely critical for the system to perform properly. That's what I'm going to talk about next.

Nick

Edit - sorry, didn't mean to be so defensive

nick, i believe the saying goes 5% of the people do 95% of the work. you my friend are in that 5%.

I thought everyone following this thread would enjoy this test that Lingenfelter did on interccooler pump flow tests. They have a bosch pump from a cobra in the test. They even include as installed flow rates. There are some interesting results for the meziere pumps. None of the MB pumps are included but it is a good reference.

http://ls1tech.com/forums/forced-ind...g-results.html

nice find

Thanks, that's a gold-mine of information, and very relevent to MB intercoolers. These tests are reported in various Cadallic and Camero forums, where there's generally a lot of interest in IC's. Everyone assumes the 20 gpm & 55 gpm pumps are Meziere, plus the Bosch pump part no is given as 0392-022-002, so the test results are very informative.

Firstly, the pumps are characterised for flow and pressure, not just flow. Some manufacturers, like Bosch & Johnson do give this information, but others don't. Meziere are the glaring exception, and it seems there's a good reason. I've pointed out in this thread that I don't have any pressure/flow information for the WP136S, and the characteristics on my chart are a guess based on open-pipe flow, and electrical power consumption.

My understanding is that Meziere make ENGINE cooling pumps, and the pressure/flow characteristics are optimized for engine cooling systems, which have much lower resistance than charge coolers. The pump tests show this is borne out. Everyone thinks of the Meziere (WP136) as a 20gpm pump, but put a load on it, and it flows LESS than the Bosch (just like the Johnson CM30). I think this should answer the question once and for all, should the stock pump should be upgraded to a Johnson or Meziere? The answer is no.

Secondly, the tests also give installed performance measurements, which show how much flow resistance there is in a presumably comparable IC system. The answer is - a lot, and it suggests that the pump's static pressure (closed outlet) is the better guide to installed performance than the open pipe flow.

Thirdly, there are a few new candidate pumps for our cars. I'm not sure I want to consider the over-rated EMP pump, which looks like a monster, but the VariMax pump looks like a great candidate. The performance looks rather similar to the Pierburg CWA-50.

I'll try to combine some of the new information onto my existing chart.

Nick

The other thing I wanted to capture is a set of links to a blog by Bruce Nunnally. He’s a fanatical Cadillac STS-V owner, and the parallels between this thread and his blog are remarkable. We’re coming from the same place, but he’s rather more professional about this. He’s been performing both analytical and empirical pump evaluation, and has some interesting installed performance measurements. He’s also gone a long way to closing the loop on predicting how pump performance on the bench equates to improved car performance.

http://caddyinfo.com/wordpress/cadillac-sts-v-intercooler-pump-bucket-test/
http://caddyinfo.com/wordpress/how-cool-is-this-intercooler-pressure-versus-pump-output-and-flow-motorama/
http://caddyinfo.com/wordpress/adding-boost-pressure-to-the-lc3-4-4l-dohc-vvt-v8/
http://caddyinfo.com/wordpress/intercooler-cooling-corvette-zr-1-ls9-cadillac-cts-v-lsa-sts-v-lc3/
http://caddyinfo.com/wordpress/meziere-55-gpm-intercooler-pump/
http://caddyinfo.com/wordpress/intercooler-pump-test-4/
http://caddyinfo.com/wordpress/cadillac-sts-v-intercooler-flow-gpm-bucket-test-3/
http://caddyinfo.com/wordpress/gonna-pump-you-up-intercooler-pumps-in-series-to-maximize-cooling-motorama/
http://caddyinfo.com/wordpress/taking-cooling-cues-from-the-zl1/
http://caddyinfo.com/wordpress/datalog-analysis-of-cadillac-sts-v-ambient-vs-iat1-vs-iat2/

What’s particularly interesting are the Laminova chargecooler characteristics for the STS-V. These show the charge cooler’s heat transfer rate vs. coolant flow rate, and he can predict the benefit of upgrading the pump. This answers another question that’s been unanswered for a long time – does increasing the coolant flow necessarily improve the charge-cooling? The answer is yes, but with diminishing returns. More flow = more cooling, even up to high flow and high pressure. I’ve been saying this all year, and this is good evidence.

The other gratifying point is the typical operating conditions of the STS-V charge cooler. The nominal condition has 30 degC coolant circulating at 20 lpm and about 50kPa pressure. The intake air is 110 degC,and the air flow rate is 350 g/sec. The heat transfer rate is 17 to 22 kW, depending on coolant flow rate (corresponding to a 25-fold increase in pumping power!). Those with good memories will recall that these figures are very similar to the pure first principles analysis I did towards the beginning of this thread. I conservatively predicted that the V12TT charge cooler passed 400g/sec and needed to dissipate 20kW, rising to 30kW for a tuned engine. So I’m feeling very pleased about that.

Nick

Dupe.

It was a while ago now, but like Lingenfelter and Nunnalley I spent a lot of time measuring pump performance. But it wasn't very productive. My intention was to test all configurations of pump and heat exchanger, working as a system.

Firstly I baselined the stock pump & HE, so see where I was starting from. Then I substituted the Pierburg pump, and tested that with the stock HE, to see how much improvement I could get from the big pump in isolation. Then I hooked-up the BMW radiator, and finally went back to the Bosch pump, specifically to see whether the stock HE was restricting flow. Knowing that the Pierburg was an engine pump, I suspected that the full benefit might only come when used in a low-restriction system, but I wanted to see for sure.

Rather than using a bucket and stopwatch, or a fancy ultrasonic flowmeter, I plumbed-in a simple 3/4" impeller water meter to the HE outlet, and measured the flow round the closed-circuit. Of course, for each configuration I tested, I had to drain, fill and bleed the system each time. Have you ever heard people say how difficult it is to bleed the V12TT system? Well its true, its an absolute nightmare. The water-antifreeze mix froths badly, and if you try to circulate it using the, it turns the whole system into a milkshake, and the flow gridlocks. In hindsight, I would have done the testing with plain water, but its too late now. I spent several whole days doing this, and really made very little useful progress.

Attachment 410008

The engine wasn't ready to be run, so I used a battery charged up to about 13.0V. Even that lost some voltage during the testing, and the flow figures suffered, so I hooked-up an 8A charger during the testing to keep the voltage up. The test above is the stock HE feeding into the water meter, and then into the stock pump, in front of the wheel.

Attachment 410009

Above, here's the X3 radiator outlet feeding the meter with appropriate pipes. The Pierburg pump is controlled by the tiny Tecomotive pump controller. Both the Bosch and Pierburg pumps ran pretty quietly, and apart from the meter, it was quite easy to avoid leaks - hardly even any need for the jubilee clips.

Attachment 410010

I eventually rationalized the whole test into two tests - all stock and all modified. I wasn't even able to do the first until I'd developed a completely new way to fill and bleed the system. You just can't can't get any meaningful results until all the air is out of the system.

And rather like the Lingfelter and Nunnalley tests, the stock results weren't as good as you might have expected. The best I could do was about 14 litres/min, or 3.7 US gpm.

With the BMW pump and radiator, I got about 50% more flow: 21 lpm/5.5 gpm. I was disappointed with that, but in hindsight its a good result.

It shows that an IC system really does need an IC pump, and not an engine cooling pump. The Pierburg CWA-100 would probably be more suitable than the CWA-200, but the VariMax sounds like a good option. On this occasion, I'll just have to live without knowing how restrictive the stock HE is, and how much more flow I got just by fitting the X3 rad.

Nick

I really appreciate your efforts. Keep it up, they benefit all of us.:y

Looking at the Lingenfelter and CaddyInfo threads on Mercedes, Cadillac & Chevy, I think there's a very clear conclusion to be had, and a new direction to look for suitable IC pumps.

IC systems are quite unlike engine cooling systems (or buckets of water for that matter). They're all HIGH RESISTANCE systems, so they need a lot of pressure to achieve good flow. Many pumps are rated for open flow, with no load, but that's almost totally useless for chosing an IC pump. If you want to use a single figure, the static pressure (maximum head) gives a better indication of improved performance, but pumps are more complicated than that, and you really need proper flow vs pressure curves, which is where the Lingenfelter tests are so useful.

There's a lot of talk about 20 gpm pumps, but 3-4 gpm is quite normal in the real world, and 5 gpm is a good figure. Its actually quite difficult to get even that high - I spent several days trying - and a high antifreeze concentration, an uncharged battery, or a tiny bit of air in the system, really knocks it back.

What this means is that pumps that are suitable for engine coolling, or that do well with the bucket test, aren't really suitable for IC systems. So that includes pretty much all Johnson, Meziere & Davies-Craig pumps. So that means that all the forum favourites, like the EWP-80, WP136, CM30 or even CM90, are no improvement on the Bosch. Many people swear by the Jabsco 50840, but even that's a high-flow/low pressure pump. In my searches, I found a couple of other likely 20gpm+ candidates - the Dayton 5PXX0 and the SURflow COMSV012D but they give much the same performance as the above:

http://www.drillspot.com/products/1380036/Dayton_5PXX0_Pump
http://www.pumpvendor.com/media/shurflo_industrial/Shurflo_Industrial_COMB_COMS_DC_series.pdf

IC systems normally run at around 0.5 bar, and if you want more flow, you need a lot more pressure. So you need to look for pumps that do 1 bar or more, and I found a couple of real candidates. They have 15-20 gpm open flow, but they achieve up to 1.1-1.2 bar, and the installed performance will be better than ALL the above.

The first is an identical twin big brother to the Jabsco 50840, called the 50860. It looks and cost about the same, has less open flow, but much more power, pressure and installed flow.

The other I never heard of before, but is rather expensive - the FLOJET DC 40/10.

http://www.xylemflowcontrol.com/marine-and-rv/general-purpose-pumps/circulation-pumps/50840-series-low-pressure-cyclone-centrifugal-pump-copy-item14187.htm
http://www.xylemflowcontrol.com/files/50860_43000-0858.pdf
http://www.aquaintl.com/product_details.php?category_id=786&item_id=3339
http://www.xylemflowcontrol.com/files/DC15_30_40_SS_950-0676.pdf

They're both monster pumps, bigger even than the Renntech pump, and I would guess similar to the stock Stewart-EMP E2512A that Lingenfelter promote so effectively in the top thread.

Nick

Just a quick thought. I'm probably worrying about nothing (and it won't affect some people at all) but one of the quick wins for charge cooling is to have the IC pump running continuously.

But what happens in freezing weather? The coolant will circulate while the ignition is turned on, and after a cold start will presumably cool the IC to below zero.

Isn't there a risk of the IC freezing up?

Nick

The flowjet pump is very interesting. I am doing a little more digging for specs. The Jabsco pumps are all only rated for thirty minutes on and thirty minutes off for only a 50% duty cycle. For a continuously running set up you might see a very early failure. They list a warning that the Jabsco pumps will get very hot. Any guess the rating on the pressure cap on the intercoolers system?

As for cold weather operation, as long as your antifreeze percentage is high enough to keep ice from forming you are in good shape. With the intercoolers sitting in the engine bay they will warm the system quickly. At best our IATs are 20 degrees above OAT. If you are using water wetter or another lubricant they will not protect against freezing.

I believe the Pierberg 100 Series pump is flow rated at .80 bar which is right in line with our operating range.

I think once the engine is running the turbos will warm the intake air enough to avoid freezing, but I was thinking of the scenario where the ignition is turned on some time before the engine is started. The IC will be freezing cold, and the intake air will be cold for a short while before the turbos warm up. The fins are close together, and I don't think it would be difficult for ice to block the whole intake, and then the engine wouldn't stand a chance of running. Anyone would be at a loss to diagnose that problem, which would mysteriously disappear.

On the upside, freezing cold air is dry, because the moisture has condensed and frozen out of it, so the problem might not be that bad. However, throttle bodies are often heated by engine water to avoid freezing. I admit that throttle bodies are a different case, as the air cools when it expands suddenly when passing a part-closed throttle, so you can get freezing of moisture-laden air. IC's are rather different, but this is mostly just me trying to cover things that might go wrong.

Yes, the Jabsco 50860 is rated for intermittent operation, and I think the FLOJET 40/10 is as well. However, the low pressure 50840 is continuously rated, as is the 24V version of the 50860, for some reason. As long as those big pumps are run under normal IAT thermostatic control, I think they'd be OK.

