Close Loop Leans out

[LittleH]

Hello guys , I get a weird behaviour with my s300. Have it for quite a few years now , no issues everything worked perfect.
We had a fire in the engine bay , but we replaced everything in terms of wiring and fixed it up (as it was)

We tuned the car open loop works amazing everything is almost spot on for the map.
We tried to enable close loop as it was (last settings) and basically the car went super lean 19+ mixture.

We use MTX-L plus wideband , s300 reads it properly and its set up properly (same values on s300 and gauge)
Its on pin EGR D12,

Now the weird thing is , it does it only on idle and light load (it messes up with the AFRs) but after the car moves a couple of meters seems to be normal again , but it will still re-do it. The ecu itself seems to be fine , the wideband seems to be working as intented.

Also the L.TRIM is 19% (even in open loop)
S.TRIM 0% (which is correct)

when I tested with close loop the short trim did kinda work but I think it was a bit laggy (per say) , I don't have a log from that because it runs really rough , but I can get on that to provide.

Has anyone had similar experience with that? and if so , how did you solve it?

I use the latest stable sManager version , but I just downloaded the beta one , to test that as well.

[Spunkster]

Post the calibration as well as datalogs showing the issue.

[LittleH]

I did a datalog but looking at it , the CL remains OFF (in the graphs) which is very odd. So that will not determine a lot of stuff.


What I did notice is that after I turned it ON , the AFRs leaned out, map went to 433mbar and rpm dropped to 610.

After I enriched the table. (It needed quite a lot of more fuel) it works great , need to tweak with the settings of the wideband but that is fine.


Here is the datalog with the cal (dont mind the maps it was before smoothing it out and fixing it up , just for testing purposes)

And the weird thing is that now that I have both maps (close loop tune as well which is fairly finalised) it need 30~% more fuel

Also it affected areas on the maps that the close loop is disabled based on the conditions , which is really really weird. and had to trim it a bit for example at wot on vtec needed 4-5% more fuel.

The map is the first one after a fire that we saw the problem not the final one. just to get home.

ECU was checked and it seems fine.
Wiring (used but in excelenet way)
Wideband , we replaced the wiring (between where the sensor connects to the gauge/controller) (MTX-L plus) with another from innovate. the sensor it self is fine , tested that as well.

Now the latest tune works fine , since I added quite a bit more fuel , just wondering why it would affect it so much and also affect parts that were out of close loop's reach.

[LittleH]

So small update , yesterday went out and did some testing , sensor threw an E8 and the car just was off its mind. , I had to put back the open loop tune and then fix the sensor (which I did) and now its working again perfectly!

I have programmed my wideband to output 2.5V when throwing an error so it wont affect the tables and it worked amazing many years.

Quick question, the Inactivity timeout does that turn the close loop on and off?
and can I disable it? for example 0s or something? I think that might cause an issue but not sure.

[LittleH]

Also something I forgot to mention , is that not sure if the ecu is intact after the fire (everything seems to be working so I think its fine)
ELD is enabled , also we did a COP conversion with doctronik coil on plug , normally it would not affect it but not sure the circuit inside the ecu if it affects it. (it was suppose to be factory replacement but again not sure!)

[Spunkster]

What injectors are you using?

[LittleH]

I am using 440cc RC Engineering Injector SL9-0440 , they are checked and cleaned , they are matched 100% and they don't have many kms after cleaning and checking them.

[LittleH]

So quick update.
I noticed that I do use the wrong dead times for my injectors.
Used the correct ones (added ~45% more fuel) in order to get it stoich (@ idle) , that worked out fine.
Also used new wideband LSU4.9 (Innovate MTX-L plus)

But behaviour still remained the same. once enabled close loop -> leaned out , tried moving around a bit but nothing.

I think maybe the problem is with Inactivity Timeout. (I get why we need it) but with innovate you can smartly bypass it , cause if it throws an error you can program what kind of voltage to output -> (Have that setup at 2.5V) in order to not have any weird changes or issues , that's how I did it before and it worked flawlesly , although that does not work with AEM widebands for example...


Now I fixed small table in open loop (stoich + safe).
But not sure, I would like some ideas on what can go wrong?
Should I take my ECU for a checkup? does anyone else have the same issues? (D12 pin EGRL)

[LittleH]

So another update. Tuned the car with proper dead times, everything is good.
Now when I enabled close loop , it leaned out (not completly) , 18~ afrs (IN Display the CL flag was still not on)
after whack of the throttle and returned to idle (CL flag enabled) (not sure why)
It seemed to be working but the weird thing is. the fueling seemed to changed (on vtec as well) which should be disabled the close loop)

Also Open loop idle AFR (14.7) , close loop idle afr 14.7 (but in display it showed correction of 33~%?)
Which is super odd.

I am really puzzled about all of that tbh , not sure what is happening. Thought about being something hardware (but that does not make sense in my brain)

Maybe the new wideband timeout caused some weird bugs? Could we have an option to have that enabled or disabled? (In my case it does not matter cause I programmed the output of the wideband in error to be 2.5V) so it won't affect the tune.

but I am SPECULATING , to me it seems something is up software-wise.