EJ207 Factory Wideband vs Innovate Wideband

Hi,

I have just completed the physical swap of an EJ207 V8 into my USDM 2004 WRX. Following startup I am trying to debug a hunting idle. I have smoked checked for leaks from the turbo inlet to throttle body and also checked the post throttle body and confirmed a good seal. I replaced the factory sensor with a new unit for the USDM EJ205 (Denso 234-9011). Perhaps a bad idea and was based on looking at a few ej207 threads in nasoic. I also have an Innovate Motorsports 3918 MTX Series MTX-L Plus Wideband Gauge that has been open air calibrated.

In order to get a better idea of what was going on I logged the engine at several rpm settings in neutral. I see what I would think are some significant differences between the ECU sensor (A/F Sensor #1) and the innovate but really do not know how closely I should expect them to track each other. I have attached a typical log file. I expect that I need to get an EJ207 specific AFR sensor but would like to get confirmation of that here.

TIA

A hunting idle just indicates an unstable torque generation. At the lowest RPM, you are making too much torque and the engine accelerates, at the maximum RPM, you don't make enough torque to maintain the engine speed. Often just adding a bit more fuel (perhaps targeting Lambda 0.95-.96) will help stabilize the idle, or reducing the ignition timing may reduce the torque preventing the first "surge". Is there an idle control system adding to the mix?

One of the problems with looking at an oscillating lambda trace is there is a delay between the combustion event, and the O2 sensor detecting the excess oxygen. This delay is increased with slower engine RPM (and the farther the O2 sensor is from the head). Also, with the engine surging, it's not in a steady state condition (Map is changing usually), so individual cell fueling changes likely aren't needed, if the fueling table is smooth in the surging area.

A EJ207 with the lambda in the downpipe can have a transport delay between the combustion event and the lambda sensor reading that event of greater than 2 seconds, so unless your tuning software is taking this into account, then you can very easily end up chasing the fuelling and making the issue worse. The firing order with the standard headers also can skew the results, as the sensor doesn't see a consistent pulse of combustion gases, rather more of a surge and ebb pattern.

I have found that the fuelling around the idle area should be quite flat with these engines, so if it is lumpy then this will exacerbate the issue. They do prefer to idle on the slight rich side as well, mine is happiest idling at 950rpm and 0.96 lambda on E85. The timing ranges around -10 tp 15 degrees. i also have the TGV's disabled.