Minor thing, but possibly related:
In the log mon.xs, at 564,5 seconds, you can see pulse width nearly doubled, followed by a cough the whole car felt. Pretty weird. This was after a short period of constant throttle and gently falling MAP.

The "main event"
At 716 you can see RPM take a small dive (input circuits? It's hopefully another topic, though the timing error is very small, inconsistent with the times I have had input issues) what's weird, and this exact thing happened before (in the other log file) where it goes pig rich (not quite full scale, but darned near it), makes very little power and stays for a while. In the earlier scale, it stayed for a while (I was in traffic and just wanted not to get run over). This was only 8 seconds, but in the other log it was quite long, and even occasionally came out of that state only to return a few points later.


Is this a software thing? I was thinking it was an injector sticking, but why does it always start at just above 5,200 rpm and why does it "fix" itself?

This is a 2000 miata, turbo (but I don't think boost is an issue) with used Lancer Evo injectors (fairly new), running MS-IIxb18. Thanks!
-Abe.

 

I'm sorry, I saw the title of your thread and thought "Paris Hilton?".

C
[/thread jack]