What we need for IC pumping is away from the mainstream for pump manufcturers, falling between the low pressure centrifugal pumps, and the high pressure diaphram pumps. Something like 1bar @ 30lpm will do, and the Pierburg CWA-100 is very close (and can be run from a Tecomotive controller, AND the motor is water-cooled). I was looking for another high flow/high pressure pump that's continuously rated, and I did come up with this promising option today - the Stuart Turner 12/50 series:

http://www.stuart-turner.co.uk/produ...250-/1250-12-v
http://www.stuart-turner.co.uk/media..._Datasheet.pdf
http://www.stuart-turner.co.uk/media...%20Section.pdf
http://www.stuart-turner.co.uk/media...ow-Voltage.pdf
http://www.stuart-turner.co.uk/media...ance-12-50.pdf

Of course the other option is to use two Bosch pumps - leave the stock one in situ and simply add another one at the HE inlet.

Nick

Interesting article on Pierburg Electric Water Pumps here:

http://www.omiq.it/resources/Documen..._64_5_6_LR.pdf

Attachment 410011

Jabsco Cyclone 50860 flow-chart:

Attachment 410012

Note that the 50870 is obsolete and has been superceded by the apparently similar 50860.

I think this is narrowly the best of all the stock charge cooler pumps - pumps 30 lpm/8 gpm at 1 bar.

Nick

Still questions
 

Ok so ive read the whole thread and still have a few questions. Before reading this thread I was gonna just upgrade the factory HE with a larger universal one of some sort along with a pump upgrade, like people have done in the past. But the pumps used don't seem to be like a good choice anymore.

First question is if I was to go the radiator route with a cap and overflow nipple does the nipple have to be sealed shut? does the cap need a certain psi rating? along with that (and this may be a stupid question) but if the all aluminum radiator had the oil cooler option could you route your ac lines to run through there to have like a slight killer chiller effect? not sure how the oil cooler portion of radiators are routed....

Second, is that it seems that the OP has proven that the pumps we usually upgrade to are no better than the factory so do we install a second factory (same specs) to run with the other inline? wired to kick on together
Or do we install a second one -<>- split to where one could run all the time and not affect the other and then the factory kick on as normal?

My plans are to upgrade these The HE and the IC Pump once I figure the best setup for me. Then add a snow performance boost activated meth injection kit solely for AIT's I wont be tuning for the injection just for the temps.

sorry for grammar, punctuation etc....lol its late and I don't care.

If you go for a radiator, the whole system still has to be sealed, but not for quite the same reasons as the engine system (which needs a high boiling point - the IC doesn't). Rad & header caps are usually set about 1 bar, which should be fine.

I guess you could try running AC refrigerant through the oil cooler, but I think the latter isn't designed for high pressure, so there'd be a risk of rupture. Even the low side of an AC system has a few bar.

You're right, most of the popular "upgrade" pumps are clearly unsuitable for IC systems, but adding a second stock pump with the first is a good idea, especiallly if you wire it to the ignition. Its got to be in series with the existing pump though, so the pressures add, not the flows. I'd suggest fitting it in the front left-hand corner, near the HE inlet, but remember the pump outlet must always face upwards, to avoid air-locks.

Nick

I've been learning a bit more about centrifugal pumps, how to match them to the system, and how to upgrade them. There are some great resources out there, and some common-sense discussion of what happens when you double-up pumps for more flow:

http://www.engineeringtoolbox.com/pu...ial-d_636.html
http://www.mcnallyinstitute.com/15-html/15-01.htm
http://www.pumpindustry.com.au/under...parallel/2138/
http://wycopump.com/pdfs/pumpclinic35.pdf

Basically you can double pumps either in series or in parallel, and both techniques will increase system flow. One will be more successful than the other, though, and it depends on the characteristics of the pump and the system - which is flat and which is steep.

The simple effect of multiple pumps is:
Pumps in series add the pressure at the same flow rate
Pumps in parallel double the flow at the same pressure

There's a recurring theme across all those websites, and that's how important it is to match the pump to the system. Each pump has a preferred operating point on its pressure/flow curve. Its the condition that its designed (or tuned) to operate at, and its where you get the maximum efficiency and the hydraulic power. Hydraulic power is the product of pressure and flow. If you multiply litres / sec by kPa you get Watts (and in practice its almost always less than half the electrical input power).

The actual operating condition of a pump is where the pump curve (which slopes down) crosses the system curve (which slopes upwards). Both curves are shallow at low rates of flow, and get steeper at high flows. If the pump isn't matched to the system, it will operate away from its optimum operating range, and you won't get the best from it.

More than that though, it will also cause the pump to break down. If a pump has too little resistance (shallow system curve), it will over-speed and may cavitate, eroding the impeller. If the resistance is too high (steep system curve), all sorts of things go wrong. The flow and pressure around the impeller will not be uniform, and will be subject to noise, vibration, radial load and axial load. All of which will wear out the seals and bearings, and may cause the pump to overheat.

That's effectively what happens when you take a high-flow engine cooling pump and fit it into a typical high-resistance charge-cooling system. You get low flow and an unreliable pump. There's a lot of good science to this, and it all makes perfect sense.

Nick

I need to correct myself for the last time. I've just bought an S65 heat exchanger for my S600, and the AC condenser mounts onto that, not the other way round. Therefore the S65, like all other cars with full-height heat exchangers, has the HE sandwiched between the condenser and the radiator.

Incidentally, this means that the S65 and CL65 use a different AC condenser to all the other models.

It opens an interesting option, though. My condenser has a steering cooler in the top, and I'm not planning to replace that. The S65 heat exchanger ALSO has a steering cooler at the top, which then becomes redundant. Can I think of anything to do with it? I certainly can! How about using it to cool the suspension oil, instead of that horrible little square cooler?

I like that idea.

Nick

Moved down

http://www.hehlhans.de/motorg55-7.htm

Info: there's an illustrated procedure here to strip down the 0392022003 pump. It uses a conventional DC motor with brushes, so it will never last for very long. I think that must explain why the motor doesn't run all the time. The Pierburg pump uses an electronic commutator, so it has no brushes, and is presumably more reliable when running continuously. Many of the other industrial pumps I listed recently are designed to be servicable, so you're positively encouraged by buy service kits for brushes, seals, bearings, etc. They're designed to be maintainable.

Lots of good stuff on ABC as well: http://www.hehlhans.de/sl55amg-abc.htm

Nick

I intended to copy some of the charts from those websites as they illustrated what pumps do pretty well. This is what happens when you use two pumps in parallel with a high restriction system (steep system curve). The combined flow rate does go up, but not very much. You're better off putting the pumps in series, so the pressures add (same as using a higher pressure pump):

Attachment 410013

Now this is what happens when you have low restriction system (flat system curve). Pressure isn't the problem, and you get much more flow. This is more effective than series pumps, as the double pressure pump curve would cross the system curve at a lower point than the double-flow curve:

Attachment 410014

PC gamers take water cooling seriously, and they can be very analytical about it. Here's a good chart that makes it easy to see.

Black line: system curve (resistance)
Thin blue line: pressure vs flow for single pump
Green line: two pumps in parallel (double flow)
Pink line: two pumps in series (double pressure)

The operating point is where the system curve crosses the appropriate pump curve. In the example given, which has a steep system curve (i.e. high resistance) the pink line crosses at the highest flow. That's because it crosses the pump curves below 3gpm, in the region where the series configuration performs better than the parallel. Look at the point where the green and pink curves cross - 3gpm/6psi. This is where they perform the same. If the system curve was shallow (low resistance), and fell to the RIGHT of this point, then the parallel configuration would give more flow.

Attachment 410015

There's lots of discussion on this subject everywhere you look, but less on what you can do with the other side of the equation - the resistance of the system. Just as pump flow adds in parallel, and pressure adds in series, the same applies to the resistance of the loads in the cooling system. Two heat exchangers in series will have double resistance/low flow, while two HE's in parallel will have half resistance/high flow. Unfortunately half resistance doesn't mean double flow - you have to plot and compare the pump & system curves to get the answer, just as you do with pumps.

It was the discussion on series and parallel pumps that got me thinking about series and parallel heat exchangers. In the case of the V12TT, the two charge coolers are already in parallel, minimising the resistance of the system, even though its obviously still pretty high. The IC's are in series with the HE, so the resistances add, and flow though both is low. My X3 radiator makes a big HE, and its resistance must be very low - it has to flow enough water to cool an engine. But the flow through the HE is obviously being limited by the restriction of the IC's, so I'm not getting as much cooling from the HE as I ought to.

Connecting the IC's and the HE in parallel with the pump means that the IC's aren't in the loop with the pump and HE - so with suitably large pipes, the flow in that loop should go through the roof. Of course this only works with a powerful pump that maintains good pressure at those higher flow rates, and the CWA-200 certainly does that. I'm not sure if that arguement would also apply to the Meziere or Jabsco pumps - other popular but relatively shallow-slope pumps. Might be worth thinking about.

Incidentally, I've been playing with the back seats in my S600 Limo this weekend. I use the word " Limo" advisedly, as the back seat is a bit disappointing for such a large car. The backrest is tall and well-angled, but the cushions are rather short and flat. I removed the cushions and dismantled the electric mechanism. I modified the mountings so the front of the cushion is about an inch higher and further forward. The length and angle are much closer to how I have it in the front, and I think its more comfortable and cosetting. A simple job that's really worth doing. No great effort or downside either.

Nick

Unbalanced Parallel Chargecooler Plumbing

OK, lets see who's paying attention....

So how can you best use a high-flow pump in a charge cooler system? I had a crazy idea last night, thinking about how to make the most of my BMW coolant pump. This is a Pierburg CWA-200, which generates about 0.5 bar pressure over a wide range of flow rates. Indeed, the optimum operating point is around 120 lpm/32 gpm @ 0.5 bar, which is much more flow than I'm using (see post 108) So I thought about better ways to take advantage of the Pierburg's capability. Even if it was flowing much faster, it would still generate much the same pressure (more even), and that's what's necessary to keep water flowing through the IC's. Regardless of what's happening elsewhere in the system (double pumps, double HE's, whatever) the IC simply needs to "see" the right pressure and flow.

Attachment 410016

So I thought about running a by-pass pipe directly from the pump outlet to the HE inlet, by-passing the IC altogether. Doesn't make much sense, huh? Surely you want all the water to go through the IC? My thinking was it would increase the flow through the pump and get it working closer to its preferred pressure/flow point. It would increase flow through the HE, but because the CWA-200 has such a flat output curve, it would maintain output pressure at the higher flow rate. Therefore the IC flow wouldn't be affected, but the flow through the HE would be increased, which is a real benefit.

Attachment 410017

The increased HE flow interested me, but the by-pass to the HE inlet didn't make so much sense, as it would tend to increase the dynamic pressure at the inlet, and reduce the pressure differential across the IC. It was getting late and I wasn't thinking clearly, but I put my head down and things instantly became clearer.

IC systems have high resistance, and its been difficult to cope with that all along. Maybe because of the IC, maybe the HE, maybe the inadequate 3/4" piping. But the thing is they're all in series, with resistance adding up. Why not simply connect them in parallel instead of in series? The two IC's are effectively in parallel with eachother already, so what will putting the HE in parallel do? The ICs will see all the pressure generated by the pump, but the flow through the HE will be higher. My HE is a BMW X3 radiator, which should take the full flow of the CWA-200 - which should be several times higher than the flow through the IC. On average, that means the coolant will pass through the HE several times, for every pass through the IC. It should reduce the temperature of the water going through the IC, due to the repeated HE passes. And as long as the pump is big enough, it should actually increase the flow through the IC, as the pump's output pressure isn't split across both the IC & HE - they both see the full pressure of the pump.

Attachment 410018

Of course, a proportion of the heated water from the IC will go through the pump and go straight back to the IC without going through the HE. But as long as the HE flow is much higher than the IC flow, most of the water going to the IC's will be extra-cold: colder than it would have been from a single HE pass.

So there's a thought - running the IC and HE in parallel instead of in series. Can that make any sense at all!?!?

The reason it might be a good idea is to make the most of what you've got. The charge coolers are an integrated part of the engine - upgrade them and you have no room for large air filters and cold air inlets, which undermines the benefit of the bigger IC's. The pump and HE are easier to upgrade, and can achieve similar improvements. An IC system is simply a very big heat sink for the charge air, and you want to minimise the thermal resistance between the intake air and the ambient air, so that you maximise heat flow outwards.

