Lean AFR on certain RPM

[NeWT]

Hi all. Very special problem, I think. At first, here is my bin and log - ftp://murchik.net (2 files there). ER34 GTT - RB25DET NEO (no swaps).

I have a lean AFR at 2200-2300 RPM at cruise and on acceleration. But on the both sides of this problem everything is linear and everything is ok.

Here is accel:



As you can see, MAF is linear growing, RPM is linear growing, Inj Duty is linear growing, but AFR has a falldown.

Maybe it is some kind of a bug in ECU? I can say that fuel pressure is ok, I can see it on AEM fuel pressure gauge in a cockpit. Sparks and coils are new.

[AUSBen31]

Have you checked you're VE map to see if there is a low section at that exact RPM?

The Neo ECU is very sensitive to VE map.

[Matt]

Also try disconnect the MAP (boost sensor) to the ECU and disable inside software (See NEO tuning guide on our General Documentation section on the website)

This can cause issues I have found since it is used to reference the VE map but when engine characteristics change, the referencing of this table from Boost and TPS can be affected

[NeWT]

AUSBen31, VE map is almost flat. Moreover, I tried to enrich this zone on VE map, and there was no result.

When I was on a dynostand, this situation was like a misfires, car was twitching at this zone. But I don't know if lean AFR is because of misfires or misfires are because of lean AFR.
After that I changed sparks, coils and CAS to the new ones, and there was no effect. If to follow a theory of a bad ignition, maybe there is some problem in a physical parts of ECU? For example, some transistors or capacitors to the pins of ECU where coils are connected?

Matt, boost sensor is disabled inside software due to NEO tuning guide. But ok, I will try to disconnect it physically. But to be honest, didn't get your idea

Another theory from Russian forum is to check VTC map, but I'm not sure VTC can affect AFR so much.

[Matt]

VTC wont affect AFR too much

Looking at your log file, only AFR is going lean, whilst injection time looks fine. This seems to indicate that there is a problem outside the ECU itself (it could be misfire due to coil issue or otherwise for example)

If I saw injection time going lower (lean) at the same time then I would say tuning issue but ECU injection output appears fine

[NeWT]

If I saw injection time going lower (lean) at the same time then I would say tuning issue but ECU injection output appears fine

Thats what I'm talking about... But coils are brand new stock and I don't know where to look futher

[Matt]

I am not sure from here, this is more of a mechanical issue than ECU tuning issue. If injection time is steady and then you get this lean patch, do you feel it cutting (like coil or plug cutting)? One of the injectors not working?

[NeWT]

Yes, I feel it like coil/plug cutting. But why it is just on a certain RPMs...

[Matt]

Might be due to spark hot hitting the gap (one of the plugs) at that particular RPM

[NeWT]

This effect was on 2 new sets of plugs.
Meanwhile I tried to play with VTC parameters and that's it! When I touch VTC region by TP, this effect moves to another RPMs or to another load region. Didn't understood it fully, but I will continue to play. Theoretically, I need to set (VTC/TP) to (Stock VTC/TP) * Knew / Kold, I made this but result isn't ideal. I'm continuing to experiment.

[NeWT]

That's was not it. I am riding my Sky not so frequently. That's why I'm too slow to post feedbacks =)
I tried some more combinations and found this situation:
1) Riding at 1500 RPM - cruising - all ok.
2) Slightly pressing an acceleration pedal - RPM rises slowly.
3) Coming to ~2400 RPM zone - AFR leans.
4) And now, if I'll rise my right foot a little bit, to let an engine work on decelleration (or cruise), AFR goes normal! I'm absolutely sure with this.
I've recorded a log. Matt, please check it. Every time I go to 2400 RPM and then go to cruise/decel - AFR goes normal at once.
That becames 100% a tuning issue, but where?
All files are here: ftp://murchik.net/Nistune/

[Matt]

Why does your tune take TP all the way to 250 in the timing map? I would make fuel and timing map same maximum value in scales (208 for fuel and timing)

r34_1.PNG

(31.79 KiB) Downloaded 3485 times

AUSBen31, VE map is almost flat. Moreover, I tried to enrich this zone on VE map, and there was no result.

VE map is a TPS vs RPM table. It will only have an effect when the throttle position is over the cells (not load based map). So it is mainly to adjust acceleration from throttle position. I would not change this map unless you have to fix an issue with acceleration

Here we see constant RPM, MAF and resulting injection time calculated by ECU. However your AFRs keep going lean.... why is that? The ECU is commanding injection properly but your fuel at this time is leaning out. This is why I think there is perhaps a problem with fuel pressure or one of the injectors at this RPM/load area

r34_2.PNG

(26.69 KiB) Downloaded 3485 times

Another section, showing leanout with AFR but a steady injection time (orange) around 1.88 to 1.75ms following the decrease in MAF load (blue), then the AFRs start richening back despite no change in injection time. Strange

r34_1.PNG

(31.79 KiB) Downloaded 3485 times

[NeWT]

I mean, are you 100% sure that "Injection Time" showed by Nistune is really true? What if it's all about NEO ECU with it's speciality?
Talking about mechanical issues, I'm very careful in assembling my car, and all the intake was checked for leakages under pressure and soapy water.

[Matt]

I mean, are you 100% sure that "Injection Time" showed by Nistune is really true?

Yes it is definitely true... it is the result of the calculations used by the ECU and used for the time control of opening/closing the injector driver. Check it with a scope

A stuck or leaky injector, or issue with fuel pressure can also trigger similar issues

[unex]

if fuel system is modified and especially if fuel damper is removed, 6cyl engines often suffers for lean spots in some places, because of fuel pulsations. Even on this forum was not so old thread about that, so may be its the same issue.