Jeedie

New to this forum, let me shortly introduce myself before diving into the issue I'm running into..

I'm a bit of a car nut, owning several cars, from sporty (Miata / Lotus Elise) to more comfortable (BMW and now Mercedes). I live in The Netherlands and found this forum while looking for a Mercedes. I found a wealth of knowledge and insights here which really helped me finding the right car for me.

Last week I bought a used 2013 S212 E300, with the 3.5 M276 engine. All seemed fine and a throughout technical inspection performed by Dekra showed no issues with the car. So I bought it in confidence it was a fine car. However, on the second drive I ran into issues when starting the engine. It misfires badly, won't rev, smells like unburnt fuel and even lacks the power to drive at walking pace. Pretty bad.

I've acquired the Launch Creader Elite to be able to diagnose issues with the car. The ECU reported P0190 and several misfire codes. The P0190 pointed at the fuel pressure sensor in the high-pressure rail. Looking at the live values coming from the sensor, the ECU reported a suspiciously steady 6.02 bar... Searching this forum, I found that the ECU substitutes the rail pressure with '6 bar' on sensor failures. So, I dug in and replaced the sensor with a new unit today. Now, with ingnition turned on, I read a more plausible 2.53 bar, which is moving about a bit.

Well, it seemed that solved the issue, so I decided to start the engine. After a few failed attempts (probable still some air in the rail), it started. However, the pressure measured by the new sensor is still way too low, hovering slightly around 3.29 bar. The fuel pump pressure is roughly 6 bar. Of course, the engine is still misfiring badly.

What would be a logical next step in trouble finding? Is there a way to connect an external pressure gauge to the rail so I can crosscheck the low pressure? Or could there still be air in the system that's acting up?

Any hints on how to proceed are very welcome, I'm handy with a spanner but new to the world of Mercedes..

Below is a live picture of the pressure reported by the ecu on the rail sensor.

chassis

Unplug the diagnostic device and open the hood/bonnet.

Open harness connectors for cam position sensors and cam magnets and look for signs of oil. This causes misfires and is common on MBs of ~2012 and later. If you find signs of oil, replace the sensors and magnets and clean the oil from the harness connector with electrical contact cleaner. Cleaning the oil might give temporary relief from symptoms but leaking sensors and magnets cannot be repaired, they must be replaced.

Check the ECU connectors for signs of oil and clean it away if any is present.

Search this site for hundreds of posts on this topic.

Jeedie

Thanks! I've just quickly checked all 8 points there are known to leak oil into the harness, those are all bone dry.. Does anyone know the normal expected fuel rail pressure at idle? Just so I can make sure that the 3 bar I'm reading is correct or incorrect..

S-Prihadi

The one in red :
Looking at the live values coming from the sensor, the ECU reported a suspiciously steady 6.02 bar... Searching this forum, I found that the ECU substitutes the rail pressure with '6 bar' on sensor failures.
The LIMP mode uses assumed Low Pressure Fuel pump target pressure ( assuming fuel filter is healthy aka clean ), which is 6 BAR at the sensor on top of the fuel filter, at the fuel tank 4 meters away from engine.


The one in green :
The fuel pump pressure is roughly 6 bar.
Make it a habit to clearly indicate LP or HP fuel pump you speak of, while I am 99% certain you are referring to LP ( low pressure ) fuel pump at the tank, but accurate information is a must
when one is to assist you.


I suspect at the moment with information at hand, you HP fuel pump (HPFP) fuel quantity valve is defective.
This HP fuel pump is a NORMALLY OPEN type, meaning when the fuel quantity valve goes bad it will not do its magic of manipulating when to pressurize the HP fuel pump,
and then the HPFP will only get at its input the amount of pressure the low pressure in-tank fuel pump ( LPFP) can offer.
LPFP fuel pressure itself will have a head loss of approx 1.2BAR when read at the test port at engine bay.
The test port is also the fuel input for HP fuel pump. See below :















You must have this fuel pressure tester to read the fuel pressure at test port.
https://www.thedrive.com/reviews/305...essure-testers


For you to troubleshoot, do it properly and start from Low Pressure (LP) side too.
LP side is the start of everything, HP side is a secondary stage.

