I guess that there are hundred of possibilities to explain misfire.

This one seems to be related to carburation (injector).
May I request your help/expertise to try to determine what is the highest probable scenario/issue ?

Symptoms :

• Engine M157 start without issues (exclusively maintained by MB, 120KMiles, no tuning)
• Cylinders run correctly (some very limited misfire one or two on random cylinders) at first start
• Depending of the external temperature :
  • If freezing : everything is OK, no issue
  • If above 32F, misfire happen for 1sec when oil engine temperature reach around 150F
  • If around 50F misfire for same duration (1sec) will happen around 130F

After the 1sec misfire, engine is restoring back to normal operation without any intervention at my side

It also seems that :

• It happen at first start (morning) and never reoccurs later in the day
• It doesn't happen when the motor is pre heated by the Webasto if it reach an engine temperature which prevent the "engine first starts sequence" (more fuel + timing retards)

Additionally

• Reported misfire cylinder is mainly the 5 (20x) but with some (up to 3) also in the 2 or 3 other cylinder (don't remember their number)
• I have already changed spark plug and spark pencil several time without true improvement (misfire could no more happen for 600 miles but will systematically comeback after (following the above scenario)
• Removed sparks highlighted one more carbonated than the 7 others (but I don't have the position of it (garage didn't numbered them)

My guess :

• It seems that something happen when oil engine reach a certain threshold depending of the external temperature. I guess it is the "end" of the "more fuel" sequence (pulling the choke when cold start in the old time)
• The transition from "more fuel" to normal fuel operation is creating instability triggering misfire for 1 sec

Your feeling ? :

• On the basis that cyl 5 is the "failing one" during 1 sec, should I investigate at the level of the injector ?
  One rationale being : Injector is leaking during the night creating the condition to have bad ignition of the gas when "more fuel" to normal fuel transition happen ?

OR

• Is it more related to the need to adapt ECU parameter to ensure that gas injection is correctly tuned ?

OR

• Any other ideas ?

I'm aware that being the cylinder 5, there are potentially other kind of issue (leaking valve,..).

I'm appreciating your help and expertise

Thanks

Terry

 

Terry, you have a very accurate account of what your engine dislikes .

I translate this with short misfire towards the end of engine warmup sequence.

Very likely not cylinder damaged on #5 because it misfires lightly before going away.


> This is caused by borderline mixtures:
Confirm lean LTFT reading ECU data
Replace 2x upstream Lambda.

Enjoy renewed power.


+++ OIL CONSUMPTION:
keep an eye on engine oil LEVEL
replace GOOD oil around 5kMi.

 

Thank you so much CaliBenzDriver,

Your translation is fully summarising my issue

Your diagnostic makes a lot of sense and gaves me the opportunity to extend my testing by logging the O2 sensor value (Xentry diag

Before misfire (cold start)


During misfire (highlighted in red)



Warm engine (after misfire)


My understanding :

• Lamba control upstream R/F stay coherent from cold to warm with "scrambled value" at the time of the misfire
• Left O2 sensor downstream seems to be unpredictable with erratic value especially at cold, misfire and with less extend warm time (transition at borderline mixture) -> I guess that we could conclude that it's the one which is defective (left bank is matching with cyl 5)

I guess that the best option is to replace the 4xO2 sensor at the same time.

There is still one question in my mind, could a poor mixture triggered by another defective item (e.g injector) generate the same kind of graf (all the sensor with "scrambled" value) ?

Relating to Confirm lean LTFT reading ECU data, how could I retrieve it using Xentry ?

Thank you so much

Terry

 

This is a little scary because the last thing you want to do with a turbo car is to go lean under boost.

 

Great, your exhaust sensors are alive. If they were kaput you would get a CEL On.

-- What's wrong with the upstream lambda is they are plugged up by combustion residues. By responding too slow they cause lean response ie. detuned accelerations.

-- Additionally you could also be dealing with oil-in-harness at the O2/Lambda connectors for upstream/downstream. Double-check connectors of 4x exhaust sensors are oil free.


> ECU RULES!
Our personal reading interpretation of exhaust sensors curves are meaningless compared to what the ECU understands. This is the only thing that matters to us.

We could easily misinterpret curves out of context.... both banks don't look even during warm up but that's an inconclusive evidence.

