RomRaider

I have a '99 E36 M3 US. It has 24# injectors, 803 MAF, and M50 manifold. I am running the matching tune from EPIC Motorsports who is supposed to be a good tuner.

The car runs very well. No strange idling issues, no stuttering, it never stalls, and it pulls strong to redline. Despite that, I occasionally get a fault code related to O2 sensors maxing out. I have been chasing this problem for quite some time. I replaced the MAF and both front O2 sensors, and I have been unable to find a vacuum leak. While troubleshooting this on another forum, someone suggested that the fuel injectors I was sold may not be the right size. This is when I found out about RomRaider and am hoping to "fix" my fuel maps if they are indeed the problem.

Since the lambda integrators are high, I would need to add fuel which I am comfortable doing via RR. My only hesitation is that if I truly do have a vacuum leak that I haven't been able to find, I might be making my engine run too rich. Is there a way to tell via logs if there's a vacuum leak? For example, would a vacuum leak affect all engine RPMs? It seems that it's worst between 1900-3900 RPM. If I make my changes and the lambda integrators are still high, does that indicate a vacuum leak?

I didn't have a look to your log but are you sure that your O2s are not swaped from cylinder banks ?

When I installed the new O2 sensors, I used the factory wire routing which seems to make the connector obvious given the length of the cable. It isn't physically possible for me to hook them up the other way because one would be too short. That said, I don't know with 100% certainty, and I will try to verify.

I am still curious about my question(s):
1) If I had a real vacuum leak, would it be apparent across the entire range (or most of it)?
2) If I corrected my fuel table based on integrator numbers, would the integrators still read high?

Factory wiring routing is kinda counter intuitive in my opinion but yes, the bank one sensor goes to the rear connector on the fuel rail and bank 2 goes to the front connector due to the wire lengths being the same.

Vacuum leaks are most apparent at idle and light load, wont see it at heavy throttle or WOT. Looks like your trims are pretty close at idle, only around +3%. Whats the condition of the MAF? Might just have to clean it but I see where the trims are about maxed out at 30% throttle

Compare your MAF scalar against the 803 MAF scalar in my MAF scaling thread. it's possible they got it wrong.

Also, your o2 sensors are not swapped. They are fine.

Sorry for the radio silence - apparently I wasn't subscribed (or accidentally unsubscribed) to my own topic, so I wasn't seeing anyone's responses.

I have since confirmed 100% that my sensors were wired correctly (Front -> Rear, and Rear -> Front).

Since I live in Los Angeles, it's hard to get safe logging runs, but the coronavirus lockdown has improved that considerably. I was able to do some logging, update my base fuel map, then log again. The integrator numbers are definitely trending in the right direction which is encouraging.



MAF and O2 sensors are brand new



You might be on to something here. I copied my MAF scaler into your spreadsheet and plotted the curves. My "mystery" scaler is in between your listed stock and 803 scalers. Now I have more questions than I had before
1) If I update my MAF scaler to your 803 numbers, would that invalidate any table that has a load axis?
2) I presume that my tables were tuned based on the mystery scaler. Would I be losing anything by continuing my tuning using the mystery scaler?

Looks like they scaled it to fit in the 1024kg/hr limit

That makes sense. I'm guessing that was probably the normal way to use a "bigger" MAF on these cars prior to the RR community's improvements. I assume this isn't a problem for me, so I'll continue to the next step which is trying to finalize this tune so I can get my car to pass smog in California.

I notice in my data that the integrator is adding fuel around idle which gives me this nagging feeling that somehow I still have a vacuum leak even though I can't find it. I've got a mityvac and a TechEdge WBo2 sensor, so I'm hoping those will help confirm if there's a leak. My thinking is that the car should be running close to 14.7 AFR at idle, but if I have a vacuum leak, then extra fuel will be added and I should see that with my WBo2 sensor.

This might be just bad MAF/injectors scalling. Your car will be running at 14.7 at idle even with small vacuum leak if integrator is able to compensate for the leak. You will still read lambda = 1 using WBo2 sensor. Only after disabling fuel trims totally you will read leaner AFRs but it still does not mean there is a leak... It is much easier to diagnose in cars running stock hardware as those are much better calibrated from factory.
If I were you I would just fine tune fuel maps to get fuel trims within +/- 1% and call it a day if you are unable to find any vacuum leak.
You can also try to check vacuum at the intake manifold at warm idle using good gauge / MAP sensor with high resolution (so something measuring up to 1.5 bar absolute would be best).