When you have the fuel pressure tester and a proper GOOD fire extinguisher , do a proper fuel pressure vs flow test, using a proper
fuel jerry can and your baby Autel. The function to use at the Autel is N118 fuel controller module, fuel pump ACTIVATION.
DO BE WARNED that the Autel can turn on LP fuel pump but it CAN NOT stop it until the 30 seconds countdown, so make sure no leak on the fuel tester hose.
Yep, that is the dumb-azz part I hate with the baby autel, but I love its other very good capability.

Cover any connector , example ECM, where a fuel spill or leak at such high 5-6BAR of pressure can wet those connectors and create fire risk.
Be careful, you engine is a V engine and a fuel leak can hit exhaust manifold and create fire hazard.
So wait till engine metal at exhaust manifold to be no hotter than 50C.

When and if LP fuel pressure is verified by analog pressure gauge at that fuel pressure test port, we can then also rule out any fuel clogging
at fuel filter itself and between the output of the fuel filter + its short plastic blue hose + the fuel pipe under the car.....and including the stainless steel braided protected flexible fuel
hose into the test port ( HP fuel pump inlet ).

By allowing fuel to drain out to a jerry can and you read the fuel pressure both digital one at fuel filter using baby Autel and analog wise using fuel pressure tester, you are testing
indirectly LP fuel pump vs flow performance on a known head pressure loss. Measure the amount of fuel you get for the 30 seconds duration of the fuel pump activation test.
Or use the fuel tank DRAIN feature of baby Autel to pump out fuel at a good measured 60 seconds.
If you have a good measuring flask of 5 liters, even better. Or you can use accurate scale to measure weight of the pumped out fuel at jerry can at 0.74 specific gravity. 740 grams per 1 liter of fuel.


When all above done and confirmed LP side is decent or very clean.
Only then we hunt down the HP fuel pump..............

S-Prihadi

Clean filter 5.8 to 6 BAR at engine OFF, ignition ON.

Real test for fuel pressure under use at WOT is below :



Your fuel filter P/N for chassis W212.255 is the same as mine.
My engine burn a bit more fuel than yours at WOT, because I am at 245Kw, yours probably 215KW or so

S-Prihadi

I forgot to add :
Baby Launch module ID for low pressure fuel pump controller N118 is : FSCU , item 16








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

Now, before you asked any question on how is the pressure behavior of M276 HP fuel pressure, let he shed some light.

NOTE : I am using my engine as base line. M276 3.0 Turbo. M276.820 is the variant.
It is probably 100% the same HP fuel pump pressure protocol to non turbo M276 3.5 NA.
To date no one log in real time at our W212 section their M726 3.5NA HP fuel pressure with RPM and throttle opening, so I am assuming it will be the same as my 3.0 turbo engine.

Why I am telling you this is, you may while doing the data logging with your baby Autel get surprised that my engine ( probably yours is the same ) ,
does not do maximum 188BAR all the way to WOT redline. Both your Miata and Lotus Elise ( Toyota engine ) is not GDI ( direct injection ) and has no HP fuel pump, so you may not be familiar
with the Bosch HDP5 family of HP fuel pump we are using on M276 / M278.

Take a look at my log, attached.

Also use M276 3.5NA engine booklet for information on how the LP and HP fuel system work, also attached.


Have fun....................