What's significant are the Long-term FUEL TRIMS at idle. Let's see what yours say about your combustion: lean?

 

I never look at the downstream sensors, they seem pointless to me. Your lambda looks fine imo, but may be dirty like Cali said, or the oil thing. I'd post O2 sensor mileage and check for oil at the ECU plugs and get back to us.
It is really weird about the glitch and temp thing. There is some cam timing changes at temp, but water temp, so might want to note the water temp it happens.
There are also some lambda changes, which I don't see in your short window, so maybe if you had a longer log. Look at STFT as well. My ST is far more responsive at idle than LTFT, and your engine is virtually the same as mine. Often my LT won't move at all, it just sits there giving me the finger while ST does all the work.
Spark can change drastically too during warm up, but that should be random, not at xx temp.

Does this issue trip a CEL, or just buggin you? Does it do this every time you transition through that temp range?

 

Thanks Chevota,

No, it doesn't seem linked to water temp. It's a combination of : for this external temp when the oil temp reach for the first time this threshold, misfire (only one time at first start of the day and if the engine did the warmup 'noisy start') First post give an idea of the temp threshold
No CEL just buggin (1 sec max).
Not every time that temp transition. Only after first "warmup start" of the day (but to be honest, i don't believe that I ever triggered a second "warmup start" the same day (oil temp staying above the triggering threshold)

I don't have the full history of the car. But I believe that the O2 sensor are the factory one (120K).
Moreover, my engine around 80K encountered some misfire for an extended time (bad coil on cylinder 3 during COVID, no spare in stock). It's not the same bank, but I guess that poor combustion could increase the deterioration of the O2 sensor (does it make sense ?)

Will try to retrieve the LTFT and STFT

Thank you gentlemen for giving me the chance to make my first steps in a domain that I'm far to master/understand until now.

 

At idle Xentry reports a ON/OFF ratio of purging between 0 to 30%. which seems to mean lean mixture.

The following info seems to be related to mixture (air fuel ratio) too :



I don't know if it could be related but :

• The injection time in the right cylinder is sometime higher than the threshold value 0.70 > Threshold 0.60ms.
• Problem looking linked to "mixture", I have tried the "mixture adaptation test"
  The result is an error detection on "Self adjustement in lower partial-load range, right cylinder bank"
  -> the value is 1.28 when max threshold is 1.20.
  The second attempt confirmed 1.25 > 1.20

I didn't find until now a scanner retrieving LTFT or STFT.

Hope it helps

Thanks

 

I see... run a quick search for "FUEL TRIM" Every OBD gas engine tracks LTFT.

 

Great. Thanks.

You will find below the printscreen in "warm" condition (30 min drive) as well as a log of my way back to home attached (.csv)


It seems that bank 1 (Right I guess) is in lean condition and unbalanced with Bank 2.
I guess that they have both an issue, value being > 10.
I have no idea from where should I start my investigation (MAF sensor, others... ?)

Hope it helps
trackLog-2024-févr.-23_18-46-04.csv
Thank you so much

Terry

 

no maf in these cars

 

Hi PeterUbers,

Thanks for the no MAF info. If I remenber well there is a MAP. Being unique, I don't see how it could be the cause of an "unbalanced air flow" in the 2 banks
Just guessing, if it is not the airflow, should I look a the "pressure pumps" (one by bank) ?

Funny my car is looking pretty most the same than yours (years, model, same engine but certainly better tuned than mine ;-)

 

gorgeous estate! Love that white (as you know)

This is well above my pay grade, you're right about the MAPs, have you scoped your cylinders lately?

 

BANK-1 LTFT: 25%: An extreme correction!!

Terry: can you spot progress being made with certainty and without replacing any good parts....

Now you know for sure why your engine can't fire reliably.


> WHY is your engine TOO LEAN?
Either your engine wants more gas or is getting too much unmetered air.

- it's very unlikely a fuel delivery with B1: 25% and B2: 10%. it's hugely unbalanced off for one rail.

If your engine has two separate rails to use its dual pumps then... HPFP may be have issues (pump, lobe, roller).
Can you read rail pressures? This will also provide a pass/fail answer.

- You can concentrate on Bank1 - It's dragging Bank2 with itself. Fix B1 and B2 will come back to even balance.


- This has nothing to do with ignition. Engine is burning so well ECU is enriching the base mixture.


