Probably get fuel from steel tanks and hence will be some rust materials. Any in ground tank and even above ground tanks will rust on the inside eventually. In ground tanks that may be 50 yrs old or more almost guaranteed.

 

Hmm, thus far when scoping the cam position sensors (I call them camshaft Hall sensor as this is what WIS refers to), I've seen the nice square wave. Just to validate that the weird spikes were caused by the current clamp, I re-ran the test without the current probe attached. Strangely, the spikes are still there. That's when parenting duties kicked in (my 8-year old asking questions about here homework), so I've yet to search under the hood where my nice square waves went...


====================


Its design is similar. Mine had a MAN filter element inside. There is an interference fit over the exit pipe, just as yours. The fit is very very snug, I don't expect much leakage through this bit. But, in GDI engine design, clean fuel is of utmost importance. Small particles tend to destroy the injectors, and the word is still out on my fuel pressure failure..

I live in The Netherlands, most of our country is below seawater level. There won't be many fuel storage tanks at petrol stations that are above sea-level. Luckily we have very good flood control, otherwise half the country would constantly be flooded.

Anyway, off the find my missing squares :-)

 

Dont ignore that front end knock. Most likely a ball joint (tie rod end type connection) that can really bite ya if it goes.

 

Thanks, I'm definitely planning on fixing it. Initially I suspected the drop links, however replacing these did not solve the issue. I'll check the ball joints with a crow-bar, see what I can find.

 

Finally got round to check the signal coming from the cam position sensor. Luckily, it's there, it was just a bad probe pin on the connector. Some wiggling and this showed up:



So yes indeed, it's okay to use the current probe with a voltage probe on a common ground.

Well anyways, not much left. Fuel supply to the HPFP is okay. Electrical signal is okay. Timing seems fine, as far as possible to match to any known-good trace (none). Camshaft lobes seems to be fixed in place. Injectors are not leaking. Guess it is time for another HPFP.

 

Ok,that is a comforting news. Just make sure you prime the LP fuel filter well so that no air bubbles enter the HPFP.
The HPFP uses fuel for lubrication of the internal mini piston. So running it dry even for a few seconds is a NO NO
I seen some people "drain" the fuel rail by running engine till it dies out of fuel...a big NO NO too for our type of HPFP

 

I will have to take a look at the WIS again, but I didn't see a procedure for priming the fuel filter. Do you prime it by activating the LP fuel pump (using XENTRY or the LAUNCH Creader tool) without starting the engine?

 

Sure, you can activate the LP pump using both. With the launch software, you cannot deactivate it, so it runs for 30s and stops automatically. With Xentry, you are supposedly able to stop the pump. It's not really possible to properly bleed the system; after removing the HPFP you can prime the pump and flow some fuel through the Schrader valve or by leaving one of the rail lines disconnected and bleed from that point.

 

I just reviewed the WIS document (attached) and there is no mention of priming the filter. Perhaps that is only for the SL?

 

No need to expect the WIS to explain everything, sometimes they miss stuff.
On certain engines disconnecting battery is a must to replace fuel filter ( in the fuel tank ) and on my engine it does not stated so, this VERY important safety step.

HPFP uses a normally open QV valve. So a 5 BAR from LP fuel pump will pushes enough fuel to "flood" the mini piston region ( the one we DO NOT WANT to be dry ) and be safe
to crank the engine.

Baby Launch weakness is that, 30 seconds LP fuel pump running without being able to stop it. Dumb-azz program they wrote. Is this still the case with current firmware ?
I no more try to check baby launch LP fuel pump manual activation using its latest firmware, I simply use my Xentry.

In Autel and Xentry you can run and KILL the LP fuel pump manual actuation anytime.
Still it is OK using the Launch though, if only for fuel priming and system is tight with no leak.
Do it twice to 3 times at least , bleed slowly each time the fuel....till no more air bubbles in fuel when you press the schrader valve.

I just replaced my fuel hose at engine side, the one with stainless steel braided jacket, I failed to pull out the blue plastic fuel hose one at the fuel tank side though.
Need to drop the fuel tank at least 2 inch. So next time I will re visit the blue plastic hose.
So I have to prime and filled back the evacuated fuel from the hoses + pipe.....as to not score my HPFP piston.

Too see our HPFP fuel pump internal


Look at the super fine piston at gasoline pumping region. Imagine that piston being dry of fuel..............
200BAR is a lot to ask from ring-less piston.

 

@S-Prihadi still you would think something as critical as this would have been corrected. But, I am using BenzNinja’s installation of WIS which is the latest but still four years old. (Apparently the latest stand-alone version was 2020.)

