Trying to assess a rich condition

[Breeze1]

And no, I'm not talking about my personal condition, unfortunately

I'll try to be as short as possible to avoid too long a post. A few months ago, I was chasing a misfiring condition and tried all sorts of things among which replacing injectors, until I found out that the MAF sensor, which I had mistakenly ruled out at first, was the culprit. I tried a used MAF sensor at a scrapyard that was visually the exact same part as mine, and boom: misfiring gone. That was a week after replacing the injectors. During that week span (between the injectors and MAF replacement, or right after replacing the latter, I can't tell) I noticed a heavy fuel smell in the cockpit, to the point where I went back at my mechanic's and asked him to check for an injector leak. We found none. The smell pretty much went away, but the gas mileage did seem to have gone down.

I thought maybe the DME was adjusting to the "new" (cleaned and calibrated by rc injection) injectors and needed a brake from the tiresome misfiring issue anyway so I kind of ignored it... From time to time when checking oil level (which remained steady at all times) I would notice a gas smell. For some reason I had doubts about the MAF sensor and checked its part number: it's a siemens vdo which has a 1 digit difference with the bmw one, and was originally made for a 1.9l turbo, 200hp volvo v40...so roughly the same power as a 328i, the exact same shape and diameter, but the same scaling?

2 days ago when replacing engine oil, we found out that the oil filter was, I give it to you, coming apart upon removal (!) and the gas smell was very obvious. I suspected that either the guy who replaced my blown engine had not replaced the oil filter as I asked and left a God knows how old and dry filter in, or that the gas smell had something to do with the filter's destruction (which I doubt after less than 7000 miles since the engine was replaced).

I changed the oil and obviously the filter, drove about 60/70 miles since, and checked the oil smell: it still smells gasoline FFS.

No error codes, the car drives perfectly and it did even with the old filter in. I looked at live data (attached) from INPA and found out that just as before the oil change, the engine seems to run rough. Lambda integrators seem to be thinking that the engine is running lean and adding fuel accordingly if I interpreted that data well.

On a log I made with romraider, lambda integrators look busy as well and seem to be considering that the mixture is up to 17% too lean, again if I read the data correctly.

I also have no cats on the car, but it has been flashed "accordingly": the DME knows there are no post cats sensors and the catalyst heating tables have been edited accordingly, but the fuel tables are still stock...

Basically I'd like to have your opinion on the above, and I'm wondering if:
- the gas smell means contamination through cylinder walls (bore wash?) and how bad it is to have driven about 7000 miles like that, sometimes hard... if it was running waaay too rich, wouldn't I have a rich misfiring condition?
- in case I fix the issue rapidly, am I looking at a mandatory oil change even if the current one has literally less than 100 miles on it?
- about the possible culprit (injectors, which are black ones apparently identical to the original ref of the 328i, since all M52TU's and M54's share the same injectors with the exception of the M54B30, which comes with violet, larger capacity injectors OR MAF sensor), can I rule one out based on this data?


I'm leaning towards a falsely lean condition caused by an off scale sensor, but I don't know if its physical characteristics and the similarly powered car it's intended to run on are any indication that it should be good for my car as well...

Edit: I should mention that romraider's data (lambda integrators banks 1 & 2) is recorded at part/wide open throttle, while INPA's is recorded at idle.

[MarcoZandrini]

How is the “new old” MAF behaving? That’d be my first concern.

[Breeze1]

How is the “new old” MAF behaving? That’d be my first concern.

It seems to be doing well. No misfiring or hiccups, except occaaasionally (mild hiccups, no misfiring per se) for the first mile or two after a cold start, but nothing that qualifies as an "issue" IMO. At low/mid/high loads, the car behaves very well and accelerates steadily and healthily all the way to redline. I'd say if something is wrong with it it'd be more a different scaling when compared to the stock one (which, again, is visually 100% identical and made by siemens vdo as well). But that's the thing: same part with a different part#, aimed at working on a car that makes the same amount of hp (200) although with a different setup (1.9l turbo), how would that affect scaling? I honestly don't know. I tried looking up its spec sheet to compare it to the Bmw one, to no avail.