Either your engine wants more gas or is getting too much unmetered air.
Now you are out finding air leaks...

 

Subscribed for the solution

 

- What's causing the ECU to extremely enrich the lean Bank1 is the $5,000 question.

-- It could be a cracked plastic pipe but NOP because the TRIM is all the way max. A little crack cant supply enough air.

A missing plenum donut seal would do that well.

Such a large trim... it sounds like B1-HPFP is working with a flat roller... (mileage?) - Check pressure for good measure - GDI engine, never want to take 2 days before questioning gas delivery.

I don't think a seasoned PCV would do that because they buildup a carbon plug as the diaphragm is getting torned out .


When was this engine last worked on for like sparkplugs job... ??

In case I hurd you have Xentry handy, you can run a quick crank compression test - This will provide excellent cylinder introductory status without plugs-out procedures.

BTW, DO WE HAVE A CYL#...? Read the misfire counter ✌️

Hopefully we font find out its from the Bank1 limp tensioner screwing intake cam timing.


WHERE do you want to hedge your bets?
> You need to go on a quest for clues (using a can of brake cleaner spray on a cold engine with fire extinguisher nearby - Yes, don't do it!)

 

> EASY STATUS:
The good thing is now you can directly see how the trims are responding to test you're doing. Read the trims for their response on warmed engine.


> VIP SENSORS:
You can study the Upstream Lambda sensors or go ahead and replace them with reputable quality parts: Bosch or MB.

Point being if you run out of certainty clues, Lambda are the sensors grading LTFT.
We have seen how sensor go nuts during the 1sec occurence... may be B1 fuel pump related but both bank response was nuts.
​​​​​​

 

Generally, the ECU can adjust fuel trims to either -25% or 25% max. Sitting at 25% enrichment means, as Cali said, unmetered air is entering the engine and the ECU is trying to compensate. Usually, 25% either way will throw a code. Did you say that you did do a smoke test to check for leaks?

 

Great analysis ! and a lot of good scenario to investigate

To answer to your question:
- sparkplug and pencil where changed recently (20K)
- it's always the cyl 5 which misfire
- one of the sparkplug was carbonate when changed (too rich ?) but no oil
- yes exhaust is smelling a lot gas at startup (whatever it will be first or following during the day).
- Engine (120K) consume 1/3 gal Oil every 5500 Miles (on average)
- Mileage 19,60 miles par gallon
- xentry report sometime late injection (70ms > expected max 60ms) on bank 1
- LTFT1 (right bank) stay around the same high value whatever it's idle or accelerate (seems independent of the throttle)

Just thinking (pure logic, far to reach your level of expertise)

1. Air leakage ?

• Air intake are distinct, each turbo being connected to each of them.
• But each flow coming back from turbo seems to be merged in the change air cooler
  

If an air leakage happens, I guess that it should be between the unique charge air cooler exit and the 4 right or 4 left cylinders (gasket). It's not obvious for me. If there is a leakage, gasket between change air cooler and both side being one "tube", if it leaks at one side, it should impact the other side too (to the same extend ? I guess yes but maybe wrong)

2. HPFP ?

• Bad news, there is only one sensor (B4/25) for both rails
• But there is a "interconnecting pipe" between rail 1 and rail 2 which connect close to cyl 5

Hypothesis :

• Y94/2 bank1 (right) high pressure valve does not feed the expected amount of gas but less. The mixture will be more O2 than gas. Exhaust O2 sensor will report to ECU the need to enrich forcing the Y94/2 to release more. A the end, the motor run smooth (like it is today) but LTFT1 is reported high >20
• The interconnecting pipe could explain intermittent late injection on bank1. I guess that sometime, bank2 is feeding bank1. It could also explain that LTFT2 is less than LTFT1 but still high (10).
• Let's guess that Y94/1 bank2 is not defective but Y94/2 is (supply less). Bank1 will consume (interconnection pipe) Bank2 gas/pressure making Bank2 reported lean too (but at a lower extend).
• I will not say that it is linked but, interconnecting pipe is just connected before injector 5. Does the "balancing" from bank2 to bank1 impact the feed of injector 5 ? (good question).

But I'm still insisting that I'm a total dummy in this domain, it's just "pure logic".

Your feeling ?

Thank you so much

Thierry




.

