Idle/low rpm suddenly going lean

Had the engine fired up, warmed up and idling and suddenly noted the AFR reading was reading lean (around lambda 1.1). Connected my computer to the ECU and saw closed loop multiplier was already at the max limit of x1.2 meaning there is an overall error of 30%!! Gave it a little gas to raise the RPM by around 1000 and the AFR and closed loop multiplier would become normal again but once I let it fall back to idle, it's lean again.

I thought maybe there is an exhaust leak and the sensor is reading wrong so I turned off CL fueling and idle. With CL off, lambda reading at was around 1.3 and I manually increase the fuel table at the idle by 20% and I could hear the engine increase in rpm. Given adding 20% fuel would increase engine speed, the engine was indeed running lean and the AFR reading should be accurate.

I started looking at my older logs and noted the lean condition had started days ago but the CL was sufficient to compensate until now (needed around 10-15% compensation). However, once the car moves out of idle and increase in RPM the lean condition disappears and fuelling would become accurate. I have looked at the actual pulse width of the injectors at idle and it has indeed increased.

This is a 2ZZ engine and I don't think the FPR is vacuum referenced and I cannot think of any reason way it would only run lean at low rpm. If fuel pressure was insufficient, it should be worse in the higher RPM range. Only other thing I have noted back when the AFR was normal, DBW idle was at 0.7 to 0.8% but now it is at around 0.5% (MAP also decreased by 10-20Kpa). However, the idle duty does fluctuate between drives previously and also since this is a not a AFM system, even if there was vacuum leak should not affect fuelling anyway.

Whats the VVT doing? Air temp similar? If the previous idle MAP was say 50kpa and it has now dropped to 40Kpa, then you should have something like 20% less PW, not more.

I know some of the ZZ engines have a baffle in the intake manifold that can come loose and do odd things.

Thank you Adam. Yes IATs were similar and VVT target/actual was at zero. Just rechecked MAP, I had the unit mixed up, it dropped only by 5 Kpa from (40 to 35 Kpa).

Yes the drop in MAP would result in a smaller pulse width but the CL compensation had countered that so overall the PW had increased. Injectors were running at 2.7ms during the worse case with compensation maxed out compared to 2.4ms when things were normal and it was still slightly lean.

Will be checking fuel pressure once I have an adaptor made.

hello can you see any other compensations going on i know it is extremely hot there at the moment so check the iat correction is set correctly

Hello Ross, IAT correction is around 1.5 to 2% for every 10C change. Intake temps had been around 35 to 45C the whole time.

I guess if I see an increase in pulse width as mentioned, the ECU is indeed correcting for the lean condition and the injectors are indeed opening longer and yet I am still running lean.

Maybe there is something mechanically wrong with the fuel system. But what makes it wrong only at low RPM (where it should be least stressed) but fine at high RPMs really baffles me. I have checked battery voltage just in-case and it was also stable.

Is it intake temp sensor in manifold? It may be heat soak related? At vacuum/low rpm the heat exchange between the charge air and sensor will be much less meaning the actual air density is higher than perceived by the system?

Higher temperatures can impact diaphragm behaviour a bit in a regulator too, operating near upper and lower effective flow limits of a regulator can be a lot more non-linear too.

Intake sensor is around 6 inches from the throttle body and does not feel hot to the touch. Intake temp is 45C max and IAT correction is only 1.5 to 2% per 10C so I doubt it would result in a 30% change in AFR.

Just tested the fuel pressure today and it was rock steady at around 350kPA so fuel pressure is not an issue. Managed to find access to an injector tester and noted roughly 6% variance between the injectors at 2.5ms. After cleaning them it got worse.......

One significant finding was the injector dead time though. Not sure whether it is the proper way to test it but I tried reducing the pulse width all the way and slowly crank it up. The injectors would only fire when it reaches 1.6ms (at 12V)which is significantly higher than the dead time set in the ECU (0.948@12V and 0.844 at 14V). I tried searching online for injector dead time data for the 2ZZ stock injectors but I couldn't find any concrete data but seems like it should be