 

Well, MB I believe would expect techy to have basic knowledge. That is who the WIS is written for and not a DIY with zero experience.
At the least the WIS explained well on how to tightened the two bolts of HPFP, 180 degrees turn one after the other bolts as to provide slow accurate straight line
pressure to the piston engine oil bath side ( pressure point to the camshaft lobe ) and to make sure the camshaft fuel lobe is at the flat valley and not at the peak hill.
Careless people can break the HPFP lower side of the mini piston, the engine oil bath side if they do not install it carefully.

You want to see WIS dumb azz mistake ?
I pointed out on my post here WIS mistake, where the dumb-azz writer does not know our M276.8** 3.0 Turbo uses 4 lobes fuel camshaft and not 3 lobes like M276.9 3.5NA engines, as such
we M276.8 3.0 Turbo would not find the cam fuel lobe flat valley if we follow WIS ........

https://mbworld.org/forums/e-class-w...l-warning.html

 

So, to be clear, activate the fuel pump using XENTRY to purge air from the fuel line. Is 30 seconds too long or too short? And, what about the Schrader valve on the fuel rail?

My fuel filter may be different than yours. My tank is NOT a saddle tank, but one that sits high over the rear axle.



I'm wondering, based on the location of the filter, whether it is self-purging. And, which line is the feed from the fuel pump to the filter and then from the filter to the HPFP? Is the larger connections from the LP pump and the smaller (bottom) connection to the HPFP?






And what is the purpose of this metal thing? Is a relief valve of some sort?

 

30 seconds is actually too long, for me. I like 5-10 seconds manual actuation of the fuel pump. But I do it more than 3 times.

Your fuel filter is different to mine....
Yours does not have the venturi/jet system, because your tank is not a two zone saddle tank.
It seem yours does not need fuel level float at the fuel filter too, as your tank is a single zone tank.

Your fuel filter output port is at the fuel filter bottom ?
So that valve you arrow red must be the equivalent of my fuel filter pressure regulator , also at the bottom of the fuel filter.. That is approx 7.5BAR if on my fuel filter.
But on my fuel filter being a venturi/jet type to return excess fuel back to the RIGHT side saddle tank to the bowl of the fuel filter PUMP, at that valve is where
the venturi/jet small manifold is.

 

Dont ignore that front end knock. Most likely a ball joint (tie rod end type connection) that can really bite ya if it goes.

 

Not much of an update this time. I've replaced the HPFP with the one I only used for ~1 minute from new. Flushed the fuel system with 2 liters of fuel by running the low pressure pump for 30s and bleeding from the high pressure rail.

Then cranked the engine with the injectors disconnected. The pressure gets to 200bar, however, it's taking >10s to reach this pressure. Then I installed the inlet manifold etc, and started the engine. The pressure then rapidly drops down to ~6 bar:


(the rise halfway in the graph, sloping towards ~65 bar is on engine shutoff, when pressure in the rail is increasing due to probably heat soak)

As you can see by the irregular rate of decline, the pump is not having a good time. It's trying, but it's not even providing enough power to provide a steady 200bar at idle.

So, by swapping the pump, the issue remains. I replaced the fuel filter, bled the system, checked the low pressure flow etc. I replaced the ECU, checked the timing, checked the timing signals from the cam(s) checked the actual power consumption of the quantity valve, it all checks out okay. What on earth could wreak so much havoc in fuel pressure while all these things check out okay?

 

Have you actually established where the 3 fuel lobes are in respect to camshaft degrees ?
I thought you wanted to calculate the degree, to make sure the timing of the QV of HPFP pump is proper at fuel lobes peak and valley.

You been focusing on the OUTPUT.
What I mean by OUTPUT is, the ECM decision/calculation as OUTPUT when to open and close the QV valve to build up that pressure.
You have not completely confirm the signals INPUTS, plural.
ECM's output is only good if the input is good.

You have the WIS I gave you on the timing , between CRANKshaft sensor signals and CAMshaft signals.
Are they good at all times ?

The fact that you engine once run well for a few hundred miles on the 2nd HP fuel pump , it showed your ECM and HP fuel pump is OK at that point in time.
So mechanically the HP fuel pump was OK. INPUTs were OK, decision/calculation of ECM to produce the OUTPUT were OK.

I still strongly believe this is signal issue and not a mechanical issue.
So go and trace the INPUTS, and yes, 2 channel scope is not enough a hardware to track this down with ease and low memory capacity of the scope is not helping too.

Perhaps you now best send your car to a professional diagnostician like Diagnose Dan or someone as good or better than him, with all the test instruments able to
perform the signals diagnostic.

We DIY dudes has a limit, either by knowledge, experience or test instruments.
But at least we try and we learnt something along the way.