 

Great Thierry, we are starting to have the topic well circumscribed - What's it gonna be AIR OR FUEL??

Let's check out the fuel pressures first.

> It's possible your Bank1 HPFP is not putting out enough pressure?
Each bank has its own pump and individual rail with a central bridge pipe. Read the ECU rail pressure at cold idle.


> As far as air something may have been disturbed during sparkplug service... perhaps something left unplugged around the intake plenum.


> Cyl#5 vs. Bank-1:
We have an extremely lean Bank1 and Cyl5 on B2 is misfiring. That is the cylinder furthest away from fuel pumps....

Perhaps your low pressure fuel tank pump is acting up. This pump has a controller hosted on the ECU CAN-C.

Pls evaluate pressure data from pump module with scanner.

See if either the low pressure supply then the high fuel pressure collapse under high demand load.
That 3-phase low pressure pump feeds HPFP under a fixed pressure.
It has a built-in filtration that can restrict flow but not pressure...


Once you have certified fuel pressures are about satisfied the issue will have to be on the air side of mixture!
Solid state MAP sensors don't go bad but left unplugged from plenum, a possibility. See that MAP data is active live.

> Single rail pressure:
Each HPFP uses a control valve to build up its pressure.
Comparing both numbers of control valves ACTIVATION will be easy to witness work differences.

 

I know it was asked earlier, but I'll ask again.

Have you tried smoke testing the car to look for an air leak?

Given how the misfires appears when the engine is warm, I recommend smoke testing it under that same condition. Something may be expanding once the engine is all warmed up.

 

That's a good idea which seems simple to apply.

My main issue is the fact that I don't have at this time what seems to be needed.
I will take a look.

Thank you so much

 

That's top recommendation.
I will try and come back with what I will find.
Still need additional time to have the opportunity to block the car for several days and work on it.
Thank you so much for your help and priceless advices.

Will keep you posted

Terry.

 

I don't believe an air leak will do that. The eng will feed said air to all cyls and simply adjust fuel to match. Your idle will be higher if it leaks enough, otherwise it'll idle down ~540. I think it's 540, if fully warmed up.
>>>> My bad, I was thinking you had an M278. You have an M157 eng, right? I looked up three different ones, diff years, and all were 600 in gear, 650 in neutral.

The STFT can go to 25, then, may fault out. Mine will fault if it hits 25 but there is some kind of timer or something else because sometimes it'll fault in a few seconds, sometimes it can be there for maybe a minute and not fault. Why I have no clue.
The LTFT on mine can go to at least 60something and has never faulted on me. The LTFT can work at idle, but it's lazy and may not, or may take a while to adjust. Normally LT is reasonably fast, a few seconds, but at idle not so much. It also seems LT freezes past a certain rpm and/or load, so when you floor it it's frozen and ST has to do all the work. Normally, say cruising at 60, and you suddenly spring an air leak, ST would almost instantly add fuel to compensate, then LT will adj for it and ST will go back to 0-ish. Like a quick fine adjustment and a slow course adj.

Before the O2's kick in, after a start, ST and LT just sit there because they have nothing to go on. Once the O2's kick on, ST & LT will do their thing. How long it takes for them to turn On should be ~30sec, but can vary, maybe up to 90sec?

Pix is two screenshots, one of LT way neg, the other waaay positive. This is because I was messing with the programing, not actually something wrong.



So it seems to me it's actual fuel delivery, which may be the O2 sensor is reading wrong and trims adjust, but it wasn't wrong and now you have too much fuel? If adding 25% then it's pushing the limits of misfire, so maybe that cyl is simply the first to go? There are some adjustments it makes during warm up, so maybe it's bordeline ok and that adjustment causes the misfire before it compensates a moment later. Let me see if I can find that info

 

Gottit. So this is Lambda for Catalyst warmup cycle. Why it's doing all this I don't have a clue, but you can see it does some weird stuff.
Note it says it's based on startup temp, then injector cycles.
I think the Inj Count is based on just one, like cyl 1 only. It also seems this comes to a stop when the O2's kick on, so open loop only. Once I go closed loop it shoots for a lambda of 1.00.



Side note; my fuel trims usually hover ~10. Some of that is the 10% Ethanol in the gas, but mostly it's just the way it is. Most M278 motors I've seen do the same, but not all.
I have one piece of data for an M157 and that particular cars trims sat at 7.