An air-to-air system is actually quite good at this, as there's very little between the two air streams - just a thin sheet of finned ally with a reasonable thermal resistance. A water cooled system is rather different, as you have to add to this an interface with the water in the IC, plus an interface with the water in the HE, plus the circulation system. All of that can only add to the thermal resistance of the whole system, so you need to minimise the resistance at both the IC and the HE.

Increasing the flow through the IC helps, but its difficult to make big improvements. Having a large heat exchanger and lots of flow WILL make a big difference to the HE however, and provide lots of cold water for the IC's. So the thermal resistance of the HE cannot be too low - even if the water was circulating at a million mph, it will still be worse than air-to-air, as it can only add to what's already there. Radiators normally run at a temperature delta between inlet and outlet of around 10 degC / 20 degF. So that's one delta for the HE, and one delta for the IC. Ideally, both will be zero, but the idea of a big HE and a big pump is to get that 10 degC delta as close to zero as possible.

Of course the V12TT uses charge cooling for packaging and piping reasons, but there are other advantages. The water passages in the HE are narrower than air passages, allowing MORE rows for a given frontal area. My X3 rad has fifty rows - more than any air-air cooler - and there's less obstruction to ambient air passing through, so other things being equal there will be more through-flow. Similar arguements apply to the IC's as well.

Nick

This is the chart that I've been looking forwards to posting for some time. I've combined the curves for the new pumps I've mentioned in recent posts, plus the Lingenfelter test results, and added them to the pumps I compared in post #24 back on page 1.

Pump performance is usually considered using the pump curve and the system curve. These are always plotted on a graph of pressure (vertical axis) against flow (horizontal). The installed operating point is where they cross. I've plotted two charge cooler system curves based on best information I can find (or measure). The lower one is copied from Nunnally's CadyInfo blog, and shows the pressure vs flow curve for the CTS-V charge cooler alone. The upper curve is double the lower one, which reflects various measurements, including mine. The two curves are like best and worst case system resistance, with the upper curve probably being most representative.

Because there are so many pumps, I've ignored the pumps that perform WORSE than the BOSCH, so I've removed the smaller Johnson and Davies Craig pumps (and I was sorely tempted to remove the Meziere, but its a useful point of reference). Still, its a very busy chart, and there were a few more I'd like to have added.

Attachment 410019


Its easily summarised though. There are three main categories of pumps:




1. Stock IC performance: Around 15 lpm @ 50 kPa:

• Bosch 392 022 010
• Meziere WP136S
• Johnson CM90
• Jabsco 50840
• Davies Craig EWP115
• Pierburg CWA-200 and several other coolant pumps like the the Dayton 5PXX0 also fall into this category
• Bosch 392 022 002 is a step behind the others

2. Improved IC performace: Around 20 lpm @ 70 kPa:

• Peirburg CWA-50
• VariMax 410110
• Camero ZL-1

3. Best IC performance: Around 25 lpm @ 100 kPa:

• Pierburg CWA-100
• Flojet DC40/10
• Corvette ZR-1
• Stuart Turner 12/50
• Jabsco 50860
• Stewart-EMP E2512A would also be in this category, as Lingenfelter measured the installed flow the same as the ZR-1
• Pierburg CWA-400 scrapes into this category, but its a coolant pump, and would be stalled as an IC pump

I estimated the Meziere 136 curve for the last chart, based on nothing but open-outlet flow, power consumption plus assumed pump efficency. It turns out from the Lingenfelter measurements that I was only 1 lpm out across the board. When you look at the variation that battery voltage, coolant temp and anti-freeze concentration make to pump output, that means I was spot-on. Do I understand pumps or what? :-)

The dual scales were fun, and I jumped through hoops in vain to make this easy to read. Opening the PDF is the best way.

Enjoy, Nick

I went a whole different direction. I wired up 2 Johnson CM30 pumps in opposite ends and the flow increased dramatically. Run after run wide open and the temps are always back down instantly. I also have the Speedriven intercoolers and bigger heat exchanger so I needed more coolant flow through them. Coolant capacity has almost doubled so I figured two pumps are better than one. Since the Johnson CM30's are magnetic pumps, they draw little voltage compared to the stock Bosch. So all I did was wire them together on the same circuit. This is what I did and its working out great. Did anyone else use this layout?


Attachment 409985

What were the iats before and after?

Curiosity has gotten the better of me, and I've just bought a new S65 heat exchanger and an "SLS Pump" - also known as the new Renntech pump - see this thread.

Here are some pictures and measurements comparing the (new) S65 HE with my (very old) S600 HE.

The S600 HE matrix has 27 rows and measures 580 x 265 x 21 mm
The S65 HE matrix has 45 rows and measures 580 x 400 x 23 mm

Almost all radiators and heat exchangers have 10mm row spacing, so I originally thought the S65 HE was 450mm high, but it appears the rows are about 9mm apart, so its only 50% deeper than the S600 HE. It is a little bit thicker though, and the header tanks are a much better design.

The S65 HE obviously has the steering cooler built in, but its done in a strange way. You'd think that the smaller cooler would be attached to the bigger one, but if anything its the other way round.

Nick

Edit:

The Stewart E2512A is also sold as the EMP or Europump WP29, and there's lots of information about it:

http://www.europump.ca/uploads/WP29.pdf
http://www.emp-corp.com/media/MarketingMaterial/WP29/SpecificationSheets/ElectricWaterPump.pdf
http://www.emp-corp.com/products/advanced/WP29-electric-water-pump/

Its a big pump - at full speed it pumps over 15gpm @ 20psi. Its quite expensive (though not unreasonable) needs one inch pipes, and it uses a lot of power, but it has more flow AND more pressure than anything else. The motor is water cooled, like the Pierburg pumps, and it has electronic control and an electronic communtator (no brushes), so it has a big MTBF. If you want the best IC pump, this is it.

Nick

I've been looking into the Lingenfelter tests some more, and the Stewart E2512 pump in particular, which was way ahead of all the other pumps. Lingenfelter described stock and modifed installed performance, but didn't give the pressure/flow curve for the "stock" pump.

Initially I shied away from any sort of overdriven product that might work great for drag racers, but not for a daily driver. However, I think Lingenfelter have been a bit misleading about their modification or upgrade. Looking at the EMP websites and documents for the pump, it seems that the upgrade performance is simply the EMP Stewart pump running at full speed. Lingenfelter are simply proving EMP's own performance claims.

So I added all the data I could find to my pump comparison chart, so everyone can see just how good the EMP WP29 is (and how bad the CM30 is - which is still promoted as an upgrade by a lot of tuners). A word of caution though, the EMP only achieves those results when used with 1" hoses.

Attachment 410020

Its a busy chart, but the PDF version is much better as usual. Note that I've expanded both axes to fit the EMP results, and I've added the Johnson CM30 back in, so everyone remembers how weak it is.

The legend lists the pumps in the order they appear on the chart, apart from the EMP curves, which are at the bottom because I plotted it against the secondary (imperial) axes.

Notice the big difference between the 12V and 14V curves - it affects comparisons between closely-matched pumps. As does anti-freeze percentage and coolant temp, to a leser extent.

Nick

That's all the theory I have on pumps; there is no more.

Now it's time for a shootout....

Attachment 410021

Thanks veeeery much for all your work.
Really its time to see some real world results :-) Recovery time, max IAT, etc.
Thanks alot, i look forward to your reviews and ongoing commitment :-)

It would be upsetting to go through all these books of informations and supposed changes for the better only to see continued heat soak. :smash:

The other thing I’m doing is re-plumbing the charge coolers. I decided not to move the pipes from the side of the engine. I don’t want the coil packs getting any hotter than they already are. I’m changing everything else at the back, for the following reasons:

1. I’m adding a small header tank to fill the system, check the level, and top it up – the usual cooling system stuff.
2. I got rid of the valves on the IC inlets, and added bleed nipples instead, to help bleed the system automatically.
3. My pump controller needs a coolant sender, so I added an in-line sensor adapter to one IC outlet.
4. The system had four 90 degree equal-Tees, which are terrible for fluid flow, so I removed them all.
5. To split the flow between the coolers, I use a pair of Y-splitters instead, which should flow better.

The plumbing down the back of the engine is tricky, so I removed it all, and traced the pipe geometry onto a large sheet. I then got all the stock hoses together, added the couplings and adapters and some new hose, and experimented with countless configurations to fit the geometry. I was careful not to cut or damage any removed parts, so the modifications are reversible. Hint: if you’re ever likely to do anything with the charge coolers, be very careful with the inlet and outlet ports. They’re made of ally, not steel, and easily damaged; just removing a hose is enough to bend them.

My original intention was to add bleed nipples in order to bleed the system when its filled, but it’s not as simple as that. Bleed nipples fit into bleed receptacles, so they can be closed off by tightening down. The Schrader valve ports won’t do this – there are no conical seats, so the nipples will always be open. I initially got some two-piece Stahlbus speed bleeder valves; perfect for this application. However, I changed my mind and decided to keep the bleed ports open all the time, not just for filling, so I used conventional 6mm bleed nipples instead, and hooked them up to the header tank with fuel hose.

The bleed nipples are on the IC inlets, and the header filler connects to the outlet, so the header tank effectively straddles the IC’s. Therefore the pressure across the IC’s generated by the pump will force coolant through the bleed pipes, so the coolant in the header will be continuously circulating. The bleed apertures are small, so they won’t take much flow away from the IC’s. In any case, my pump is working in a flat part of its operating curve, so it will probably supply the extra flow without losing pressure. Now, any air in the system can be clearly seen, which is something you can’t do with the stocksystem – what’s going on inside has always been rather a mystery.

Has the system been properly filled?
Has any air got into the system?
Is there a coolant leak?
How hot is the coolant?
Is the pump actually working?
Is there even any coolant in the system?

The header tank gets round most of these problems, but it does need to be used in conjunction with an HE that can be bled properly, without requiring a vacuum bleeder. Not that that’s the end of the world of course – have you seen the price of those Chinese rotary vacuum pumps these days? You can service both your AC and IC!

Nick

Attachment 410022

Attachment 410023

One of the things that convinced me it was worth upgrading the V12TT IC was realising that even modest cars like the Golf 1.6 TDI had large air-air IC's; around the same size as the engine rad, in fact.

I hired a Golf 1.4 TSI DSG last week, and it was very pleasant surprise in all respects. Looking under the hood, it had the tiniest turbo I've ever seen, but it also had a water cooled IC. All the recent VW/Audi/SEAT/Skoda engines seem to be moving in that direction.

According to Nissens, that little 122 bhp 1.4 TSI engine has a 620x410x16mm HE - about half the volume of the air-air IC, and amounts to 4067 cc volume. The stock S600 HE is 3228 cc.

Food for thought.

Nick

So do you have results yet in which pump does the best IAT reduction and shortest recovery times?
I'm very keen to know what is best to be bought by me. :-)

Thanks alot Nick ;-)

The best pump depends on the system, and in particular what HE you're using. The information needed is in post 118.

For my system, the CWA-100 pump will almost certainly be better than the CWA-200, as it should raise 0.85 vs 0.50 bar pressure, though you can get more flow from cheaper pumps.

I think the best bang for the buck is the Jabsco 50860 (for intermittent operation) or the Stuart Turner 12/50 (for continuous operation), though they may be noisy, and need periodic maintenance.

Pumps like the VariMax, Pierburg and EMP/Stewart benefit from electronic commutators, so they should be more reliable.

Its taken me six months of evenings and weekends to find out the hard way how to do it - a lot of that being down to figuring out what radiator to use, and how to fit it. In hindsight, I guess I would have skipped getting the Tekomotive pump controller (clever though it is) as the wiring took so long. I would simply wire any pump to come on with the ignition, and use the stock relay to switch it (unless using a Pierburg, which you can effectively wire up as if it was the relay itself).

In the meantime, I've been without a properly functioning IC system for six months, which is starting to get very frustrating....

Nick

Thanks alot. Where can i get the Pierburg CWA-100 Pump? No results in germany so far. Maybe i check with Pierburg directly, as 500€ at MB Dealer is ridiculously expensive IMHO.
And is it "CWA-100" or "CWA-100.2" ???