Jeedie

Some very useful information, thanks a lot! Your logging confirms the idle fuel rail pressure to be a lot higher than what I'm witnessing. A sudden, complete failure of the HPFP as a cause is pretty unlikely, so I'm currently. thinking along these lines:

- My new fuel pressure sensor is broken. It's not unheard of for new components to arrive broken. Due to the relative low cost of the sensor, I'm going to try another one to rule this out.
- If the sensor is not the issue, the actual pressure in the rail must be very low. So I need to look for:

-- Reduced flow into the high-pressure system, either the LPFP failing, or the quantitive valve not operating, or not being controlled by the ECU.
-- Excessive loss of pressure through the injectors. Maybe one or more injectors are over fuelling massively, or perhaps an injector is stuck open.

I'll check flow and pressure on the low pressure side as described above, this would be a pretty straightforward check. It's hard to imagine this is the actual issue, as there isn't even enough pressure to idle. So the total needed fuel amount is very low. At the same time I'm getting a strong unburnt fuel smell from the exhaust, indicating it's over fuelling.

Hopefully my baby launch scanner can show enough information to assess the operation of the injectors. An audible test to confirm opening/closing of the injector is straightforward as well.

Regarding the quantity valve, the engine booklet mentions a PWM signal controlling this valve. I get below info from the launch scanner during startup / idling of the engine:



There might be an electrical issue in getting the PWM signal to the valve, or the valve itself might be acting up. I'll see if I can get hold of a scope to check this signal. Is there an ECU pinout diagram so I can check the conductivity from the ECU to the quantity valve connector Y94?

chassis

Which retailer did you purchase the fuel pressure sensor from?

Jeedie

It’s a Bosch item, just like the original. Only the last 3 digits of the part number are different, xxx102 to xxx115.

chassis

Thanks. Which retailer did you place the order with?

Jeedie

A local supplier in The Netherlands, mijnautoonderdelen.nl (my car parts).

S-Prihadi

Post #7 here is about HP fuel pump. Do note your engine is the 3 lobe type camshaft fuel pump lobe and mine being the 3.0 Turbo M276 is using the 4 lobes one.
https://mbworld.org/forums/e-class-w...ml#post8554309

When you have read well my post, you will discover that scoping the quantity valve of the HP fuel pump will bring more mystery as you must know the cam's fuel lobe timing
of the lobe highest peak vs the pulsing of the PWM signal to the fuel quantity valve. I have not measured my camshaft actual fuel lobe peaks (4) physically in relation to camshaft zero degree,
so the mystery is still a mystery for me.

Attached M276 engine computer wiring schematic, 2 files. Y94 is the fuel quantity control valve


This was my test on the HP fuel pump ( post #7) and low press fuel pump.
Post #7 https://mbworld.org/forums/e-class-w...ml#post8554309

If you ever want to remove the HP fuel pump, some notes here : https://mbworld.org/forums/e-class-w...l-warning.html

Now one last caution.
3rd party scanner update speed at best is only 1HZ. So when you are looking for FAST data speed between what is the actual command from ECM to what you are seeing on the scanner
display, you best use fast C4 Xentry like the one reccomemnded by BenzNinja, otherwise you may diagnose wrong.
Case to note is when doing VVT Test as per MB Xentry.

S-Prihadi

GDI injector is inside the combustion chamber, next to spark plug. Not like port fuel injector at the intake manifold which get assistance from intake "vacuum" pressure.

If GDI injector pressure is only 5BAR because HP fuel pump fuel quantity valve is bad and no pressure boosting occurs, strong un-burnt fuel smell is not purely from over fueling.
Cylinder pressure at TDC is approx 10BAR if I remember correctly.
So the GDI injector with a lousy 5BAR has to overcome cylinder pressure build up and could not do proper fine mist spray, would probably be fuel dripping like a runny nose during a cold
and combustion will be bad, hence that strong fuel smell. Extremely poor fuel combustion.

Jeedie

All brilliant info that will help a lot in solving this issue. I'll follow the steps and hints from above.

