engine stumbling and stalling at idle when warm, new engine

[ericdand]

Hi everyone,

I just put a new engine from a reputable builder in my RX-8, and though it started easily, it's not able to idle right.

When it starts up, it idles well, but once the coolant gets up into the mid-hundreds and the RPM starts to come down from 1100, the problems start.

The problem, as seen from the data logger:

• When the car starts, the short-term fuel trim (primary sensor) reading goes quickly to -25%, occasionally coming up to around -20% for a moment then returning to -25%.

• When the car warms enough that the desired RPM goes under 900 or so, the actual engine RPM starts falling from 1100 down to the target of 800-something, and the actual equivalence/air to fuel ratio (lambda) starts to lean out from 0.9 to 1.0.

• While the RPM and AFR are changing, the measured short-term fuel trim, injector pulse width, and throttle position do not change. (side question: why are RPM and AFR changing, if none of the engine's inputs seem to be changing?)

• The engine RPM and AFR both reach their targets after 2-3 seconds, then exceed them. The engine suddenly starts to falter. RPM drops to 650 and AFR suddenly leans out to 1.2-1.3.

• After a moment, the engine catches itself: the ECU cracks the throttle open a bit, the short-term fuel trim (primary sensor) suddenly spikes up past 0% to 5-10% for a moment, and the engine RPM surges back up to 1100.

• The RPM starts to fall again the AFR leans out again, then they both exceed their targets again, and the whole process repeats.

• The problem repeats with a regular period of around three seconds: two seconds of the idle gradually dropping from high idle to low idle, then a second of sputtering and then catching and returning to high idle.

• Sometimes the engine fails to catch itself, and it stalls.

• The car also does the same thing if I give it a bit of throttle: the RPM just oscillates at the higher RPM (so no stalling, at least, but it's still doing it). For example, if I hold the throttle in to get the RPM to around 1700, it'll oscillate between 1800 and 1500 or so instead, in the same way as at idle.

• I took the car for a gentle spin around the parking lot, and it didn't run too badly. Not sure what the fuel trims were (no logging while driving around), but it stalled while I was coasting back into the shop.

I have attached three data log files. The first is a cold start, the second two are hot starts.


About the car:

• I'm on a completely stock tune (AFAIK), not yet used VersaTuner to remap anything.

• I have an AEM intake, so the VFAD is deleted (green plug not plugged into anything).

• The secondary air injection system is deleted: there's a block-off plate on the exhaust, and a rubber plug on the vacuum solenoid.

• There's no catalytic converter (only a midpipe), but the primary and secondary O2 sensors are both still connected.

I do have some suspicion that the ECU is just being confused by the AEM intake and the missing cat, and perhaps I should just drive it and let it learn. But I really don't want to do much driving before I know for sure that all is well.

Diagnosis so far:

• I did do the twenty-stomp ECU reset procedure before first starting the new engine.

• I suspected the throttle body, so I swapped to a spare that passes all the tests from the workshop manual.

• The spark plugs are new, and the coils are known to be good.

• There's a good electrical ground all over the engine and engine bay, including to the ECU. The connections from the ECU to the ignition coils are good.

• I found that my old fuel pump was starting to fail, so I replaced it with a new one. I have a fuel pressure gauge in the car (sensor teed at the red quick connector next to the engine), and the pressure is now a rock solid 60psi the entire time.

• The fuel injector rail holds pressure, so the injectors don't seem to be leaking or stuck open. But I can't say conclusively that the injectors are not faulty in some other way.

• I cleaned the MAF, and the data readings from it are as expected.

• I swapped the primary O2 sensor for another (used) spare, and the behaviour did not change.

Questions to help me learn:

• Does anyone have a good reference for learning about the warm-up and idle procedure that the ECU follows?

• In VersaTuner, does "short-term fuel trim (primary sensor)" indicate the ECU's actual fuel trim, or does it indicate the reading from the O2 sensor? If the latter, is there any logging of the ECU's actual fuel trim values, apart from the injector pulse width?

The real question: what is wrong with my car?

[0124alexrx8]

Have you checked for vacuum leaks? Exhaust leaks can do that, so do intake leaks.