SICAMG

Just ordered the clamp and will install and test soon. Then go a couple lb's more boost and see how she responds. Thanks for all the help !!

SICAMG

So it is back again. No fuel boost/map under boost. Was going to do a data log one more time before I pull the engine for more mods and she went real lean and kicked on the fail safe. It has not happened since May of this year and I have the map clamp on as well and was no where near max boost so that is not the issue.
It appears that the ECU is not seeing or miss interpreting the boost reading and it just stays in a naturally asperated? AFR mode.
Stop car, shut off ,restart all is well until next time !!!! And it has been 6 months since the last time !! The ECU is seeing something it does not like ….or is not responding to what it see's {boost}.

sjc246

Definitely sounds like a boost spike. This is most likely ambient temperature related. You will see slightly higher boost when it's colder out (denser air). When I had this problem, it only happened on cool days.

What voltage did you set the map clamp at? I believe the general consensus is that you shouldn't go any higher than 4.5 volts (and that seems to be working for me). You should double check with a different (known good) multimeter to make sure it is adjusted correctly (and, of course, make sure it's installed correctly). While unlikely, I suppose there is a possibility that the setting could have shifted even if it was set properly to begin with (heat cycling can do strange things to electronics).

Do you happen to have access to an oscilloscope? If so, I would connect it so that one channel is pre map clamp signal and one is post map clamp. You should be able to see the spikes and dips in the boost signal that you can't see with basic OBD2 data logging and you can verify that the post clamp signal isn't exceeding your set value. See post number 21 in this thread as an example (he has the map signal clamped at 4.6V with a piggyback): https://mbworld.org/forums/w211-amg/...p-running.html.

SICAMG

I am at 4.5 volts. Tonight I will drop it to 4.25 and drive from there. Funny you mention temps. we have seen a drop from 75* average to 29* this morning. So yes can be temp related but boost numbers still seem to be consistant around 16lb's.
Again this is such a random thing....last time was around six months ago.
Also I do not have a scope so will rely on meter readings.

sjc246

Definitely worth a try. I believe 4.25 volts is approximately 16.5 pounds of boost so definitely don't go any lower than that or you could go lean due to not seeing the actual boost. Perhaps try 4.3 to be a bit safer. Definitely keep an eye on your wideband for the first few runs to make sure you don't go any leaner than normal.

SICAMG

Hey ...You know what I was just thinking before I go in and make some adjustments....What if it is the LOW side for the vacuum that is causing this.
I am on a 3rd gear down shift roll every time so maybe the vacuum reading is off the grid and is causing this ???
It will never throw the MIL light but will have the implausable map sensor code in memory.

Just a thought as I will tune that out as well for Saturday.

sjc246

I don't think so. I don't know of any mods that will increase vacuum and you would need to make a lot more than stock to set an error for the sensor. What, specifically, is the fault code? Assuming it's caused by over-boost, the fault should be something along the lines of "MAP sensor short to positive" (although it may have different text if you're using a generic OBD reader instead of an SDS or at least something with manufacturer specific data). If the fault were due to too much vacuum, the fault would indicate "short to ground." It does it in third gear because boost is higher in third than in first or second (in theory, it should also do it in 4th and 5th if you were going fast enough).

SICAMG

Ok after poking around with the clamp I have found that 17lb's of boost is 3.6 volts. Why so low I do not know. So it is now set at 3.4 for 16 lb's. Will test and let you know.

sjc246

How did you determine that?

SICAMG

This is what I did . Put a pressure gauge, Mity Vac, on the map sensor,have a spare sensor so just plugged it in,brought gauge to 17 lb,got a reading of 3.6 volts on the meter.
I would assume that when the ECU see's 17lbs the map is reading 3.6 volt's going by this mock up test ????. This does not make sense and seems like a very low number.

This is somehow not the case. The car would go to limp mode instantly as soon as you jump on the throttle and make boost,fail safe kicking on every time.
BUT........ at least I am able to delibratly make this happen and know for sure that this is the culprit as far as lean/ limp mode goes and it does not throw a cell light any time either like normal
So I set the clamp back to 4.5 volts and everything was fine. Today I will knock it back a tenth and try it again.

So again I am delberatly putting 17 lb's of pressure/boost on the sensor, seen on STAR, and seeing 3.6 volts coming out on meter for a voltage reading. Strange isn't it?

