RomRaider

Hello everyone I've been driving a Carberry for a couple of months now, and I've run into two problems like this:

1. During acceleration, especially at low revs and high gear, when passing the 1800-2500 RPM, my CEL lights up, as if a detonation has been detected. Not every time, but often enough (every ride). It is easy to repeat the problem, for example, accelerating in 3rd gear from 1500 rpm to 3000 - almost always, in the specified interval, the lamp lights up. At the same time, judging by the logs, there is no detonation at this moment, ever. FKC and FLKC (and Knock Correction Advance) are zero, or contain small corrections in other places. It is also interesting that I used to have light corrections in this interval in FLKC, but now they are not there either.


What is it and what could be wrong?

2. I'm trying to figure out how to manage AFR by editing the table Primary Open Loop Fueling (Anti Lag Conditions Not Net). For the test, I set the AFR 12.06 in the boost and rpm ranges that I use. However, when writing logs, I see real AFR 10.5, despite the fact that the values Fueling Final Base (AFR 12.08) and Primary Open Loop Map Enrichment (AFR 12.07) contain the values that I have set.


What am I doing wrong? And why does the fuel and air mixture not obey the tabular values?

hard to say for sure without seeing your rom, but you usually need to tune the VE or MAF scale so that your wideband o2 reading matches your target AFR - also generally an AFR target of 12 on boost is gonna mangle your engine sooner rather than later. Has the knock stopped because your IAM has dropped maybe?

Thanks a lot for the reply! I used the base ROM from this post, with 100% SD:
viewtopic.php?f=28&t=16566
but I have already modified it quite deeply, both based on my previous settings on tuned stock ROM, and based on road tests and logs. I think the VE map is done, but I'll try to revise it.

Do I understand correctly that if in OL mode I have a difference in the Estimated and real AFR form wideband o2, then the ECU, after several iterations, will make a long-term correction to A/F Learning, to fix it? Or is this true ONLY for CL mod and the discrepancy on OL can be corrected only manually by adjusting the VE table?

Regarding the mix in 12.0 AFR, I understand that this topic is very controversial. We have a pretty powerful Subaru community here, I am often asked for help with stock firmware, and I saw a car on the EJ20x, with 1.3 bars of excess, working perfectly on a mixture of 13.6 AFR, on 98 gasoline. I was impressed by this and I want to test something similar.

in open loop operation there are no adjustments to the fuel trims - because the factory o2 can't be relied upon when the manifold is filled with high exhaust pressure and partly because the ECU isn't up to the job. However your A/F trims can still be applied from range D so it's important to make sure your trims aren't too wild. If your VE was correct your AFR on the wideband (note, wideband o2, not factory o2) would be a match so if your seeing them not match then your VE is not correct in those areas.

Thank you very much! Now it will be much easier to figure it out. I will try, additionally, to make a flowchart of the logic of the fuel system in order to better understand everything.

Well, I figured out what was going on and solved that, the problem was with this particular range in the VE table, no false detonation or anything like that. The screenshot shows that the work at this moment is on an open loop (status 10), the expected mixture is 12, but rich as much as 10.5. The problem turned out to be overestimated values of volumetric efficiency, as you say.

The CEL continues to blink sometimes in strange ranges without detonation, but the ECU sees the real detonation and works out as needed, apparently some kind of glitch in the algorithm of CEL. Maybe I can take some information about it to developers?