Good point, and thanks for your feedback. I'm still baffled that such a common set of mods have proven to be so much trouble, but I'm glad RR exists!

I'm starting to wonder why I don't just take a stock M3 ROM, and apply a linear scaling to the fuel map for my larger injectors instead of trying to reverse-engineer what the previous tuner did. In thinking through this, I believe this is not straight-forward because the previous tuner scaled the 803 MAF to 1024 which means that it will always read a lower airflow than is actually coming in. I was thinking I could account for this, but the MAF scaling won't be linear (unlike the injectors), and so it's not as easy as multiplying the entire table by some constant.

So I see at least 3 options:
1) Convert to 2048kg which will make the MAF airflow read accurately
2) Stick with 1024kg, but somehow account for the MAF scaling in the fuel map
3) Start from scratch using the method in the big fuel tuning thread

As an additional note, I am trying to get this car to pass the SMOG sniffer test (I'm not worried about the visual test). That is why I was thinking that I could start from the stock map and just make adjustments since my car only has bolt-ons and is still N/A.

Does anyone have thoughts on the above?

I took a break from this due to some life events, but now I'm back trying to get this thing working. Since my last update, I:

1) Removed 803 MAF and went back to stock E36 M3 MAF
2) Flashed stock ROM, but with Fuel/Base map scaled for 24# injectors and leaned out slightly at 125 load and low-to-mid RPMs

I then drove for ~50 miles, all my emissions monitors were completed and no check-engine-light. The LTFT ended up around 4%. I went to get smogged and failed, but not by a huge margin. Here's my results:



To summarize the above:
1) I failed the HC test, but not by a huge margin
2) My CO% numbers are 3x higher than average, but still passing
3) My NO is below the average and well within passing range

I chatted with the smog tech and he felt that the high values for HC and CO% are due to some unburnt fuel in the exhaust, and a rich mixture. It also seemed odd that my car was adding fuel (+4% LTFT) even though the smog machine detected a rich mixture. This further supported the idea of a vacuum leak that I couldn't find. I then did the following:
1) Replaced all spark plugs, spark coils, and front O2 sensors
2) Bought a smoke machine and found 2 vacuum leaks (oil cap and throttle gasket)

The car seems to run slightly better now. The idle is more stable, and drivability seems slightly better. I just finished driving about 30 miles, hoping that the LTFT would go lower, but that doesn't seem to be the case. LTFTs are right around +4%. I don't know if that's a good sign or bad sign. I double-checked and cannot find any more vacuum leaks (I checked the brake booster too).

At this point, I'm going to schedule another smog appointment and see if my results are any better. The smog guy likes me, so I'm also going to ask if he'll let me log the smog test with RR so I can see exactly what's going on.

If you reading rich even though the ecu thinks it's lean, I would look into exhaust leaks rather than efore the engine

Thanks for the tip. I did find a small leak between the exhaust manifold and the mid-section. I tightened the nut 1/4 turn and it seems to have went away.

I also realized that any tuning I had done previously was invalid because of the vacuum leaks. I went back to a stock Fuel/Base map and scaled the IPW down by 20% to account for my larger injectors. Now the LTFT fluctuates between -1 and 1, and often ends up at 0. I think the original problem might be solved, HOWEVER I cannot seem to get the SAI emissions monitor to complete. All other monitors are fine. The SAI had passed a few weeks ago when I last got smogged, but not since I fixed all the vacuum leaks and the Fuel/Base map. It could just be that I need to try a few more times. I'll datalog my next attempt to see what's going on.

I logged this attempt and I'm not seeing anything obviously wrong. When the car starts, I can hear/feel the SAI running for ~90 seconds. During this time, the logs are showing lean which should be the acceptance criteria for the SAI monitor. The log covers the 5 minutes I was idling. I also drove around for 5-10 minutes afterwards just in case, but it didn't seem to help. No codes/lights. Without anything to fix, I'll repeat this process until it hopefully passes.