Hi I’m currently seeing weird lambda spikes paired with a weird async pw spike.
What is causing this? The async pw spikes up to 800+ms for a split second and lamba shows 0,25 for that time, after that lambda looks normal again.
I can provide log and map if needed.
I have to recheck on the weekend. But it’s a separate K-Type sensor. I can’t remember where i connected it.
It shouldn’t have anything to do with the Lambda. (would be 4.2 and directly connected if i remember correctly.)
I also see this behaviour but with lambda reading spiking to 63?
This seems very strange to me, either the EMU is behaving wrong or the log is wrong?
The logs are from my EDL-1, i will check if i have the latest firmware flashed to that.
Please attach orginal edl log file. What firmware do you use ? In the past there was and error in async enrich that can cause wrong injection time.
Actually I tested your project on the bench and EGT readins are stable. However I did use precise voltage source as EGT. I will perform test with the real EGT sensor.
i have managed to get a egt to can controller and ill test it in the next few days and tell you more feedback. i have also tried different ecu and a different egt sensor and i still have the same outcome,
when test the sensor to a standalone gauge it works fine
I’m running client 3.055 Exp.2 and EMU Black firmware 3.055.
My EDL-1 is on 1.21.
Log containing the problem is attached.
03192836.BG3 (6.8 MB)
Loving this update. DSG seems so much better with combined load. I also don’t seem to be getting the near stall issue if I coast off throttle a little and then try to accelerate. I suspect this is the change to PPS that fixed this.
I have a couple more suggestions for you.
Can you please add a minimum RPM to the launch control so that it doesn’t enable ignition retard before a minimum RPM? Currently if I floor it, I get a random backfire and slower RPM increase due to the ignition retard being enabled. Allowing a min RPM of say 3000 will allow revs to shoot up quickly, then retard can kick in till my set RPM of 4000.
Secondly when reviewing the Graph Log, and moving the mouse around on the log all the other active tunable maps track with that mouse movement. Love that feature. But if you could add a Click function to select the active cell in the cross hairs that would be great. It would allow me to jump straight to fuel or ignition maps and +/- the selected value instead of clicking it.
Cheers,
Ian
Ah no. This was EMU Software 55 a or b. Noticed it while tuning.
I will attach the log soon.
This is a cheap K-Type Sensor. Could it be faulty? As soon as the engine is off it reads constant values without oscillating i think. I will retry on the weekend.
Btw: is there an output Test where i can fire up my injectors for like 10sec to fill up glasses to see if all of them deliver the same amount of fuel?
I found the cause of the my old problem of oil smoke from the exhaust. The engine was running lean on fuel... to run with little smoke (0.84),
I modified everything possible and found a half-solution by changing the trigger and dead time Injector on the VR6 BUB.
I'll share the logs with the original trigger and the modified one later.
I’ve just seen there is a log channel about edl write errors. Every time I see the weird spikes in the log, it says edl write errors. So I assume this is more of a logging error (EDL, SD card etc.) instead of the emu behaving wrong. I would also assume I would feel (even if it is very short) a lambda spike to 0.25 while crusing..
You can use Functions to do that. Connect switch to switch input and make a function that will trigger Injector output for 10s after the switch was pressed.
Your observation is perfectly correct. It is a logging issue (problem with sector write), not a lambda issue.
You can try to use different sd card.
Here is a list of supported cards with their performance:
The only issue that comes into my mind is some internal connection of the sensor to sensor inconel enclosure. You can try to remove it from exhoust manifold, start the engine and check if there are still oscilations (make sure the sensor doesnt touch any metal parts of the chassis).
In the meantime I will make some test on the bench with the real K-Type sensor.