CL correction/AFR targetting

[Mattt]

Hi all,

I've been a visitor to the forum a fair amount over the years, however on this instance I haven't been able to find any solutions or ideas.

Basic rundown is the car is a Subaru Liberty RS, running the G4+ V1-2WRX. Has a Link 4 bar map sensor direct to manifold and using an AEM wideband LSU 4.2 through the AN6 (I know not ideal, am looking to update to the CAN controller).

The car has been tuned and running well for the last couple of years. A couple of weeks ago went for a spirited drive and about half way along the ECU activated boost cut, I checked all the vacuum lines/intercooler joins and none had come off or split. At this point the AFR at light load went rich intermittently and it was generally a bit all over the place.

Since that day I haven't been able to find any physical issues under the bonnet and boost control is working again. The fuelling is still a bit wild at times, mainly at low load when CL fuel correction is working. On light cruise the AFR target and actual are not very close at times, this has been like that the whole time. Is this to do with the VE map being inaccurate? The idling AFR going rich is a new issue though. 

I've attached a couple of log files, hopefully someone here can make sense of what is happening.

Cheers,

Matt

Drive.llg Idle.llg

[Adamw]

its a bit odd how the MAP goes dead flat when the fan turns on?  This is where the lambda jumps off target also.  Not really sure if this is a symptom or cause since the idle becomes much more stable at the same time.  I wonder if the lambda is not reading right and it is actually much leaner than it shows.  Can you try turning off CLL to see if it idles better.

 

[Mattt]

Hi Adam,

thanks for the reply. I have attached a couple more logs, the one idling with CLL turned doesn't seem to make any difference. Theres also one with CLL back on and engine speed above idle (1800-2500rpm) incase theres any more info in there.

On normal driving when the CLL isn't in lockout it seems to be constantly incr/decr fuel and doesn't really settle close to the target AFR at times, is this normal or should it be more of a flat line?

If theres anything specific you need to try and narrow the issue down let me know, I'm at a bit of a loss at the moment. 

Thanks for your help.

Idle CLL off.llg log.llg

[Adamw]

Can you do another log at idle, CLL off, then in main fuel settings add a master fuel trim of +10%, leave it say 10 sec, then try -10% trim for a further 10sec.  

[Mattt]

Hey Adam,

attached is the requested log file. I logged the cold start also (sorry its about 10 min long) and it idled super rich until the engine temp reached 79deg and then it jumped back to 14 or so AFR, I'm assuming thats something to do with enrichment turning off at that temp? Unfortunately I stopped logging before this happened though so its not on there, I can log another cold start if it'll help.

Edit: I played around with the CLL limits while the car was warming up which is why its a bit wild on the log file. It took around 31% CL correction for AFRs to close up to the target.

Thanks,

Matt

 

Cold start 6.6.22.llg Idle master fuel 10%.llg

[Adamw]

Can you attach the tune as well.  In that log with the master fuel trim, did you do both +10 and -10 trim?  Im still not sure I believe the lambda is telling us the full story since it can change quite wildly with no change in injector PW.  Suspect a couple of cylinders are misfiring or something sometimes.  In most of the logs so far it seems to run smoothest when Eff PW is above about 0.67ms, even though that shows very rich at times.  

[Mattt]

I did both +10% and -10% with the trim off in betwen. I have attached a screenshot.

I also noticed the PW of injectors when I first started logging, being that they're 1000cc I wasn't sure how much extra or less PW it takes for a big variation in AFR.

Attached is the tune file also.

Thanks,

Matt

RS Map.pclr

[Adamw]

Can you update firmware to 5.6.8, I dont think that is the problem but the old 5.6.4 that you are running is from before my time so I dont have a good memory of any problems it had.  

Then with it idling with CLL off, can you try changing the "Min Eff PW" to 0.68ms.  This will make it rich as hell but so far in all your logs all the odd lambda control and hunting idle is always when the PW is below about 0.67ms.  

[Mattt]

Hi Adam,

I updated the firmware and set the minimum PW. I've attached a couple of logs, one with the thermo fan cycling. The engine has started idling higher than target also but that's probably a separate issue again.

Thanks again for your help,

Matt

Idle 2.llg Idle eng fan cycle.llg

[Mattt]

Hi Adam,

quick update on this and a couple of questions if thats ok. I ended up going away from the analog input for lambda and installing a CANLAM along with a CAN gauge. It seems to have fixed the issue for the most part, however the tune has seemed to go a bit out of whack. I'm seeing leaner than targeted AFRs when going into positive manifold pressure and accel enrichment seems to be lagging when the pedal is moved (going lean for a split second before richening). 

I'm going to contact a tuner and get it on a dyno for diagnosis but I thought you may see something abnormal in these logs that I can investigate. Theres no exhaust or intake leaks and from what I can see all the sensors are working. My only thought is maybe there was an underlying issue when it was tuned with the analog lambda input and now moving to the canbus setup its seeing different values.

Last thing, since installing and configuring the CANLAM the maximum AFR value has changed to 148 or so and I'm unable to find a setting to change it to a lower number?

Thanks again,

Matt

20.9.22.llg cold start 20.9.22.llg

[Adamw]

1 hour ago, Mattt said:

I'm seeing leaner than targeted AFRs when going into positive manifold pressure and accel enrichment seems to be lagging when the pedal is moved (going lean for a split second before richening). 

It may well have always done that, and now with a faster CAN lambda you just get to see more of it.  I suspect you might need a little more accel hold and possible accel sensitivity and clamp.  

 

1 hour ago, Mattt said:

Last thing, since installing and configuring the CANLAM the maximum AFR value has changed to 148 or so and I'm unable to find a setting to change it to a lower number?

You cant change it, with analog widebands they need to limit the range to try to get acceptable accuracy and resolution.  With CAN you dont have that limitation so the range doesnt need to be limited.  

For your cold start I would increase the CLL ECT lockout to say 50-60C so that closed loop is not working when its cold.  If you attach your tune I can probably give some suggestions to solve that big oscillation you have shortly after start up.

[Mattt]

Thanks for the reply Adam, I have attached a copy of the tune file.

Cheers

RS tune 20.09.22.pclr

[Adamw]

I made some small changes in your map below they may improve some of it.  

RStune20_09.22 idle and accel tweaks.pclr

[Mattt]

13 minutes ago, Adamw said:

I made some small changes in your map below they may improve some of it.  

RStune20_09.22 idle and accel tweaks.pclr 164.68 kB · 0 downloads

Thanks heaps Adam, I'll load it on tomorrow and see how it goes.