Or i go with the Jabsco, it sounds really good and i have done a bit of research in germany about their products - they are top notch and they are market leader in pumps in the Marine sector :bow::eek:
They call their pumps themselves as "heavy duty" and "long running" - whatever that means.... :zoom:

Check this: http://xylem.anbeca-software.de/Prod...ide%202013.pdf
Page 19 are the pumps we need? High Pressure or High Flow? I guess for the small IC lines, "high pressure" is better, right? :-) :naughty:
****... http://www.xylemflowcontrol.com/file...pspgs10-19.pdf
Motor Life Time of just 2500 hours... Pierburg and others have 6000 hours MTBF, IIRC

Saying goodbye to the s600, did not want to waste any more time on a 10yr old car, bought a new M5. Good luck to all trying to get reliable power out of these v12s.

Sorry to hear that, but if you can afford a new M5 (or E63 or RS6) then why struggle with a ten year old car? From reading your other posts, it sounds like you gave up on your ABC. I would never consider running an S600 without serviceable ABC.

Nick

I think I have found a source of a large quantity of Pierburg IC pumps for a low price.

I'm just trying to establish whether they're CWA-50 or CWA-100.

Nick

thats the intercooler i designed for my vr5 turbo running 2bar of boost :)
i has 4 chambers, so water needs to flow 4 times through it before leaving the intercooler. heres a pic of the prototype install on my car. i drove a
pwr radiator before. but i was way to small for my ammount of boost:y
tell david chris from germany sent you, and he should give you guys a small discount... he is a cool guy.

do you think, a 2 pass radiator is better than the 4 pass. i am thinking about reducing the water flow in the radiator
so water have more time to cool down.

https://fbcdn-sphotos-e-a.akamaihd.n...49672317_n.jpg

at the moment i am running the ewp80 pump, but i will switch to the cwa-50 i think.
i can get it here for 249euros.
greetings from bavaria :)
chris from carlicious-parts

I got in touch with David Wang at Winner back in June, and was impressed with their speed, value and flexibility. I even drafted up a drawing of an S600 custom heat exchanger, but didn't take it any further as I was running out of available time. Next time round I think a custom HE from them would make a lot of sense.

Do I understand that they made that HE specifically for your application? Is that a modified VW Golf? What's the quality like? I'd like to think an S600 HE could become part of their standard product range, and wouldn't mind being a guinea pig.

I'd prefer a dual pass cooler, as the flow resistance will be high with 4-pass - definitely too high for an EWP80.

Nick

the HE you see on ebay and on my car is one and the same. it was my drawing and david did the rest.

its a modified jetta mk4 wagon with 450hp and 600nm running a 2.4l 5cylinder engine on 2bar of boost. due to the 19mm connectors i cant recognize any pressure drop in water flow. but i agree with you, maybe i will switch it to a 2pass design. first i need to get a bigger pump and bigger chargecooler.

same thing, different problem. bigger He -> to small pump, bigger pump -> to small chargecooler, etc. etc. :rolf::smash:

about the quality. its no pwr radiator if you ask for that :) but lets be realistic. that baby is hit by 1000´s of small stones, bee´s, and other **** that you hit at 200 miles. so it dont need to look perfect. for that pricing the quality is more than ok. i dont need a polished bling bling HE just to show who has the biggest balls :crazy: :D

so if you want a good HE for fair pricings. go with the winner cooler. and as you said, they are really fast on email support and realization of custom projects. as you see on the pic above, i drove a pwr HE before. and they cost an arm and a leg just because of the company name and 100% straight weldseams. i dont need this.. you dont see my HE anyway.

heres a pic of my engine bay. no benz, but a 1200kg wagon with 600nm is more than nice :zoom:
http://abload.de/image.php?img=motorraumo3uwv.jpg

http://abload.de/image.php?img=bobo59humj.jpg

btw. did anyone of you tried "water wetter" in the cooling system. maybe it will help dropping the intercooler temps a bit.

That's quite a project. I assume you gave Winner the design for the HE mountings, too?

That being the case, here's a proposition for Mercedes S600/CL600/SL600 owners:

Nick

yeah of course i gave it to dave. its no problem to weld on the 4 mounting points. even i could do this :y:D

for me it was not that complicated. i need to drill holes anyway. but for replacement intercoolers of course its more important. should be no problem at all!

Getting the right dimensions for a custom HE is tricky, but of course I already have a template - the S65 HE!

I think I've also realised the answer to an old question - why are the regular HE & condenser bolted at the top, but mounted with slots in bushes at the bottom? Any why the funny arrangement with the PAS cooler on the S65 HE?

Car parts are usually designed in a very specific way for very good reasons, even if we don't know what they are. Makers don't like spending more money unless they really have to.

I had originally imagined that you could bolt the HE to the rad at four points, but I guess you can't . It would probably put too much stress into the HE, as the main radiator expanded and contracted during heating and cooling cycles.

So I think I can take dimensions off the '65 HE for a custom '600 HE, but I'll have to add some compliance to the top mountings (where the condenser is bolted) to allow for expansion & contraction.

Nick

OK, I finally finished the IC installation, but it doesn't quite work like I hoped.

Here's the new plumbing behind the IC's. The controller thermostat is plumbed into the RHS IC outlet, and the header tank filler is conencted to the LHS IC outlet. The bleed pipes to the header are connected to the IC inlets.

Attachment 410024

The header tank is mounted between the heater inlet and the engine compartment partition. Its bolted to a bracket that I mounted to a stud on the wheel arch. The hoses run through a grommet in the partition.

Attachment 410025

I mounted the pump controller in the ashtray, and ran the power and control lines to the pump controller through the RHS fuse box and through the dash. The pump is powered from the battery, and the controller is powered from the ignition. The connections can all be found at the fuse box and the small TT/AMG aux relay unit.

Attachment 410026

I removed the metal ash tray from the drawer and mounted the controller on brackets. The wires were connected with terminal blocks, and dressed to allow the drawer to open and close normally. I did try using crimp connections for all the wiring, but everywhere I wanted to use it, it was difficult to get the crimper into the space available. Screw-down terminal blocks and a screw-driver aren't so professional, but are much easier to use.

Attachment 410027

Unfortunately, the header tank didn't fill the system anything like as well as it did when I had it connected down at the HE inlet. It took a very long time to fill the system, and since I've spent so long on this already, I can't justify the huge extra time to compare pumps in-situ, sorry. I'm just going to go with what I've got.

Since I still have lots of IC parts, including the stock HE and a flow-meter, I might try to set-up an IC pump test rig, but off the car (and configured without any air-locks!).

The other thing I might try is to connect the header tank to the HE inlet again. The direct route runs past the LHS turbo, which I want to avoid. But today it occurred to me that I might be able to run the fill pipe through the wheel arch, and route it downhill all the way. I think that might be the best way to do it.

Nick

Well I'm grumpy now, this is getting hard work. I've been struggling to get the pump running under controller operation. I had them running together with a direct connection, but not with the controller fitted in the ashtray. I assumed it was my wiring, so I reworked and re-tested all of that. Continuity, isolation and voltages seemed to be right in every case, so I disconnected the controller and ran the pump directly from the ignition. Still didn't work. Seems that sometime during the installation the pump failed. :(

I do have a spare CWA-200, but it has a different configuration that would mean changing the installation. Since I have a CWA-100 now, I may as well fit that instead. I'm sure that will be a better IC pump. And I was really looking forwards to driving around in a V12TT on full song today, for the first time in 7 months.....

Nick

I just found another Pierburg pump controller, by SFR Electronics in the Netherlands:

http://felixvandaal.nl/index.php/en/...cwapc-3-detail

http://felixvandaal.nl/downloads/MS_...al_EN_v1.0.pdf

http://felixvandaal.nl/images/CWAPC/DSC06705.JPG

This has "proper" digital display of target and actual temperatures, and looks easier to use than the Tecomotive unit, though the documentation isn't up to much. I think I would be inclined to use this, but its a bit late in the day now. Its about twice as expensive though - about 250 Euros.

Nick

Obviously, I've been scouring all sources for info on charge cooling.
I stumbled across a batch of bankrupt stock, from a company that (used to) convert vans to electric drive.
Pierburg pumps are used to cool batteries, as well as intakes.
So the other thing I need to do is find out for sure what these are: CWA-50's or 100's?

Attachment 410028

Nick :)

I know I've got lots of other things to do at the moment, but I just had an idea. Using an engine rad as an HE isn't exactly a drop-in solution, and even when there is something the right size, the inlet & outlet ports always foul the hood, oil cooler, impact bar or whatever, and make fitting difficult. True heat exchangers always seem to have their ports in-line with the matrix, rather than perpendicular. Because of that, I had to modify the headlamp and rad support brackets on my car, which would have been good to avoid.

Now, if I have to do some sort of mod, then why not mod the rad instead? All I need are 3/4" inlet and outlets, so what's the chance of being able to fit them to an egine rad, and block off the stock ports? Maybe I can do that by cutting, gluing, welding or clamping a tank connector, or something like that. Plastic rad tanks can be repaired, so the capability must be there somewhere. Getting access to the inside of the end tanks would be desirable, so I might try to remove the tanks off my E-class rad, now that I've butchered it about enough to scrap it. How hard can it be....?

Nick

:y did you found out something. where did you get those from?

They were a batch of new, unused stock from a bankrupt auction. I think the company was Azure Dynamics, who used to convert Ford Connect vans to electric drive. The pumps were part of teh battery cooling system. There were some heat exchangers for sale as well, but they didn't look very useful.
First I need to find some connectors, then I can test them.
So if you've wondered what your radiator looked like inside, this is it:

Attachment 410029

Attachment 410030

I bent the fingers back using screwdrivers and pliers, and the end tank tank came off. It's sealed to the radiator matrix with a rubber seal that's simply compressed in place. I don't think it looks too difficult to modify with end connectors on the tanks. That would make installation much easier.

Nick

HELP!

I'm having problems integrating the AMG pump. I can't get it to work using the Tecomotive controller - it just runs all the time (and runs very well, too). Maybe the Tecomotive only works with the CWA-200, so I may yet go back to that.

I tried running the CWA-100 off the stock supply, and guess that's OK, but I really wanted
to have it running continuously. So I connected the stock relay to the ignition, but that didn't work either.

Does anyone know how the V12TT pump relay works?

Is it enables with a switched 12V control, or a switched ground?

Thanks, Nick

Edit: Scrub that, I think I've got it sorted.

I only got drive my car with the upgraded IC system in anger for the first time this week. Once warm, I drove up all the long hills I could find, and ran WOT as much as I could. Although I haven't starting monitoring IC coolant temp or IAT yet, I did jump out and feel the IC's. They were cool to the touch, not cold, not warm, probably around 60F. The IC HE and all the pipes were the same. I never got that with that sort of driving before.

Nick

Nick I am interested in buying a couple of the pumps...
 

If you think they are CWA100s I might be interested in a couple as spares for my SL65.

execmalibu@gmail.com

Jeff

Hi Nick,
I've been following your posts for quite some time. Thanks for the great work you have accomplished so far.

Can you please, draft a diagram showing the new layout of the intercoolers piping as well as the new header tank. Where can I get one of those tanks. Thanks.

See post 127 or thereabouts for the plumbing layout.

http://www.ebay.co.uk/itm/Alloy-0-5-...item232658abd4

I bought one of these, but hold fire until my system works properly, otherwise this MIGHT not be the right way to deal with system bleeding.

I'll be testing and comparing the CWA-100 pump soon, as well.

Nick

The Money Shot
 

Attachment 410031

After 18 months of ownership
After 12 months of research
After 7 months of work...


Tonight I finally have my charge-cooler system working the way I always envisaged, with the following modifications:

1. BMW X3 engine radiator as heat exchanger
2. AMG / Pierburg CWA-100 circulation pump
3. Tecomotive customised pump controller
4. Header tank / swirl pot for filling and bleeding

For various reasons - bad wiring, wrong procedure, bad pump, wrong spec, whatever, I'd never yet been able to get everything together at the same time, in particular with the installed controller actually controlling the installed pump the way its supposed to be. I especially wanted to implement continuous running and variable thermostatic control of the pump, with the driver being able to control and monitor the IC temperature.
I was also worried about being able to fill and bleed the system using conventional means. To begin with, the pump seemed to froth the coolant, and I wanted to run it slowly and fill and bleed the system by itself. If you look at the pictures, you can see the tanks inlets have clear jets of coolant, with no air or vapour bubbles. Those two jets take coolant from the highest points of the system - the old "fill" ports on top of the IC's, which I had drilled out and fitted with bleed nipples. Those bleed outlets have drilled holes only a millimetre or two in diameter, so are less than one percent of the main pipe's cross-sectinal area. Yet you can still see good, strong jets into the swirl pot (which swirls quickly enough to generate a deep conical votex). Those pictures were taken with the pump running at mid-speed setting, so the unrestricted flow in the 3/4" pipes must be strong.