Can I do my own 4 wheel alignment and RoadForce Balancing ? Nope, unless I buy the Hunter complete system , which I wont

 

Hi S-Prihadi, thanks for another input from your side, keeps the process in my head going.. I did get the lobe position and did a rough check, which seems to check out ok. Eventually I realised the hint in WIS where the lobe is at a valley at the 40° crankshaft mark. That of course gives a nice deadline.

I fully get what you're saying. Ideally I would plot at a minimum both inlet cam timings, crank position sensor and quantity valve current. With 2 channels, I have to find a single baseline (probably CPS) and get multiple traces with another sensor to build the complete picture. I'll try to capture this, it'll be difficult as the engine runs very rough and only for a few seconds currently. So the timing information in a capture lacks a constant rpm to compare to.

I also tend to agree to your point in the HP pump being mechanically OK. So it should really boil down to timing of the QV versus lobe position. Either one (or both) being off would lead to this behaviour. The changes of (identical) mechanical issues in both HP pumps are very slim indeed. If only there was a straightforward way to gauge lobe position vs QV control with a scope, that would immediately validate this scenario. The closest to this is cam position from the tone wheel vs QV control.

Actually, using the string method, it's pretty forward to get a proper alignment check on camber and toe settings at home, as long as you have a very flat driveway (or garage floor) as a baseline. Try it, it's really easy and good fun and can be way more accurate than a Hunter system, unless that Hunter system has been calibrated recently and is operated by a well trained employee. I've been doing my own alignments on my Lotus for years this way, and I think this method is still used in F1 (and most races) today. Check https://www.listerbrosracing.com/pos...heel-alignment for example!

Castor adjustment needs a bit more equipment, but I'm not sure if this is adjustable at all on the W212 chassis.

 

Figure out if your Crankshaft and CAMshaft position sensor is still original/genuine or someone has replaced them with aftermarket ?

Of late I been seeing techy ( video ) getting aftermarket CRANKshaft sensor which at a glance has signal near identical to genuine one but it is not and
engine can run, but at times engine goes banana.

Sometimes the techy does not know the previous repair history of the car to that level, so that sure made techy spend more time figuring out...poor soul.


Manual wheel alignment not for me. I rather use Hunter 4 wheel alignment as I need all caster/camber data too, to see my suspension wear and tear with 5 minute degree accuracy.
Its cheap here in Jakarta for such service only US$20 at my favorite shop. I do wheel alignment and balancing like per 5,000KM or less when I hit pot holes

 

I checked the basic here, I have WIS running locally as well so I can access document GF.07.10-P-1006MMQ (this is the one you're referring to I think?). As far as I can see, the inlet cam / crank signals don't line up like shown in the diagram. Also, the diagram doesn't fully make sense, as it's showing a Cyl#1 TDC after 2 'dips' on the CPS, line 'a' below (inlet cam is line 5).


The text refers to line a as TD Cyl#1. If you compare this to document AR05.20-P-6010MM, basic position of camshaft at 53°(past Cyl #1 TDC) check, you'll find that this doesn't line up with the diagram above. Further, in video , you'll find his trace doesn't match the WIS document as well.

So first, I'm trying to find the correct baseline for CPS signal versus Crank Angle versus timing markings on the crank pulley. As the fuel pump is either on a flat bit of the cam lobe, or 180° opposite of that, I should be able to match the CPS counts to the lobe position of the fuel pump. Idea here is capture CPS versus QV Current on my scope and validate this timing.

 

The problem is the VVT / phasors in action.
Them 4 units can do up to 40 degrees shift/adjustment, depending on COLD or WARM start, assuming we want only idle RPM waveforms.

I have scoped my CAMshafts signals but not together with CRANKshaft signal.


This is VVT monitoring using Xentry.

The RPM of the 120 second data, engine START from COLD.




....
Intake CAMshaft when cold start :




======

Exhaust CAMshaft when COLD start


.



==============


3 minutes into idling after COLD start and the spikes up//down you see is I revved up the engine a bit, the rest is idling.





.....


My camshaft adaptations data :




...

I will try in a day or two to capture CRANKshaft signal along with Intake and Exhaust CAMshaft and HPFP QV actuation signal. Total 4 channels.
But I can't capture more than 30 seconds I think, if I want high resolution and 4 channels together.
Let see the result soon.

 

Hmm, I'm aware of the VVT influence, that's why when scoping the intake cams, I'm leaving the VVT solenoids disconnect to prevent them from messing with the timing. In the YT video I posted, the capture is during cranking, before ignition. I'm assuming the VVT system to be inactive during that time, but I cannot find info to validate this assumption.

Regarding your cold start excel graphs; I can see the setpoint remains stable in the first bit after startup. The actual is deviating a little from the setpoint, but within a small margin (2°). These deviations are smaller than the CPS can indicate (that's only showing a 3° timing interval, expressed on a full camshaft rotation). So there is some interpolation going on the the ECU to indicate these small deviations. And these deviations are partially due to chain slack, small accelerations / decelerations in the cam due to the valve springs and deviations in crank rotation speed due to the various compression / ignition strokes slightly impacting crank speed.

