AFR weirdness

[elaw]

I'm helping a friend tune his MS3 setup and have noticed something weird with his AFR readings, I'm wondering if anyone has any ideas on a cause?

A log file edited to show the problem is attached. If you look at the AFR you'll see it's a bit noisy but also has periodic stepwise "jumps". A lot of the steps last around 9 seconds which to me sounds like the cycle time of the radiator fan (the issue is most noticeable at idle). An example of this in the attached log is the period from 114.933 to 124.511, then the following period up to 132.942.

TPS, MAP, and PW are all stable during these events. EGO correction varies but isn't changing in steps like the AFR reading is.

My thinking is it's 1) variation in the supply voltage to the injectors or 2) variation in the supply voltage to the wideband controller. But my friend tells me (I've never seen the car in person) that the MS and the injectors are all powered from the same relay, and the Batt V trace in the log is absolutely steady. And I'd have thought a wideband controller could tolerate some voltage variation without doing this? He also assures me the grounds are proper... the MS has multiple grounds to the engine, and the O2 sensor heater is also grounded to the block at a different point.

Does anyone have any thoughts on this?

graph.png

[Matt Cramer]

Exactly what sort of O2 sensor is he using?

[elaw]

I've asked him and will post back when I get an answer. I'm about 80% sure it's an AEM.

[ol boy]

Place a DMM on the power supply to the injectors and see if it's stable and matches the value shown in TS.

Sent from my SM-G920V using Tapatalk