warm-up fuel automatic adjustment

[AndreyB]

Computers are cheap and humans are priceless so I hope that warm-up fuel automatic control would be a nice feature to save some time.

Currently 13 AFR target below 80C is hard-coded, here are the logs & video:

https://svn.code.sf.net/p/rusefi/code/trunk/misc/logs/2003_dodge_neon/2016-02-09_18.20.20_warmup_pid.msl
https://svn.code.sf.net/p/rusefi/code/trunk/misc/logs/2003_dodge_neon/2016-02-09_20.07.47_warmup_pid.msl


[video][/video]

[AndreyB]

https://svn.code.sf.net/p/rusefi/code/trunk/misc/logs/2003_dodge_neon/2016-02-14_12.42.33_warmup_pid.msl

Something is wrong - looks like this warm-up auto-correction has killed a second set of spark plugs.

Everything started really promising - I've left the car outside yesterday knowing that we will have a pretty cold night. This morning CLT was showing -20C/-4F. After some pretty long cranking the car did start! "debug f1" field in this log shows how idle air valve was gradually closing, RPMs were gradually going down. PID logic of CLT correction multiplier was keeping fuel squirts about 4ms and AFR was around 13.5 (target AFR is now increasing while CLT value is growing, no record of target AFR unfortunately), and then something went wrong around time stamp 390s: AFR indicator goes up to 17 and then all the way to 20 and 20 is the end of range and it goes off limits. What I suspect happens in reality is engine was running richer and richer but the sensor does not register that and spark plugs were getting more and more unburnt fuel on them, and then the car dies at around 520s, and then I cannot start it again.

Still an open question what is the root cause of the issue. AFR were pretty stable while CLT correction was going up, CLT correction was actually compensating for "fuel: base" which was going down (probably due to increasing tCharge)

[puff]

how long does it usually take for o2 sensor to warm up? (afaik, it doesn't provide accurate readings until it's warm enough)

[AndreyB]

This WBO has it's own heater so hopefully it's not so bad and hopefully the WBO is operational. The engine was running give or take OK for about 200 seconds.

I've noticed MAP spikes at 318secs, 354 secs and 362 secs. MAP spikes from 28kPa to 31kPa, this causes fuel squirts to go from 4.1ms to 4.6ms, this should have shown as lower (richer) AFR values but instead we are showing higher (leaner) AFR values. Looks like the fuel squirt spikes are causing the mixture to stop burning, WBO shows this as lean while in reality I have excessive fuel in the cylinders? I do not have a better explanations.

What is the root cause of these MAP spikes is another question. I have no idea what's going on. Oh. and by the way car would not start anymore - on brand new set of spark plugs, on neither stock ECU not my ECU.

[AndreyB]

What is the root cause of these MAP spikes is another question. I have no idea what's going on.

Actually trigger synchronization counter increases at same timestamps, trigger errors probably cause a misfire which causes the MAP spike which causes AFR spike.

[kb1gtt]

I'm reasonably certain your RPM doesn't change as fast as the graph claims under the normal conditions. I believe your seeing a certain level of noise on the RPM signal which is adding some jitter to your signal, which is causing the ECU to think the RPM is changing more than it really is. Occasionally this gets to bad, which causes the fuel to jump, which results in the misfire, the misfire creates exhaust pressure which creates the map pulse.

I know I sounds like a broken record, but I think you have a GND problem.

[AndreyB]

Well, I have to admit - the issue recently was empty fuel tank. For some reason I was so convinced that I had a full tank that I did not even bother looking at the fuel gauge which is actually totally operable. Anyway.

Warm-up PID control works good enough! On this log this logic keeps AFR at 13 while CLT is below 80C - and it actually does work.

https://svn.code.sf.net/p/rusefi/code/trunk/misc/logs/2003_dodge_neon/DataLogs/2016-02-15_20.55.45_warmup.msl

[puff]

[Number-One]

No more beer to your table