I'd particularly like to thank Tobias Mucke from Tecomotive for modifying the controller for IC duty, and for providing superb support all along.

I think I'm done now.

Nick :)

Attachment 410032

Thanks Nick for your elaborate details. Is it possible to list the additional plumbing parts needed such as bleed valves, piping, elbows, etc. Your work has inspired many of your followers, I believe we all need to get it right down to bolts and nuts :)

I haven't compiled a parts list (important though it is), but I did draft a top-level procedure. Of course, this relates to the complete mods that I have done, and other people won't want to do all of them.

For example, on an S65 it would be difficult to replace the HE, as its sandwiched between the condenser and the rad, and it carries the PAS cooler.

On the W216/221, the radiator is a different size to the W215/220, so the BMW X3 rad might not fit.

People can pick and chose which mods to incorporate, and work out the parts required. They're mostly independent mods, though if you want to use the Tecomotive pump controller, then you obviously have to use a Pierburg pump.

For many things, its more down to materials rather than parts. For example, you'll need lots of 8mm and 19mm hose, jubilee clips, tie-wraps, automotive wiring and MB 325.0 antifreeze.

There are two main things to concentrate on:
1. How to connect inlets and outlets of the radiator and the pump?
2. How to plumb the temp sensor for the pump controller into the IC circuit

I'll try to put a more definitive parts list together, as one of the worst things is starting the job and realising that it takes a fortnight to get an essential part. Here's the old procedure, though:

Nick

1st Feb 2014: seven months and many hours and pounds after starting work, I finally got to drive Mercury Mercedes with the IC system working how I wanted it. I didn't have an OBD2 reader hooked up, so I couldn't see IAT's today, but I could see the IC coolant temp from the Tecomotive controller.

The initial impression was that the motor was always "on", and I wasn't taking a chance or patiently waiting for the pump to switch on. The car felt eager and responsive, rather than slightly languid and reluctant to get going, like it had to be poked with a stick.

In normal running, the controller's crude temp display showed 2 LEDs out of 7. The LED's correspond to the target temp range that can be set - 0 to 55 degC. I had the controller set to position 8, which was 20 degC. On WOT the temp generally went up to LED3 - say 20 to 25 degC. On one instance - as hard as I ever drive - standstill to a high speed up a long hill - it showed LED4 for a few seconds. I think that's about 28 to 31 degC, and that's with a target temp of 20 degC. It went back down in a few seconds.

There's lots of tunes to play, and the next thing I'll do is set a lower target temp, and then change the display to show pump speed. When I got home I left the engine running and felt the charge coolers, which were again cool to the touch, even after hard driving. The IC header tank was swirling nice and slow, with no sign of bubbles in the coolant, which I was very pleased about.

Because this endeavour has been so difficult, I've had doubts all along whether I've taken the right approach. There are many short-cuts which might have achieved the same objective - different pump, different HE, skip the Tecomotive controller, etc. But I'm happy now that I've done the right thing. I think many people might prefer a custom-made HE to make the installation easier, and I wouldn't argue with that. If I did it again I'd be tempted to use an E-class radiator instead of the X3 rad, and use that to cool the ABC oil instead of the silly stock cooler. But in hindsight I don't think I've taken the wrong path - I think this is how all V12TT's should be.

One other thing. Before I put the under-tray back on, I measured the current consumption of Pierburg pump. When the AMG pump was full-on, it measured 4.9 to 5A in my system, and I got 4.9A from the Transit pump, so I take it that my batch of coolant pumps are indeed CWA-100's, rather than CWA-50's. Only difference is teh Transit pump actualy responds to the Tecomotive controller. That's good news! Anybody else want one?

Very happy now

Nick

Hi Nick,

is there a Part Number with this format "7.01360.??.0" on those pumps you got?

Paulo

Hi Paulo,

Yes, the number is 7.01360.48.0

Can you tell me what that means?

Thanks, Nick

that is pierburg part number. It means they are CWA-50s. Does the BMW connector that you bought for the CWA-100 fit them?

Paulo

I've been trying to find that out for months - how do you know? Do you work for Pierburg?
I believe the BMW pump is a CWA-50, and has the same conenctor as the AMG pump.
The CWA-200 is quite different, and has a different connector (which I completely forgot about until I brought the AMG pump home).
The Tecomotive kit comes with a CWA-200 connector, so I had to get the BMW connector, which fits both the AMG and Transit pumps.

Any ideas why the AMG pump won't talk to the Tecomotive controller?

Cheers, Nick

Hi Nick,

I have been trying to source the connector for these Transit pumps, so I have been talking to a guy at piersburg.

I will ask him about the CWA-100.

One other question, are there any numbers on the BMW connector? I'm sure they are not produced by BMW.

Paulo

If anybody’s interested, here are the connectors and pins for the CWA-200.

These are the parts I ordered from my BMW dealer:

1 x 12527549033 Plastic Socket Housing
4 x 61138366245 Rubber sealing grommets
4 x 12527545858 Individual Socket Pins

Here are the parts that I bought (two sets - one for the car, one for the pump test rig)

Attachment 410033

You can probaby crimp the cables to the pins, but I soldered them instead:

Attachment 410034

And here's the assembled connector:

Attachment 410035

It's not very easy to read the numbers on the connector from the last photo.

Can you tell what they are?

Paulo

A close up photo from those numbers and from the markings on the other side would be really helpfull.

Paulo

So you have some great results already with the CWA-50 pumps?
I cant imagine how good it would be with the CWA-100 :-)

These are the markings on the connector:

PA66 GF25
872-859
7 549 033-01
1 2 3 4

The first thing I did when I got the connectors was search for markings and part numbers that would let me find another supplier, but I found nothing. The only meaning I could get from them was the plastic material. I suppose I was put off going to a BMW dealer to buy a connector, but don't be. They only cost a few pounds and took two days to get in. I believe Audi and Mercedes use this pump on their most recent and most expensive turbo cars, so maybe you could get the connectors from them, but you definitely can get them from BMW.

These are the connections that I made, and which do work with the Transit pumps:

Pin 1: Ground : Red
Pin 2: Ground : Yellow
Pin 3: Control : Blue
Pin 4: Battery : Black

For some reason the pin-out is reversed with the CWA-200 pump.
Nick

I hate to throw yet another option in the pot, but I was answering someone's question on a Jabsco pump, and came across another model. I thought I'd already seen and disregarded it for some reasons, but maybe I looked at the wrong version. Jabsco make a wide range of industrial and marine pumps, and I've previously mentioned that the obvious one for IC pumping is not the popular 100lpm 50840, but it's high-pressure big brother, the 80lpm 50860.


50840-2012

• £ 232.27 ex. VAT
• Connections: - ¾”BSP internal threaded ports
• Dimensions: - 163mm long, 120mm wide, 122mm high
• Fuse Size: - 10 amp
• Maximum Current: - 7 amp
• Output: - 100 litres/minute (22 gallons/minute) @ 0.1m head
• High output, continuous rated centrifugal pump
• Robust stainless steel body
• Can be used with shut-off nozzle
• Non-self-priming, requires flooded suction
• Maximum recommended total head 4m
• Packaged Dimensions: L:24.00 x H:13.00 x W:18.00cm
• Actual Weight: 2.98 Kg (Approx. 3.48 Kg packed)

50860-2012

• £ 253.38 ex. VAT
• Connections: - ¾”BSP internal threaded ports
• Dimensions: - 205mm long, 120mm wide, 122mm high
• Fuse Size: - 25 amp
• Maximum Current: - 21 amp
• Output: - 80 litres/minute (18 gallons/minute) @ 0.1m head
• High output, intermittent rated centrifugal pump
• Robust stainless steel body
• Can be used with shut-off nozzle
• Non-self-priming, requires flooded suction
• Maximum recommended total head 12m
• Minimum recommended total head 5m
• Packaged Dimensions: L:31.00 x H:14.00 x W:16.00cm
• Actual Weight: 4.90 Kg (Approx. 5.40 Kg packed)

There are lots of circulation pumps on the market, but they're usually high flow/low pressure pumps. If you want more than one bar output pressure, you move into the world of positive displacement / diaphragm pumps, which are noisy and less efficient. However, I overlooked some Jabsco rotary vane pumps called "Utility Puppy", and there's a 12V model called the 23920-2213 (UK) or 23920-9213 (US) which maxes out at 50lpm, but can raise 1.2 bar pressure at 25lpm. This would make it the 2nd best IC pump after the EMP WP29, and it has a couple of advantages of its own. The motor is continuously-rated; its self-priming and it tolerates air in the water, or even no water at all (unlike centrifugal pumps).

23920-2213

• £ 320.00 ex. VAT
• ‘Utility Puppy 3000’ self-priming pump 12 volt d.c.
• Connections: - ¾” BSP inlet and discharge ports.
• Dimensions: - length 220mm, width 125mm, height 105mm
• Fuse Size: - 25 amp
• Maximum Current: - 21 amp
• Output: - up to 50 litres/minute (11 gallons/minute).
• Max 12m head
• Built-in dry-running protection for up to 10 minutes after initial prime
• Rapid self-priming from dry up to 2.4m
• Carbon-ceramic shaft seal
• Non-rusting bronze body
• Continuously rated electric motor
• Simple, fully serviceable design with few wearing parts
• Packaged Dimensions: L:29.00 x H:14.00 x W:16.00cm
• Actual Weight: 4.65 Kg (Approx. 5.15 Kg packed)

http://www.jabscoshop.com/marine/pum...12-volt-dc.htm
http://www.google.co.uk/url?sa=t&rct...60983673,d.d2k
http://www.jabscoshop.com/files/UTIL...C%20doc553.pdf

Just a quick observation from driving around this week.

I expected the Tecomotive controller would let me test and monitor the IC system, but I wasn't expecting many surprises in the performance department. With the target temperature set at 15 or 20degC, the actual IC coolant temp moved little. Under full throttle I guess the coolant temp went up 10degC at most, then came down very quickly afterwards.

However, the highest coolant temps that I saw were under quite different circumstances. After driving a few miles - sufficient to get the engine fully warmed-up, I sometimes parked for an hour, and went out again later. In those cases the engine compartment would heat-soak the IC system, which then reached 40degC (and the IC & HE were physically warm to the touch).

For a mile or so, the IC pump would come, but with the engine temp at 70 or 80degC, the engine fan ran slowly, and there wasn't much airflow while driving slowly to begin with, so the IC temp took a few minutes to go down. Speed-up again - and normal operation and temperatures were resumed.

It has only been that particular circumstance where I've seen higher IC temps - no amount of fast driving would ever get it so high. That suggests that the stock IC system will also see similar behaviour - without being able to see what's going so clearly. Of course it will also cool things down, but with the thermostatic control set to higher temps than I used on the Tecomotive, it might not necessarily bring temps down so quickly.

So just an interesting and unexpected observation.

Nick

Managed to source a CWA50 from fleabay. could be one of the ones listed here.....:)
7.01360.48.0
anyways - I tried to decipher the plug connector for it - I'll need one of these....
Is this photo for the CW50 or CW100-200?
I'm also interested in setting up the comms from my Arduino rig, but it may be easier just to vary the voltage to the motor through +/- than start trying to play with comms...

End use - SLK32 Intercooler setup.

If someone can send a link to an operating manual for the pump - much appreciated

Any thoughts?

Heat soak and low speed temps
 

That other controller you referenced from SFR Electronics has a relay output for a fan. Does the controller you are using have that option? If so maybe a fan on the intercooler radiator that is controlled/based on IC temperature would help when stopped or at low speeds (especially after a heat soak event). If it had a timer on it you could even let it run on after the engine shut down (but not for long or the battery would be dead).

Yes, the Tecomotive controller also has a fan controller, though I don't use it. The main engine fan won't run when the engine doesn't, but the tecomotive can be configured to run anytime.