The MAF sensor currently installed has the ref# Siemens 5wk9 604, whereas the original one is 5wk9 605.

[MarcoZandrini]

I’m wondering if that one number difference makes a world of difference. Does Siemens have an office specializing auto parts in the US?? I’d google the part #s to determine the difference. It could be the cause of the problem. Also, the DME has been reflashed but with what? The MAF may be perfectly fine but if the DME code is whacky you’ll be chasing your tail forever!!

[Breeze1]

I’m wondering if that one number difference makes a world of difference. Does Siemens have an office specializing auto parts in the US?? I’d google the part #s to determine the difference. It could be the cause of the problem. Also, the DME has been reflashed but with what? The MAF may be perfectly fine but if the DME code is whacky you’ll be chasing your tail forever!!

The car has a stock euro 328i CA software on it, with only the post cats sensors and the SAP disabled on it. I have googled both numbers: 5wk9 604 (the one currently on my car) refers to a Volvo v40 1.9l turbo (200hp) whereas the stock part number (5wk9605) obviously refers to my car and a few other inline 6 Bmw's. So I know the MAF sensor I have is not the original one, but I wanted to know if that would really impair the readings, given that part# aside it's the exact same part, same width, everything...

I'm know hesitating between one of these possibilities:
- the fact that only cylinders 3 and 4 are rough is a clue to a possible injectors leak on those cylinders (although they have been double-checked as mentioned before), but that is nothing more than a guess. In fact I don't even know if an injector leak would cause roughness on a given cylinder (a pro could have the answer to that question)
- although visually identical and aimed at a car that roughly makes the same power output, the MAF is indeed scaled differently and causes a falsely lean condition to be detected and unecessary fuel to be added, ending up with a rich condition.

If the two cylinders that show as rough were on the same bank and I had only one lambda integrator showing 7% leaner than stoich I would definitely rule out the maf and look into the injectors leak. But now I have one rough cylinder on each bank and the same value for both lambda integrators because that is all the more fun ya know

So the way I see it the roughness could be caused by either an off scale MAF (only affecting one cylinder on each side, unless the "tau" effect is higher on one side of the banks than the other? ) or by an injector(s) leak, but then again I'm not knowledgeable enough to determine that just by looking at those INPA/Romraider screenshots and thought perhaps there is something in there that I missed but is obvious for a pro.

[Breeze1]

More on my "investigations" here (I think I ruled out the MAF sensor) :
https://www.romraider.com/forum/view...143098#p143098

[MarcoZandrini]

Have you checked compression on those 2 cylinders?

[Breeze1]

Have you checked compression on those 2 cylinders?

I haven't, no (I don't have the tool for that). I'll replace the spark plugs tomorrow just to be sure and dig up the flow report provided by RC injection. I didn't pay attention at the time because I thought they were stock injectors anyway (they are the same on all M52TU's and M54's except the M54B30), but if memory serves well I read flow reports above the OEM indicated 208cc's. I'm suspecting that's were the rich condition comes from. If it's confirmed on the report, I'll lean out the fuel tables by 5% (based on the lambda integrators' 1.07/1.08 readings above and the adaptative multiplicative value which as far as I know are an indication of LTFT) as a start, and see how it goes.

I honestly doubt having compression issues on 2 cylinders because the car just doesn't feel like it. It sounds and drives healthy, doesn't drink oil at all... I made several road dyno logs and they show 220+hp at the crank before tuning, which makes sense with the mods I have (M54B30 intake manifold + CAI and cats delete). That still doesn't explain that fuel smell in the oil though...

[Breeze1]

Back to square one: RC Injection sent me a copy of the injectors' flow sheet and they're rated at 209/210cc, versus 208cc stock value, so basically the same. Since the MAF sensor seems to be on scale as per the comparison I've made between the fuel table and the logged values, I'm going to consider that this abnormal fuel consumption was due to the oil filter not having been changed by the guy who replaced the motor (I don't see how an oil filter could be destroyed in 7000 miles, no matter how rich the car ran), change spark plugs, keep driving and logging for a while and hope that things get back in order.

Edit: I know for a fact that the fuel filter has been replaced recently because it was one of the things I tried to get rid of the misfiring back then...

