I'm in Raleigh NC, thanks though. I'm checking around locally to see what I can come up with.

Well it wasn't the ecu. Picked up a used ecu from a local racer. No change.

So I'm really at a lose for ideas now. What about the temp sensor for the ecu? The car clearly warms up in an open loop and then transitions to closed up as it warms up. So it would seem the ecu knows the engine temp.

Ground issue? Not sure what to check.

It's also not the EMB. I disconnected the EMB and took it completely out of the loop. I feed the NB O2 signal directly into the factory ECU like stock. Naturally it runs much richer as there's no compensation for the injector size. Idle is around 11.2 AFR. Light throttle cruise, it will suddenly jump to 14.5-15.5 AFR. A few seconds/mins later, it will come back down to 11.5. So same functionality, just richer because of the injector size.

With the EMB taken out of the loop, I couldn't log the injector duty cycle, but I was able to log the AFR and fuel pressure through the Zeitronix unit. There is zero change in the fuel pressure when it leans out or when it richens up.

To be honest, at this point I know exactly when/where it will go lean. Cold start, idle, and light cruise leaving my neighborhood, everything will be fine. Pull out on the main road, normal acceleration, and about a second after I shift into 4th gear, it will go lean for the first time. After that it will go lean at any speed/gear/rpm.

miataforsale

Factory NA sensor would be the next step IMHO. BTW, can you feel it go lean or is it only the wideband telling you it's going lean?

I'll throw this out too. During the drive without the EMB, I got a check engine light. Codes 17 and 26.

17 Oxygen sensor unchanged after 1500 rpm

26 Solenoid valve (Evaporative canister purge)

I assume 17 is because it is running so rich? The O2 sensor never sends a lean signal?

** So I decided to check the O2 sensor output with a voltage meter. Car running, ~11.4 AFR. {I suck with multimeters} Set on 2000mV (2V right?) I get a reading of 004

That equates to 0.04 volts right? But I should be more like 0.90 volts at that rich AFR, right? So maybe I do have an O2 sensor problem??

So... I think I may have fixed it. Based on my last post I became suspicious of my NB O2 signal.

Two key points here. I previously mentioned 1) I suspected the PO had wired the 4 wire O2 sensor incorrectly. and 2) I ran a new signal wire and a new ground wire.

I did not however check the O2 heater wires. I made the assumption these were right. Classic mistake.

The PO had wired both of the heater wires to 12V. That's incorrect. One is a 12V and the other is ground to complete the circuit. Also, both wires had pulled out of his electrical connector, so they weren't even getting the 12V he had intended.

So I corrected that and added my EMB back into the loop. A quick test drive revealed no more random lean condition. I've still got some tweaking to do but hopefully this is behind me and I can move on to what I had originally intended to work on 2 months ago... in-boost tuning.

sigh, it's been a long night.

Good to hear.

Wow, that was a convoluted path, good job getting to the answer, though!