One snag, though. The X3 radiator was chosen to fill every millimetre of space available, and it does. There's no space anywhere for another fan, either in front of, or behind, the radiator pack.

I don't think its a big problem though. When you stop for a while, the engine cools down and the IC warms up. When you start off again, you wait a mile for the high temp system to warm up and the low temp system to cool down again. I wouldn't floor it with an engine that hadn't warmed up.

Nick

Post shut down circulation
 

Could you configure it to to keep circulating the pump for a while after shut down. That would help reduce the heat soak - better for long term durability of everything and you start at a lower temperature when you get back in the vehicle agaon so you have better power on initial re-start.

Have a look at the pictures in posts 89 & 90. There really isn't any room for another fan.

Nick

Not sure that would work very well with an electronically controlled pump. According to Jason Haines at Lingenfelter, it's the brushed motors that respond to voltage. However, they posted curves for the monster EMP pump that had more flow at higher voltage. Maybe you'd get some control. When I was testing my CWA-200, I did get more output, but not much.

http://www.lingenfelter.com/LPEforum...pump-test-data

I found the engine cooling fan wasn't working at the time.

I just replaced it with one out of a 320CDI - its the same 850W motor and controller. Lost a chunk of my hand in the process.

Nick

Howdy,

I noticed that the 2013-2014 Shelby GT500 is using a Pierberg pump. Anyone know if it is a CWA 50 or a 100?

I found some picture and price references for it. One vendor quoted $251 just for the pump.

See info below.

http://www.svtperformance.com/forums...connected.html

And at the bottom of this one. Also check out the massive stock heat exchanger.
http://www.svtperformance.com/forums...-h-e-pump.html



I had a thought that since there are harnesses out there to upgrade your older Shelby which uses a Bosch pump just like ours that there might be a chance the electrical plug is the same. Anyone have a "Ford " bosch pump they could compare our plug with? It would be sweet that you could possibly buy the new ford pump and the adapter harness would make it plug and play with the MB connector.

This is from the VMP website.talking about the harness ($59)
The 2013-2014 Shelby GT500 uses a higher flowing intercooler pump than previous years, it is an upgrade for 07-12 GT500s, 11-4 5.0L, and other vehicles that use a Bosch pump. The bosch pump flows approx 4.5GPM with an aftermarket HE, the 13 pump is good for 6-7GPM. On the road course we saw lower peak temps after running WOT on the straights with our VMP TVS Supercharged 2012 Boss 302.

Our kit includes:

3ft Extension/Adapter harness
OE Pump Connector
Adel/Cushion clamp for mounting 2013 pump 3225T8

hey guys. i got a problem. i want the cwa100 pump run in a normal car without the pwm impulse.

how can this be realized? ground to 1 and plus to 4 wont work.
ground to 1-2 and plus to pin 4 wont work either :(

any more suggestions?

Connect 12V to pin 3 - that's the control signal.

Pin 4 is the high current power pin for direct connection to the battery.

Nick

Sorry, I don't think you can tell without being able to decode the manufacturers part number.

Nick

I still need a source (prefered in europe) that sells the CWA-100 pump. Pierburg here in Germany dont answer me to my emails.

I can sell you a CWA-50 or two. I have a few....

Nick

There's one CWA50 left on ebay.com from a guy that had a few....maybe nick...;>)
I bought one.

b22b

so 12v to pin 3 and pin 4, and ground to pin 1 ?

Have you got a link? I've never found anyone advertsing a CWA50 or CWA100.

Cheers
Yup. It doesn't switch on immediately. It waits for PWM control signals, then defaults to full speed. Takes a few seconds.

Nick

ah ok. so pin 2 is empty? or should i ground it?

I don't think it matters. I grounded mine.

http://www.ebay.com.au/itm/261415699...84.m1497.l2649

b22b

i can get all the pierburg stuff

I got your original reply via email.
Nice link thanks. Is that the cheapest option for 300€? :naughty:

Got the CWA50 running at 100% duty cycle last night.
ON the CWA50 - You can find Pin1 by checking resistance to the pump chassis - zero - ground = Pin1
You need your negative battery terminal on Pin 1 and your +12V on Pin 4 + another lead from the 12V to the PWM signal pin3 (Preferably through a 1.5kOhm resistor to limit current).
Pump takes 2 seconds to start.
I'd assume the CWA100 and 200 would be very similar.
Although the signal pinouts are different between the CWA50 and 200...

CWA200:1 – VBatt, 2 – Signal, 3 – Signal Ground, 4 – Ground
CWA50: 1 – Ground, 2 – Signal Ground, 3 – Signal, 4 – VBatt
compliments of tecomotive

12v on pin 1??? thats the ground!

apologies for the ambiguity....corrected in previous post

I've been in touch with Felix van Daal at SFR Electronics. He speaks english, and is very helpful.

Their Pierburg pump controller, like the Tecomotive controller, is intended for engine cooling applications, so the minimum target coolant temperature is too high for intercooler applications (60 deg C in this case)

However, like Tecomotive, SFR will ALSO supply a pump controller with customised firmware to drop the minimum temp to whatever we want.

This is a great opportunity, as the SFR controller has a better user interface and display. It displays actual and target temperatures, plus actual pump speed, all at the same time.

When the coolant is below target, the pump doesn't switch off completely, but runs at 25%. This is perfect for the V12TT, to avoid heat-soaking the coolant pipes down the sides of the engine.

Not many people have picked up on the significance of the pump controller, focussing instead on bigger pumps. I think the pump controller is perhaps the biggest improvement to the IC system in day-to-day driving.

What a great toy.

http://www.sfracing-group.com/images/CWAPC/DSC06705.JPG

Nick

Hello Nick.

I appreciate your hard work!!!! Excelent and great work. I have read al the pages. I own an SL600. I have Speedriven upgrade IC coolers and want to upgrade the HE also.
Can you tell me which modell of pump this is…. I now it's a BMW / Pierburg pump but which model?

Thank you.
Juha

Hi Juha

That looks like a Pierburg CWA-50, commonly known as "The BMW Pump".

Is that what Speedriven supplied for your car?

Nick

Yes… Speedriven sell this pump… and this pump is very quiet when it's run.. Like a OEM pump.
I have it at my car now and want more cooling. … I have not decide yet which HE I'm gone buy.
What do you recommend me to buy? I'm not sure if S600 and SL600 have the same space at the front and if both have the same OEM HE.

Juha

If you want more cooling, one of the simplest solutions is to add a second pump in series.

Otherwise there are lots of good pumps featured in this thread. Just as long as you don't use any engine cooling pumps like Johnson, Meziere or DaviesCraig. You need something that will generate around one bar of pressure. The best of all is the EMP WP29.
But increasing flow with more pumping power is doing it the hard way. I found you could get much more flow by reducing the flow resistance - fit a larger heat exchanger.

Which is best for the SL600: I have no idea. I've done a LOT of research on the W215/W220, but I know nothing of the SL, sorry. I guess an SL65 HE would be a good place to start.

Nick

btw. the cwa 100 and 50 wont run full power with the pinout you guys postet :( it runs more in some kind of safe mode.

without controler it should be very hard to run it full power.

my friend is looking at the electronics of the pump at the moment. maybe he find a cheaper resolution to let it run full throttle. otherwise i will pin directly on the engines cooper wires :)

?????? I was running it at 22lpm and 0.6bar as described...not sure what is wrong with your setup - or do you feel it is capable of more output?

I'm away from home for work this last month and wont be home for another month - but looking forward to getting back to things.....

I put 12v directly to the Cooper wire and ground and it Powers much more 😎🙈😂

What is the copper wire???

Nick

sorry :) i meant directly on the cooper winding of the engine. i pulled out the electronics and go directly on the pins of the electro motor.

Thats interesting. I order the CWA100 pump now at the german shop you posted originally for 300€/piece.
The shop says its producing 41lpm at 0.6bar or more than 60lpm (thats one liter every second!!!) at zero resistance.

Cant we use a fake PWM controller at that one pin to let it run all the time at 100%? Or maybe they are wrong connected?

follow what I did for the CWA50.
https://mbworld.org/forums/6004400-post191.html

Thanks. What would be the easiest method testing throuput of the pump installed in the car, using cars electrical wiring?
I really need this pump doing 100% without PWM controller.

Nick are you there? Any contacts to Pierburg that can help us?

Never very far away.

I don't know what Napkin's source of information is, but I heard the same thing from SFR Electronics. I've never been able to find the CWA-100 manual that I stumbled across last year, but I did find this little gold-mine:

http://webpages.charter.net/n8nxf/EV...ant%20Pump.pdf

and this from Pierburg, if you have some time...

http://www.pierburg-service.de/ximag...mtzextrako.pdf

Nick

it cant be any easier than what I posted.
3 connections to the pump - 1 through the 1.5k ohm resistor....jobs done.

Ok well the first link helps IF the CWA100 is same to the CWA50 pins.
What connector Type do i need? I want to wire it clean. Are the stock lines thick enought to support the 100W++ powerconsumption ?

So you are 100% sure it then runs at 100% speed? Napkin the Seller of the Pump in Germany said otherwise.
Maybe he can chime in and tell how he wired it, hopefully he just made a mistake.

What was your power consumption when you did that?

Nick

Lots of informationj at the end of this thread:

https://mbworld.org/forums/w211-amg/...w-about-3.html

These are the parts I ordered from my BMW dealer:

1 x 12527549033 Plastic Socket Housing
4 x 61138366245 Rubber sealing grommets
4 x 12527545858 Individual Socket Pins

The 61133333333 part number is not meaningful - it just means some undefined socket.

Here are the parts that I bought (two sets - one for the car, one for the pump test rig)

Attachment 400186

You can probaby crimp the cables to the pins, but I soldered them instead:

Attachment 400187

And here's the assembled connector:

Attachment 400188

Thanks alot. As i write this reply i just received the pump. Its tiny, amazing that this little pump should perform twice as good as the Bosch 010 pump. Wow. Speechless.
Number on the pump is: "7.02500.25.0"

Napkin added the correct BMW connector aswell, however there are 2 contacts missing.
Can anyone confirm that the Pierburg Number "7.02500.25.0" is indeed the CWA100 and not the CWA50? :-)

Regarding the holder/bracket. Any ideas? There is one from the SLS AMG with partnumber: "A1975040146" with EPC picture see here: http://www.mbpartsworld.com/p/__/RET...975040146.html

What do you think?

EDIT: My connector looks different, just for your information guys. But it fits right into my pump.

There's a little nugget of information that I think I missed first time round. If you look at the chart below, that corresponds very closely to what Pierburg say the CWA50 will do, installed in a typical IC system (second system curve). That sounds like its running at full power to me - with a simple 12V conenction to the control pin.


Attachment 400150

I wonder why we would need a 1.5K Ohm resistor at Pin 3?
Can i use other ones? I dont have a 1.5K Ohm here

With 1.5K Ohm at 12-14.4V you almost completely annihilate the current. So i dont get it.
From Nicks document it should be "OK" to use 12V to Pin3 without any resistor.

How can i measure if the pump really works with the power as advertised? Where to measure the power flow? At the fuse for the pump?

Different question, should i upgrade the stock 10A fuse with something higher for the CWA100?

1.5kohm is fine because the controller input impedance will be high.
Its only there for protection against fault conditions.
The exact value isn't important.
Measure the power consumption by measuring the current through either pin 1 or pin 4.
Don't forget to measure the actual voltage across 1 and 4, as the pump itself will pull a battery down slightly.
10A is probably OK for the fuse, but it might be a good idea to upgrade to 20A.

Nick

yes - 1.5kOhm to protect signal electronics in pump
1.5kOhm on 12V will only allow 8milli Amps - more than enough for signal electronics.
Can be any value between 1k and 2kOhms

I usually apply 1-1.5kOhm on any signal Input/outputs.

Hi Folks, i confirmed via MS Autoservice (Part of Pier Burg group), hat my pump is indeed the cwa-100. It was a bit of a problem and i was asked where i got hat pump from.
Anyway , i installed the pump today. Wired Pin3 with a 1KOhm resistor. Pump runs approx. 3sec after the relay switches the power on.