S-Prihadi: Regarding the PWM signal to the quantity valve: usually these PWM signals run at a much higher frequency than the lobes on the camshaft. At idle, the camshaft is running at ~350rpm, roughly 6Hz (I just noticed from your pictures there are 4 lobes, so 24hz) on pulsing the HPFP. PWM signals can easily run at 200Hz for a quantity valve. Translating the PWM to the valve opening, you only have to look at the time spent either at the top of the pulse or the bottom of the pulse. If it pulses to ground for 40% of the time, the valve will be 60% open (as it's normally open). There should be no need to map the pulses onto the pulses of the cam lobe.

Regarding the fuel smell, your logic also makes sense. If there's no mist, there can be no, or very poort, combustion. Can you run these engines without the intake manifold installed to check the injectors?

Jeedie

And thanks a lot for those ECU diagrams, that helps a lot to check conductivity of the wires! I'll have to make sure to get WIS/Starfinder, it's a pretty complex machine compared to my other cars.

BTW, the Lotus I mentioned is not even a Toyota engine but a Rover K series, introduced in 1988. It does have a nice variable camshaft and fuel injection, but it's from a completely different era regarding engine technology.

chassis

Best practice regarding electrical parts is to purchase directly from MB main dealers, or main dealer online shops.

S-Prihadi

I see, I have always thought Toyota engine was the choice since day 1.
Lotus have assisted Toyota for handling improvement on their Supra line since the 80s

S-Prihadi

In red :
Nope, the intake plenum is part of the intake manifold and also is where the throttle body is at.
Google Utube, you will find lots of example.
My 3.0 Turbo is different for its intake system and much easier to work when it comes to acces COP/spark plugs and injectors.

Translating the PWM to the valve opening, you only have to look at the time spent either at the top of the pulse or the bottom of the pulse. If it pulses to ground for 40% of the time, the valve will be 60% open (as it's normally open). There should be no need to map the pulses onto the pulses of the cam lobe.
Your are thinking of a perfect round fuel lobe, so to speak.
For your 3 lobes camshaft, only 60 degrees of the 120 degrees per lobe is when the fuel quantity valve is allowed to power up or "close" and create the pressure boosting.
When not powered up it is normally open, so fuel is allowed into the compression chamber of HPFP.

I give you a 4 lobes example, out of 90 degrees per lobe, only 45 degrees is use-able to create the pressure boosting.
So delay in fuel quantity valve actuation due to say poor winding condition, will result in wrong timing to close the fuel quantity valve.

Jeedie

So, today I've (once again replaced the high pressure sensor, now with an original MB part supplied by my local MB dealer. Symptoms are similar, so it's not the sensor. I'm still reading a very low pressure on the fuel rail. Interestingly, this sensor reads about ~2 bar higher than the other one. It seems they are not very specific in the values they read, but of course sensitive enough for the ECU to control the mixture, combined with MAF and lambda values.

Next step is to measure the low pressure side of things, I'll post the results here once finished.

Jeedie

Well, results from the low pressure part are in. Pressure shown at the engine side (Schrader valve) are roughly 4.5 bar with decent flow. I get 1.3 ~ 1.5 bar pressure loss from the LPFP to the top of the engine, roughly in line with the values reported above.

So clearly there is enough fuel pressure and flow on the low pressure side. Next step is to get the HPFP replaced, it'll take a few days to get hold of the right parts.

Jeedie

From reading the instructions on replacing the HPFP, I get the impression the 2 bolts that are holding the HPFP to the housing are one time use only. Reference step 30 in the removal instruction PDF posted above. Is there a MB part number available for these bolts so I can order the right ones?

Edit: Just found the part number on FCP Euro, the bolts have MB number 0049903712.

S-Prihadi

J,

This HPFP of yours will be a good case study for all of us

2 cautions :

01. Do not ever run dry a HPFP out of fuel. The fuel is the lubricant for its thin rod sliding pistion.
No more one is allowed to run engine till it died in the name of easy pressure release/reduction of HP fuel circuit like a port fuel injection system.
Doing that will make HPFP fuel dry at its compression chamber and scoring of the thin rod sliding piston will happen.