SICAMG

OK just came back. Did some major data logging. Found that boost is up to 17.5lb in 35-40* temps !!! NEVER have seen it that high but never ran this pulley combo in the cold as it is normally 15.5-16 in 75-80 temps.
So I have clamp set at 4.35 volts and also had to reset the fail safe numbers because that was off the grid as well.
So long story short all seems to be well at 17ib, and clamped at 4.35 volts. Hope it stays !! LOL

sjc246

That's what I was afraid of. The cheap gauges on something like a hand-held pump are almost always VERY inaccurate (if they work at all). It may be off by 5-10 psi (or more sometimes). That kind of test is useless unless you can find a quality gauge that reads specifically in that pressure range (for example, even a good quality shop air pressure gauge may not be accurate at that low a pressure if it reads up to 150 or 200 psi).

Assuming this is, in fact, a 2.5 bar MAP sensor (as many people have claimed), it should read from approximately -14.7 psi (29.97in HG) at 0V up to approximately 22.08 psi at 5V. From that, assuming a linear response, we can extrapolate that 2V = 0 psi, 2.5V = 3.68 psi, 3V = 7.36 psi, 3.5V = 11.04 psi, 4V = 14.72 psi and 4.5V = 18.4 psi. However, the sensor may go slightly beyond the -14.7 to 22.08 range so the values are probably at least slightly different.

That's good. I hope it works out for you. Please update the thread if you find out anything else.

SICAMG

Well back at it again and although the car is running perfect,today it was 35* so I thought I would do a few more data logs before the engine and trans comes out for some work. Sure enough...limp mode. Check data logs and boost spiked to 18.8 lb's !!!
Did this 3 times.
Got busy with other stuff so the temps where up 22* more by the time I tried testing again. Humidity from 51% to 90%
I did turn the map clamp down a little which really should have not did anything since it was fine before but thought maby a hard spike was pushing past the set limit for a split second.
Logs showed around 17.5 lb"s on 4 pulls and there was no issues. So.....this is really a cold temp issue bringing on more boost but the clamp should take care of this. Don't understand what's going on ?? May have to grab a 77 pulley for the winter temps if need be.

sjc246

This is interesting. I just had the same problem again now that it's cold here (about 45 degrees in 3rd gear) with my clamp at 4.5v so obviously that's too high or perhaps the spikes are getting past the clamp. OBD logs show over 17.5psi. What voltage are you clamping at now? Is afr still good or is it running leaner with the clamp set so low? (If you clamp below the average peak boost, ignoring spikes, it will go lean.) When I get a chance, I will lower mine and log before and after clamp values with a scope to see what's actually happening. OBD logging is useless for this because the resolution is far far too low. Hopefully I can also get enough data to determine an accurate voltage to boost chart.

SICAMG

I did not put a meter on it but when testing the "turn ratio" to voltage output I should be at 4.4 volts.
Temps will be low 40's Monday morning so will try it on the way to work. The spikes have to be getting past the clamp some how...maybe it can not react fast enough ???? Who knows........
I also sent an email to Tony at Race IQ to see if we could tune the boost limit out some how but doubt it.
And why I saw such at high boost reading is strange as well !! This is for sure a very cold temperature condition creating more boost.

sjc246

A voltage clamp is not a computer that processes information: the clamping is done directly by hardware so it doesn't really need to "react." I will verify this with an oscilloscope once it's cold enough and I have time (possibly Thursday). I suspect the limit is slightly below 4.5 volts.

While it's certainly "possible" to "tune" this out, I seriously doubt that any US tuners (if any at all) would have a clue how to do it (although the MB/Bosch engineers in Germany certainly could do it in a second). This is not a "boost limit" per say, but rather, part of the vehicle's self-diagnostic system. The ME (ECU) always monitors all inputs for abnormalities so that it can detect a possible problem and try to find a way keep things working despite it. In this case, the ME knows that it should never see a signal over 4.4? volts from the MAP sensor because there is no way to get nearly that much boost (unless the car is modified). As such, if it does see such a high voltage, it assumes that the sensor must have failed with an internal short circuit to the 5V reference voltage so it immediately starts ignoring the MAP sensor and using a "limp home" map using only other sensors to control fuel delivery (mainly the throttle position). Because our cars need a LOT more fuel at full throttle than with stock boost levels, they runVERY lean when this happens. (If there actually was a short to positive and the ME did not react to it by ignoring the bad MAP signal, it would be commanding WOT/full boost levels of fuel at all times which would likely just stall the engine immediately, but could wash down the cylinder walls and cause engine and catalyst damage if it managed to keep running for any amount of time.)