However it only consumes 6.8 to 7 Ampere at 11.6V. 11.6V despite charger beacause with Star Diagnosis, it lets the main cooling fan roar which sucks from the battery pretty hard.
Thats calculated 81.2 Watt. The CWA100 is a pump in the 100Watt Range. According To my information the cwa-100 consumes up To 135Watt. The
CWA-50 consumes typically already 6.5 Amps.... So my conclusion is that it doesnt run at full power. :-(

In the car after a long bleeding session, its hardly a difference noticable against the Bosch 010 pump. It does recover a bit Bettler and quicker from very high IATs though, but not anywhere as close as what i expected.
I want it to run Art 100% but how?

Thanks alot
- Christian-

I think the answer is to use the SFR Electronics pump controller. They say they can get it running at 100%.

The IC system is very sensitive to air in the coolant. Are you confident that you bled it successfully? Does the pump run quietly without any whooshing noises?

The HE is a particular problem, as the inlet is at the bottom, and the outlet is in the bottom half of the HE. An unforgivable mistake by Mercedes, and one of the reasons they say you have to use the vacuum bleeder.

The Mercedes system uses a compressed-air driven vacuum pump, and even that isn't really adequate to evacuate the system. The pressure doesn't go low enough to boil the coolant.

Nick

Hi Nick thanks for your reply,
On the C215 CL55 AMG cars, there is no need for a bleeder since the system auto-bleeds itself according to my information/FSM. It has a bleed line at its "almost" highest point that gets right into the stock overflow tank.
Also the CL55 AMG only has on big HE at the very front of the radiator/condenser sandwhich, NO additional bottom HE like on the W220 S-class or SL R230 series.

I let the pump run for around half an hour via Star Diagnosis and the overflow tank lost quite some amount of liquid and often bubbles appeared. I waited until all that was over, topped the overflow tank with fresh water and that was my bleeding session.

Yes the pump runs veeeeeery quiet, no signs of any gushing or bubbles inside.

While i was working with STAR, i did the minus 10degree Fan mod. That IMHO helps alot getting the IAT cooler during slow city drives. But anyway...

Here are my findings in IAT after the CWA-100 installation on bright daylight:
Yesterday during daylight we had 27°C (80°F) ambient temperatures.

With AC on and set to 22°C (71.2°F) cabin temp, the IATs where:
- City cruising (no boost at all!) : Between 43°C to 47°C (109°F to 116°F)
- Autobahn cruising (around 130-160kph): Between 42°C to 44°C (107°F to 111°F)

With AC off, IATs dropped slightly:
- City cruising (no boost at all!) : Between 40°C to 44°C (104°F to 111°F)
- Autobahn cruising (around 130-160kmh): Between 38°C to 40°C (100°F to 104°F)

So thats AT BEST 20°F above ambient when cruising on the Autobahn with AC OFF.

The funny thing is, that when getting out of the car and touching the large HE, it always had ambient temps. No heat no nothing....

Yesterday night, the Ambient Temp dropped to 16°C (60F) and i was going to get some food. So i cruised a bit outside and inside the city and my IATs remained at 32°C (90°F) and NEVER get any lower. And that was with AC off.

Howerver there are three things that improved over the Bosch 010 pump:

- City IATs where always above 50°C (122°F) before now in summer, but now with the CWA-100 pump they dropped a few degrees. Also i believe the -10° blower mod did help since it runs now when cruising inside cities.

- IAT peak at Autobahn WOT run remains lower. I never saw 60°C (140°F) IAT yesterday, despite doing many 100-200kph++ runs directly after each other with AC disabled.
Before, driving that way, it would have been as high as 68°C (154°F) at such ambient temperatures.

- Recovery time is better. Starting at 39°C (102°F) IAT, doing a WOT pull from 100kph (62mph) to 230kph (144mph) raised IAT to 54°C (129°F). Keeping that speed or reducing it because of traffic, lowers IAT within ~20 seconds back to 39°C (102°F).
Before the CWA-100 pump upgrade it took at least 1-2 minutes (a "felt" value though).

So my question is, why do i have such high IATs? In the W211 AMG forums, people say they always have around 10°F above ambient when cruising and since they are in the USA i assume with AC ON.
Is the C215 bad designed for charge cooling despite having a felt 3 times as large stock HE than the W211?

Have you tried connecting the pump control line to the ignition, so it runs all the time?

The stock pump is switched off when the IAT drops below 35 degC.

Nick

https://mbworld.org/forums/m275-v12-...75-freaks.html

No i haven't and i wont. I'll let a tuner lower the kick-in temperature via tune (thats possible on the 55K) or let it run constantly via tune.

Today i did another bleeding according to FSM and yes there were still bubbles in it.
Basically you take pliers to block hoses, you block the return line and use a transparent line on the bleeding nipple.
Boy i tell you, even though that was on the return side and so after it passed through that bad 3-pass intercooler, you could still use it as a garden hose, that much of a pressure it creates.
10L of water from the bucket (we flushed the system while we were on it) where gone in a few seconds, so quick does that thing flow.
Impressive pump! However we could not fully 100% bleed it, because my battery collapes because that fcking main fan runs all the time at 50% (425Watt!) when activating the pump via Star Diagnosis.

I don't know nearly as much about the 55K as the 600, but it looks like the IC system is better. It shares a header tank, and there's a vent pipe from the IC to the header, which is great. However, there ought to be a vent pipe from the HE as well. Since the HE inlet isn't at the top, there's no way to bleed the HE without a vacuum filler. It's quite likely that the pressure drop across the HE matrix is greater than the head of pressure up the inlet header tank, in which case the air in the latter will be forced out. It will take less than 1 psi to fill the header, but I bet the pressure drop is at least that much. However, there's no chance that all the air will be forced out of the other tanks. Nomatter how you fill it, there will always be air in the HE. The W215/220 600 & 65 are worse, but MB seemed to learn their lesson in the W216/221 - I think they have the fill ports at the top.

Having said all that, I think changing the switch on temp in firmware is definitely the right thing to do.

Nick

Hi Nick, please see the attachments in english. Its a difference like night and day.

I trust the MB FSM. If they say it has to be bleed that way, then it has to.
I could not find anything else at the HE.

CWA 50 coolant pump - lower priced source?
 

There is a lot of info on this thread about Pierburg coolant pumps (BMW uses them). They certainly seem to have significant better flow rates than even the replacement Bosch pump.

I've seen the CWA 50 pump (or BMW part) priced from $250 to over $500.

I've found that these CWA 50 pumps are quite popular with the guys who are developing and driving electric cars. And, I found a source for the pump at a significant discount. I paid $75 plus reasonable shipping costs for a new one I recently bought and just received.

The fellow selling these has more, and is willing to have his e-mail given to those interested in buying (but he is leaving home - US west coast - in a few days and may not be responding to e-mails for a while). IF you are interested in purchasing this significant upgrade for your supercharger coolant, contact me via a direct e-mail and I'll provide his e-mail address.

I'm sure it will be first come first served until his supply is exhausted. State 'Pierburg CWA 50' in the subject, so I don't miss it and delete it without responding!!

My e-mail address is: kittyandgary@verizon.net

does anyone have Pierburg CWA100 2 for sale?

Speedriven Intercoolers w/scorpion intakes
 

For Sale are 1 yr old Speedriven Red Intercoolers with dual Scorpion Intakes. Traded car and removed them. PM me for pricing.

we sell them. just write us an email info@carlicious-parts.com

Pierburg CWA Duty Cycle
 

Greetings from South Africa!

I'm busy upgrading the charge cooling system for my S4 and read this thread with much interest. I received the proper datasheets from MS Motor Services for the CWA pumps, describing the control of the pump duty cycle. I attach two pages here for interest. This also shows that while it is possible to run he pump with +12V on he PWM input the pump isn't designed tomoperate like that and instead of running at 100% duty cycle, it will run at about 97% as indicated by one of the other forum members.

It is probably academic as 97% would be ample I think.

Thanks to Nick's efforts I was able to order a tinyCWA controller from Tecomotive with custom firmware for manual control of my pump speed. The OE pump on this car - a CWA50 - runs at 50% duty cycle normally and reprogramming the algorithm in the ECU is possible but I haven't been able to source the necessary security access code for that.

I'm fabricating a custom heat exchanger locally which has about 2.5 times the surface area of the stock unit, and with the help of the controller I will optimise the pump speed for acceptable control of IAT for our conditions.

I'm currently recording progress here: Audi Club SA - B8 S4 Development.

I got some solid ideas reading this thread - including keeping the stock Pierburg pump as opposed to what was initially suggested by others...great work Nick!:y


https://cimg3.ibsrv.net/gimg/www.mbw...e60ceb74d0.jpg


https://cimg4.ibsrv.net/gimg/www.mbw...a59d3659f5.jpg

Greetings from USA,

I bought a "BMW" pump and I would like to know if it is a CWA50 or CWA100 ? Certainly, it looks like a Pierburg pump.

Can you tell just by looking at the pictures?

The PN# is 11517568594-03 on the pump and the seller told me that the cross-ref number is 1157588885.

Thanks for your valuable input.

looks more like a 100 or 200...the 50 is smaller.

Its definitely not a CWA-50 or CWA-100. Its an engine cooling pump, rather than an intercooler pump. You can tell by the large inlet & outlet diameter. From the configuration of the electrical connector, I'd say it wasn't a Pierburg pump either. It looks like a Siemens/VDO pump; the type that's used on the 3.0 turbo engine. I don't know how you would control that. Can't help much I'm afraid. I'd write down the part numbers and put Google to the test.

Nick

I have the EWP-80 sitting around for almost a year now. Has anyone here ever mounted one in their E55 AMG? I'm curious of you guy's way of how you did it. Since the OEM way of attaching it with the OEM Bosch pump screw is not possible and before I go all out of my way designing and building a mounting adaptor. I hope someone has a solid idea or way of how he did it.

Pictures are always welcome.

I used a drainpipe mounting bracket / downpipe clip.

Nick

Do you have some pictures?

Parallel or Series?
 

Nick,

What did you end up doing - parallel or left the system HE and IC in series?

Thanks,

-Pete

Hi Pete,

I'm still using a series configuration, but I'm thinking about changing that. I use a CWA-100 pump, which is quite high pressure / low flow, but I'd like to go back to the CWA-200, and make the most of the engine radiator HE.

Before I do that though, I'd like to get some better instrumentation in, and at least have an in-line flow meter in series with the IC's. I do like the idea of having a really high flow through the HE (literally off the chart), and getting the thermal resistance down low, but I'd want to be sure that I wasn't reducing the flow to the IC's.

With charge cooling, you really want both elements of the equation to be low - ie: low thermal resistance for both the HE and IC. However, they don't have to be similar, they both simply need to be as low as possible.

The only slight problem with asymmetric shunt cooling is keeping the HE loop as short as possible. Most radiators are cross-flow, but a double pass rad like the E-class 6-cylinder cars might be a good idea. That would keep the feed and return hoses short. However, that radiator flow resistance would be higher, which undermines some of the advantage. The alternative is to stick with the cross-flow rad, and have the feed pipe running across the rad. Problem is there's not much room down there.....

Nick

Quick summary from another thread about how to install pumps in general:


I learned a lot about pumps and cooling systems in the last year or so.
Cooling systems have been developed over a hunderd years, its a problem that's been solved.
However, Mercedes seem to have forgotten all about it with the V12TT, and treated it as a special case, which is very unfortunate.

There are some long-established golden rules for installing (centrifugal) pumps, including:

• The pump inlet pressure should be as high as possible.
• Therefore the pump should be located at the lowest point in the system.
• Locate the pump after the radiator, to keep it as cool as possible.
• Mount the pump horizontally, to minimise bearing end loads.
• To avoid air locks, the outlet should be at the top of the pump, and must not point down.
• The pump's pressure/flow characteristics should be matched to the system resistance curve to achieve best performance.
• Don't run a pump into an excessively high or low resistance, or it will fail quickly.
• Keep air and contaminants out of the system.
• Finally - feed the pump from the BOTTOM of the radiator, to minimise air ingestion.

It's the inlet that should be at the top of the HE (but maybe that would interfere with the LHS engine air intake on the V12TT). Either way, there ought to be a bleed facility at the top of the HE, whether its an automatic bleed to a header tank, or a manual bleed port like the end of the return pipe (near the ABC pump).

But yes, people often see an improvement in the performance of their IC system simply because its been bled properly, rather than because of the modifications they've done. I can't think of any other reason for the popularity of marine circulation pumps and engine cooling pumps.