Would you agree that the cam position during cranking and first 10s of idling can be assumed as being not impacted by VVT?

 

The reply of BLUE
Well, when you scope for good known waveform to keep, you need a proper running engine waveform. But we can also do 2nd time with the VVT all disabled, for data sake.
The fuel lobe peak will stay the same spot no matter what VVT sprocket does.


The reply ofGREEN
Our VVT is a full dynamic positioning type not 2 stage version of oldie VVT.
Its challenging using engine "LEAKY" oil pressure alone forcing from left and right side of the VVT vanes to maintain the degree accuracy, and this oil is PWM controlled by CAM magnets manipulating mechanical oil solenoid.
So when engine is cold, shaky shaky value is expected as the tension held by the oil is not like stiff hydraulic cylinders/actuators.
Sure the chains and others wont be zero tolerance, so the ECM always set the adaptation value or compensation value. That value is equal to chain stretch...sort of.


The reply of ORANGE
You did not look carefully at the excel capture I posted, the RPM. For exhaust, as soon as engine started it goes from -25 degree to -5ish degree.
865/120 = 1 second is 7.2 data points. So in less than 3-4 seconds as soon as engine get to its cold idle speed, the exhaust VVT already completed its action.





When I scope the waveform I will do 3 types of captures :

- CRANK ONLY NO START, using Xentry Compression Test, so I do not cause DTC unplugging anything.
- COLD Start from dead engine to perhaps 20-30 ish seconds into idling as per maximum scope memory shall allow.
- WARM engine running/idling

Let see what is the ECM strategy will be for all these 3 parameters.

 

Okay, I done the scoping for my database and yours.

01. 912MB, this is the link below, only 2 weeks storage from today and they will delete.
https://file.io/hhutKuaoV8Xj

Ignore the month typo, should be APRIL and not MARCH hahahaha




.
02. If you want to use Pico DSP/Filter, keep 70khz, that way square waveform details are not lost at the edges. Or do not need filter at all is still okey.
Noise only come from HPFP QV only, that's it.

03. For Compression test 1 & 2, I can not log the VVT on Xentry because I am in Compression Mode page.

.
04. Where you see Xentry CSV file, that is VVT monitoring.


Where you see 2A , 2B, and 2C of Pico file and Xentry CSV, that means when I am scoping, I also uses Xentry VVT monitoring.
Xentry VVT monitoring is always 120 seconds worth, otherwise I can't save it.
Pico I set at 100 millisecond , so 1 page is worth 1 second and maximum 32 pages per file = 32 seconds worth of logging.

The sequence of files number 1 to 5 is the same test sequence I did.
01A.1 and 01A.2 are Compression test I did twice.

02A is cold start, Pico file worth 32 seconds.
02B is the same cold start of Xentry VVT data worth 120 seconds
02C is another Pico file where its data is at approx 80th second of Xentry CSV 02B file.

03A Pico file is my recording when idling speed has gone down to 700RPM approx
at the same time I also record using Xentry VVT, the file name then is 03B.

For 04A/04B and 05A/05B same logging method as 03A/03B.
Since Xentry VVT file is worth 120 seconds, while Pico is 32 seconds, there will be many idling data in the Xentry VVT file.



==========

Below is another folder where I convert the Xentry CSV to Excel and with German to English translation of the data PID. Just a sample if you also do not speak German , I done the translation already.




===========

The probing I did. All using test connectors, no needle probing.


ABOVE :
QV valve pin 2 (positive ) is to Channel B-Red positive.
QV pin 1 ( negative ) is to Channel B-red ground. Not engine block.




============




BELOW :
CrankPos pin 2 (signal ) is to Channel A-Blue positive.
CrankPos pin 1 ( ground ) is to Channel A-Blue ground. So not to engine ground or shared to other scope channels ground.





===============

Below : Same connection for Bank 1 INTAKE CAM-pos sensor and Bank 2 EXHAUST CAM-pos sensor. I can't access my Bank 1 Exhaust CAM-pos sensor easily, so I use Bank 2.
Channel C - Green is for CAM-Pos INTAKE at Bank 1. Its pin 2 (signal ) goes to positive lead of scope, its pin 3 ( ground ) goes to scope ground, not engine block.
Channel D - Brown is for CAM-POS EXHAUST at Bank 2. Wiring the same as INTAKE.








======================




.



.




.




.





Have fun troubleshooting your signals..........

 

Compression test in Xentry, only the result and as seen by Pico





.




Play around with my Pico file and we can discuss it later when you study them enough on how to best scope your engine signals.

.