Prime and bleed well the HP circuit when installing a new HPFP using LPFP actuation by baby autel.
Since HPFP is a normally open fuel quantity solenoid, using the LP fuel pressure tester as bleeder is not too bad.
If you can bleed a tiny bit at the pipe at Bank 2 ( LEFT SIDE ) fuel rail, that is even better. That was what I did.

02. Do not ever disconnect power wire to injectors of a piezo injector type found on M276/M728 when engine is running.
Older version of GDI injector still uses magnetic type solenoid, we use piezo type and it is very different in its electrical behaviour.
Magnetic type solenoid will self return to fuel close/shut position when power is removed, because there is a spring.
Piezo injector does not use spring to fuel close/shut position, it uses reverse power generation behaviour of a piezo stack and the ECM is the one
allowing a piezo to discharge its power and only then the fuel flow can be shut off. Piezo stack is like a generator and a load at the same time.
Piezo takes power to change from A "shape" to B "shape" and need to dump the build up power in them to go back to shape A.
So, removing power from a piezo injector while engine is running can cause fuel flooding in cylinder like an OPEN injector, with worst case scenario of liquid hammer
at combustion chamber.

You can read here : https://www.pfjones.co.uk/disconnect...s-running.html

If you like scoping, you can take my scope capture of our high voltage piezo injector from Pico forum, here : https://www.picoauto.com/support/topic23230.html

Screen shot :
The top image green channel (C) is current, the down ( reverse ) going current is when piezo dumped its power into ECM
Top image is cleaner and more accurate because I use a differential probe.




Do us a favor.
Please investigate further when you have removed the HPFP for the nature of the defect.
Is it a dead windings of the fuel quantity valve or a slow winding ?
Is the mini piston shaft scored big time or it is broken ?

Thanks

Jeedie

Well, partial result by replacing the HPFP. Today it took about 4.5 hours of spannering to replace the pump. The quantity valve connector on the old pump was badly damaged, hanging by just a wire:



The previous owner must have tried to 'fix' that with a bit of epoxy or something. A classical bodge job..

Anyway, after starting the engine it still showed very low pressure on the fuel rail (still roughly 5bar), but for a few seconds, I suddenly got good pressure (~50 bar), and the control of the quantity valve showed very different values. Idling at that moment was just fine! But then, just by itself, it dropped back to 5 bar, stuttering, and it hasn't changed since.

I suspect a wiring issue between the ecu and the quantity valve now. From the wiring diagrams posted by S-Prihadi (thanks a lot!), it looks like there might be a connector somewhere between the two. Does anyone have access to star finder to check whether it's a direct connection, or if there's anything halfway?

S-Prihadi

Nope, no intermediate connector in use. If it does have one, there will be indicated in the wiring schematic as X something.
My own engine is also directly wired to ECM connector for Y94 fuel quantity control valve.
However, I did make splice and place a Deutsch connector in parallel for future scoping work.

See blue marking of extra interconnect connector as X

Above, the X1 means interconnect connector
Below the X86/4 and X86/3 are interconnect connectors






You best check the ECM connector the big size one, the 96 pins one. Y94 uses Pin 53 and 5 the female terminal which is at the connector with wires.
That is Kostal MLK 1.2 terminal, its width is 1.2mm and thickness not to exceed 0.6mm for the subtitude male blade terminal u will be using if you do not have Kostal MLK 1.2 male terminal.
Do a loaded shake-the-wire test , load it to 2 amps with a 20 watts bulb.

S-Prihadi

If u want to keep the car for sometime, you can prepare it now for easy scoping in the future, like what I did :
https://mbworld.org/forums/e-class-w...eparation.html

Back probing ECM connector with those backprobe needle is not too good a habit, and also not a very good contact quality when engine is vibrating.