Regards, Nick

I just stumbled across what might be the ideal HE for a tuned S600 (apart from the S65 HE, that is). My wish list includes:

• Standard production radiator with aluminium matrix, and plastic tanks (so its cheap)
• RHS Oil cooler for auto transmission (so I can run ABC oil through it)
• Drain plug for bleeding
• RHS High inlet
• LHS Low outlet
• Width: 600mm
• Height: 500mm
• Thick: 40mm

The Mercedes W124 diesel auto radiator is very close, but the inlet & outlet are in the extreme corners, so its difficult to clear the grille & hood when they're closed.

The Range Rover Sport 2.7 radiator seems to fit the bill, so this might be my next project. That's 12 litres of HE matrix volume, exactly the same as the engine radiator, and twice the engine capacity, it meets every criterion for me. The frontal area is also very large - even larger than the BMW X3 radiator that I currently use - so it should maximise airflow.

I think I might use the CWA-200 and SFR controller this time round.

Fingers crossed.

http://admin.nissens.dk/FileExplorer..._Product_Large

I heard a VW Golf III HE will fit without modifications in an W211 and has better cooling than our stock one.

Yes, its bigger than the stock S600 HE. I think the dimensions are somewhere in this thread.
Nick

what role does the pump on the drivers side (US model) have ? my SL65 has two pumps :) or is that for something else. The bosch 010 pump is there but there is also another pump part number A0001405385. Also "Nick" I have removed the 010 pump and am in the process of reinstalling it, whats the best method you found to feed it with least air pockets sir?

There's only one IC pump. I don't know what the other pump is, maybe the emissions air injection pump. Bleeding the IC system properly is almost impossible. There's no
easy way to be sure of getting all the air out. I would use a cooling system refill kit, and drive it with a rotary vane vacuum pump (as used on AC systems). They're not expensive. Alternatively, you could use the method given in the other M275 thread

https://mbworld.org/forums/m275-v12-...-included.html

Driver side also has AIRmatic pump next to the fog lights.

The pump on the right is the one I dont know what it is ? Nick?

SL65 doesn't have an airmatic pump, it has an engine-driven tandem pump.

Nick

You legend, thanks bud. I bought the 010 pump and my car had it already :( do you know if the 010 pump is a standard on the Sl65?

I misread it's was about an SL. My bad.

All the turbo and supercharged cars introduced in 2002 started off with the Bosch -002 pump. This later became the -010 pump, which is slightly more powerful. Mercedes gradually changed all their supercharged cars over to the Pierburg CWA-100 a few years later. They're not interchangeable, as the Pierburg has an internal pump controller, while the Bosch is controlled by an external relay. Sorry, I don't know when either of the crossover points were. I think that would be quite hard to find out.

Nick

All good Nick, thanks once again

Are you sure? This would be very good for us CWA-100 users like myself. I still dont know if my CWA-100 runs at 100% or not. I direct wired it with an BMW connector. One person in this thread here or the other said, he removed pump cover and directly attched the wires to the pump without connectors and therefor bypass that PWM singnal stuff. :nix:

It seems like the Pierburg pump is working in an E55 and runs about 3x as much water as the EWP-80.

I'll order a Pierburg which is in the Black Series and give it a try.

Nick , I have my half front end of the car, I also bought a Bosch 010 pump accidentally , I want to run a second HE cooler, a few questions.
1. do you suggest I do this, will it improve cooling?
2. if I did should I run the second pump I bought? if so how far apart?
3. Should I wire them to run on always, or leave the system as it is per factory?
4. is the coolant type the same as the cars motor coolant?

sorry for so much questions. But you seem to be the man to talk to regarding the V12 cooling.

Thanks in advance

The answer is on post 228.

This is all the information that I have:

http://s1.teamlearn.de/QuickPlace/b-...0waterpump.pdf
http://webpages.charter.net/n8nxf/EV...ant%20Pump.pdf
https://mbworld.org/forums/m275-v12-...ml#post6395913

1. Yes, two intercoolers will improve cooling, but I would run them in parallel rather than series, to avoid reducing the flow through the IC's.
2. Yes, running a second pump of the same type is always a good idea, as long as its installed properly, and doesn't cause airlocks.
3. I wouldn't run a CWA-100 always, as it tends to froth the coolant, but the Bosch is OK. Only problem the commutator bushes wear out, so the motor doesn't last very long. The Pierburgs don't have that problem.
4. Yes, the IC uses the same coolant as the engine, though I would run a lower proportion of anti-freeze.

Nick

How do you suggest I run them in parallel , each inter cooler has its own cooler or just run both coolers of a dividing y pipe splitter ? Also about the pumps should I wire them to run continuously as the car is turned on , or leave as factory set up ? Thanks heaps again

This is what a parallel intercooler configuration looks like, schematically:


Attachment 368753


And this is what it looks like in practice. On my car, at least. Mercedes use metal pipes with right-angle T-junctions, which limit flow.


Attachment 368754


You can find the Y-pieces by doing an eBay search for "3/4 Y connector"

Nick

I have two pumps , plus a spare Ai condenser which I've been told is best to use as a cooler for the IC system including the stock unit, however the inlets on the condenser are tiny , would putting a y pipe connector with a reducer work ? One would think it would restrict the flow

It was one of the first things I thought of three years ago, but I wouldn't consider using an AC condenser, as the coolant flow would be too low.


Nick

What about if i air die grind out the pipes from the AC line (tiny piping) to the standard 3/4 piping, would that fix the flow issue ?

https://cimg0.ibsrv.net/gimg/www.mbw...ef33f04b34.jpg

Sorry i forgot to add into the digram the heat exchangers too, but you get the gist of my idea

Yes that would work well. Just make sure you get all the swarf out of the pipes.

Nick

Could anyone chime in with CWA-100 part numbers, please (BMW preferred)

Is BMW E60 - 5 Series = Electric Water pump 11517586925

an intercooler or engine water pump?

The pump in the picture is an engine pump.

Nick

I used Pierburg CWA-50, -100 and -200 pumps on my car, largely because I couldn't find any CWA-400. In fact, I didn't think there were any, but I found this on ebay.de last night:

http://www.ebay.de/itm/Tecomotive-ti...IAAOSw14xWLrEw

I used to think they were superseded in the BMW line-up by the VDO pump, but that's not the case. They do exist and they are commercially available. They use different electrical connectors, but I don't think that's a problem.

I'm just trying to figure out how to reliably identify one, and what the part numbers should be. You can't believe everything you read, especially from someone selling car parts. I think the best guide is that this pump was used on the BMW 4 cylinder 2 litre turbo engine, used in the 328i, 428i, 528i and similar models. I'm not 100% sure yet, but I believe the BMW part numbers are:
11517596763
11517604027

I believe this is the Pierburg part number for the CWA-400. I've seen other P/N's used, but I don't think they're right. The -66- dash number appears to be the most common, but there a dozen others.

7.03665.66.0

http://www.tecomotive.com/ebay/Pumpen_Uebersicht.jpg

The CWA50 and 100 look alike. Some have quick-fit ports, and some need hose clips. You can't distinguish them by appearance.

The CWA200 has 4 bolts holding the impeller housing, and two bolts holding the electronics housings. The CWA400 doesn't.

The CWA50 & 100 use the same connector, but otherwise they all use different connectors.

Nick

To complete the picture, this is the Siemens / VDO / Continental pump that's used on the 3.0 litre turbo gas engine in the recent 135i, 335i, 535i, 640i and similar.

Common BMW part numbers are:
11517563659
11517588885
11517632426

I believe two VDO part numbers are
A2C59514607
A2C53326031

Since this is used on the bigger engine, I presume this is even more powerful than the CWA-400 pump, but there's not much information about it.

http://ows-cdn.tecdoc.net/vdo/home.j...37ef248a4c0a90

It looks like the connector is the same as the CWA400, but hard to be sure.
Hopefully the control signals are the same, otherwise not much interest in the context of this thread.
I've just bought one to play with, so will find out soon.

Nick

http://i.ebayimg.com/00/s/NjAwWDgwMA...WCO1D/$_57.JPG

A datasheet at last - hurrah!

http://www.easycarparts.nl/Controls/...9514607_DS.PDF

Operating point 0.80 bar at 9000 litres per hour = 150 l/min

So about the same as the CWA400.

Nick

Excellent Info.Thank you very much,especially to Nick.

Now my Pump seems to be dead, checked Fuse and it was o.k.

Want to get the CWA100 now. Is the latest info still without a control module Ground1 ,Battery4, 1.5 kOhm resistor3 and it works with (nearly) 100% power? What are the connecting parts and mounting brackets? Anywhere good prices for the pump, and is the cwa100 worth the price in daily driving over the cwa50 which is also used in many VW/Audi applications. Many thanks!

Send me a PM.

Does anybody know how much Volts and Amps our Cars use for powering OEM intercooler pump? 12v 14v or 24v?

We have a 12V battery on board.

I bought EMP WP29 pump it operates 12v and 24v I want to install it on my 07 cl600 with 24v is there anyway to attach + - wiring on to battery and amplify it to 24V?

2 x 12V and you got your 24 V. Problem will be that the rest of your car won't work anymore bc it blows up ;) but hey you got your 24V pump going :)

But on a serious note: I'm sure the specs read it can operate between 12-24V and not 12 or 24 V.

on 12v it flow 25 gpm and 29 on 24v thought i could squeeze 4 more :D has anybody here ever installed EMP pump on their mercs?

I believe the WP29 comes in either 12v or 24v versions; they're not multi-voltage.

Have you downloaded the manual? Lingenfelter are a lot of help with these pumps.

http://emp-corp.com/support/marketin...MarketingID=99
http://emp-corp.com/support/marketin...arketingID=107

Nick

I am about to buy this pump so it either 24v or 12v? on this one it says 24v on the back so I wont be able to make this pump work on my car? its w216 cl600:nix:
This is what I found on CTSowners forum
Quote : Stewart EMP WP29 Water Pump

Selling a used Stewart EMP WP29 Water Pump, 24v version with 1"inlet/outlet. Used it with 12v for about a month and it works great! Need to install a G-force intake which routes exactly where this pump is mounted and can't find anywhere else to easily mount, so my loss is your gain. Looking for $310 shipped, also comes with brand new Deutch connector. These retail for almost $500
Will be taken out today so i can post pics later on.

What would be the difference between these two? They have different serial and part numbers but they are both 12v-24v

Every car has a different serial number aka VIN. And Two different products have two different part numbers. It's like comparing Snickers with M&Ms. :)

yes man you are so righ LOL I KNOW THAT THEY ARE LIKE SNICKERS AND M&MS OR MAY BE MARS AND BOUNTY THATS WHY I ASKED THIS QUESTION IF ANYBODY KNEW WHAT IS THE DIFFERENCE BETWEEN THEM AND YOU DEFENETLY DO NOT KNOW IT OR DO YOU?


Soo Thats the purpose of the question these two products look the same have same name (EMP WP29) both 24v but have different part numbers so there is must be some difference between them? may be one is better than the other?

Don't yell! You're question fits right into the nonsense category.

If you know why different products have different serials and part numbers then what's the real question? One apparently runs with 12 the other with 24V. Not more not less without doing the homework for you. Read the manual(s) and you will find your answer. We're not here to read and comprehend it for you. That's your job.

And if you can't take a joke, too bad for you.

I dunno. They both say 24V pumps. Probably different revisions of the exact same product. I have that pump in 24V...got a steal on it because everybody's after the 12V ones. Haven't gotten a power supply rigged up for it yet, but I plan on installing it eventually. Best intercooler pump money can buy in my opinion (the 24V version has a lot higher pressure capability...it's just a matter of figure out how to power it, lol.)

First off all they are same products Both EMP WP29 and if YOU would see the pictures attached to my question 24V are written on back of both.

Second Ok I understand if they have different serial numbers its like when two CL600 2007 for example having different VIN codes .( same products different Numbers)

BUT in mechanical engineering world same parts have same part numbers and these do not have same part numbers so there is gotta be some difference between them. I read all the PDFs and documents and I am currently waiting for emp-corp.com customer account approval to get access for more pdf documents wich you can look up with part numbers . There is only install manuals online. AND PLEASE you have really bad sense of humor and I am going thew nicotine withdrawal its my second day and dont get on my nerves please have a good day .

these are like snickers and m&ms :y:y:y:y:y:y:y:y:y:y:y:y:y:y:y:y:y:y:y:hammer: :y:y