SICAMG

I really want this wrapped up before the engine is out next week. This way I know it is all perfect so if I do have an issue with the motor build I know I did not "create" this problem if you know what I mean.

SICAMG

OK raining here so no new info but...….. riddle me this. If my/your clamp is set at 4.5 volts and we see 18lb's on the data log coming from the obd port than it, the ECU, is seeing the 18 lb's because it is reading it.
The only way to really do this correctly is to hook up a meter with peak recall , see what the map reading is AT boost and adjust from there.
Again even though it is set at 4.5 I am still reading 18lb's and that 18 could be at say 4.1 volts.

drothgeb

You try testing the output of the clamp to make sure it works? Maybe hook a 9v battery to your MAP sensor leads on the clamp and check what the clamp output to the ECU is with a meter to make sure it's actually clamping?

I don't actually know what the sensor input of the clamp is rated for but if a 9V is too high you could rig up 3x AAs in series for a 4.5V test rig.

sjc246

Thanks for the suggestion, but that shouldn't be necessary. The clamp is "set" at a certain voltage by connecting the input to the 5V reference power supply from the ME and adjusting the knob while watching the output with a voltmeter. You turn the knob until the output voltage is as low as you desire. I suspect the limit is just slightly below 4.5V and I'm sure we can find it with some trial and error. I will try 4.4V and report back when I get time to try it (needs to be cold and not wet).

Using a meter with peak recall will not work. A very good meter may be better than OBD logging, but not by very much. What you are trying to see is a spike for a fraction of a second. Multimeters display (and record) an average of the value over at least a few tenths of a second so you will never see that spike. With OBD logging, you only see snapshots of the reading (a few per second at most) while the ME is reading the value nearly continuously (in theory, you may see the spike, but it's very unlikely). The only way to see it with any kind of accuracy is with an oscilloscope or an analog data logger with a very high refresh rate. 18psi isn't likely to be 4.1V on a 2.5 bar sensor. It should be slightly under 4.5V (a typical 2.5 bar sensor should output 4.5V at around 18.4psi). This sensor may deviate slightly from "typical" but probably not by very much. I'm hoping to log some points with a scope to figure out exactly what it is.

SICAMG

OK so if you see 18psi that means the ECU saw 18psi because you are logging it. No??
If you see 20lb's than the ECU saw that as well and so on and so forth. So the map reading in voltage for that psi is still "seen"
When you did the math ...4.5 volts "should" = 18.4 psi. Our cut off as we are told is 17lb"s...long before 4.5 volts.
Maybe I am reading into this to much.LOL

drothgeb

Okay, that's what I was getting at, using a known voltage source to verify the output of the clamp is actually clamping. You guys are moving it off of the 5V reference to the MAP sensor and onto the MAP sensor output once set though correct?

Yeah if you are logging 18PSI through obdii or 20 or what ever, the associated voltage output from the map sensor is being seen by the ECU.

SICAMG

Drothgeb the unit is adjusted with a meter on the output side. You adjust your voltage to where you want it ...say 4.0V and than reconnect everything. The ECU will now be "input capped" and only see 4.0V because that is all that is being sent out from the MAP sensor clamp . You can set it where ever you want . 2.5V - 5 V ...what ever you need.
When I did mine I used a pressure gauge,pumped it up to 18lb saw what the normal voltage was,and clamped below that. However as I found out this was not very accurate.

sjc246

Yes. Assuming that the logging application is reading and converting the data correctly. This isn't always the case, but I haven't heard of any such issues in this case, so it is probably accurate. The problem is that you only see some of the values that the ME (ECU) sees (a very very small percentage). As a simplified example, lets say the ME sees ....17.0, 17.3, 16.8, 18.2, 17.2, 19.4, 17.3, 16.9, 17.1, 17.0, 17.2.....and many more, you may only see the value 17.2 but none of the other values because the OBD protocol only supports a few readings per second (depends a lot on the specific car and tool being used), while the ME sees hundreds if not thousands of values in that same time (and the OBD limit is a total for all data, so the more data you log, the lower the resolution of each one)

I have heard 17psi, 17.5psi, over 18psi, etc all supposed to be the limit. We have also been told that the limit is 4.5V. Most of this information is just an approximation or guess; it is not clear if any of it comes from actual knowledge of the coding or any significant amount of trial and error testing. I suspect that to know the answer for sure we will need to do the trial and error.

SICAMG

As soon as the rain stops and temps go lower I will test more. I only have till this Saturday because I want this engine out and done before the next 1/2 mile event.