[328 Power 04]

Everything you showed us says you're running lean, and it's adding fuel to help run around 14.5:1.

Your long term fuel trims are at +5%.
Vacuum leaks, MAF, or maybe the oxygen sensors.

[Breeze1]

Everything you showed us says you're running lean, and it's adding fuel to help run around 14.5:1.

Your long term fuel trims are at +5%.
Vacuum leaks, MAF, or maybe the oxygen sensors.

It ran without o2 sensors for a few hundred miles after I removed the cats so it could have cause a genuinely rich condition. I put them back in 2 weeks ago when I was able to get rid of the misfiring. Saturday upon replacing the spark plugs I found they were dark brown, which would indicate a rich condition (possibly caused, as you said, by what actually is a lean condition). The gas smell in the cabin seems to have gone even after several WOT pulls yesterday. About running lean, "never say never" as they say, but given that every single hose/boot has been checked, as well as the intake manifold and that I have ruled out (to the best of my knowledge: https://www.romraider.com/forum/view...143098#p143098) my MAF sensor being off scale, I doubt it but I'll certainly keep that in check.

Saturday (on a cold start) even before replacing the spark plugs, STFT's were back to 1.01 according to INPA.

Now that oil+filter+spark plugs have been replaced and adaptations cleared, I'll drive it a few dozen miles and make another live data check and log and see how it's gotten.

To (hopefully) conclude this until further confirmation, I'd say that the car running on OL (with a non stock airbox) and virtually with no oil filter (I've learned a lesson to never trust anyone based on that, not even a guy you just bought an engine from) could have caused a rich condition.

A quick word to say again how great RC Injection's team is. The lady (nicknamed "Marc") and one of their techs have been very helpful in trying to diagnose where that issue was coming from.

[Breeze1]

This is live data on a warm engine. STFT's are still in the vicinity of 1 at idle, and while cyl3 is almost back in line with others and much less "rough" than before oil & oil filter+ spark plugs change, cyl4 still shows a very rough condition (it doesn't show on the pic but it goes back and forth between 0.7ish to 1.0+ as seen on this pic. Given that LTFT's are now different between banks 1 & 2 I can't help but link that cyl4's rough condition to a higher LTFT on bank 2.

At this point I'm hoping that things will slowly get back in order and keep monitoring live data every 50 miles or so...

[Breeze1]

Update on this. After the oil+filter+spark plugs replacement, I drove about 70 miles, made another live data check with INPA, and found out that the lambda integrators were now showing 0.95 instead of 1.05 before. So I thought "finally data seems to agree with this rich condition" (see comments above about gas mileage, fuel smell...) and multiplied fuel tables (base+ part load banks 1 & 2) by 0.97 to lean them out by "only" 3% to be on the safe side, drove again, logged with INPA again, found out that the LI were back to 1.00 and the multiplicative (LTFT) values down to about 3 and 4% respectively. The fuel smell in the cabin after a WOT run seems to be gone as well.

BUT I then logged a drive with romraider to see how the LI would look like in actual driving at low and mid loads. AND the LI in romraider actually skyrocketed, going from a max of 16-17%, with most occurences at or below 10%, to as much as 24%

Which leaves me with a question for those of you familiar with romraider: is there a table that overrules the base and warm fuel tables? Because I thought the fuel enrichment table, which I left untouched, was expressed in relative terms to the base table? So how could the LI increase instead of decreasing if the target IPW has decreased??

I wanted to further investigate and I started looking at vanos values. Went back to a log made before leaning out by 3%, and found a delta between vanos targets and actual values, either advanced or retarded by mostly 1 to 2% on both intake and exhaust sides. I also noticed in INPA that the reference (initial) intake cam position was showing 121.88 instead of the DME's value which is 120, so retarded by 1.88%...

If that ref point discrepancy at idle somehow translates into actual driving, and considering that less air than expected (the intake valve opens later) enters the engine, while on the other hand exhaust gases probably exit the combustion chamber at higher velocity because of the decat situation (less backpressure), is it possible that I'm looking at a falsely lean situation with rich consequences, resulting from inferior than expected cylinder filling?