Lean spikes

Hi,

Car is a 2001 Mazda MX-5, located in Canterbury, New Zealand.

ECU is a Megasquirt 3, and the wideband is the AEM UEGO 30-4110.

Here's the detail:

Recently I've started driving around and getting sudden lean spikes. This is absolutely something that can be felt all through the car. The transmission shakes like crap (visible in the gear stick), and the engine clearly misfires. My own expectation would be that if it goes lean, it's probably a wrecked gasket causing a vacuum leak somewhere after the MAP sensor line, or a damaged injector (or injector harness). I've replaced all the appropriate gaskets, so it isn't that, and checked the injectors (including completely rewiring the injector harness itself). I also replaced the fuel pump and fuel filter (mostly because it was due, but it seemed like a good idea to eliminate this possibility). It doesn't seem to be any of my original theories.

Naturally, I try google. Which gives me all kinds of response, including the potential for false spikes in either direction from the AFR gauge. I doubt it is this, but AEM don't do free air calibration so I'm not exactly certain.

Now, from here I realise a couple of things (been ongoing for at least a couple of drives at this point):

Since changing the injectors (which worked absolutely fine, with no issue, for at least six months!) I've also changed the ignition system - new coils and sparkplugs.

The new sparkplugs are in a much higher heat range (intended to swap over *after* I had turbocharged, not in the weeks before).

The issue *only* occurs after the car is at full temp.

Perhaps I skipped a step in my understanding of combustion when I overlooked the finer details of ignition. Could this be the cause or should I keep looking?

James

The AEM gauges can give false readings when the sensors are failing, this doesn't seem to be the case with yours as you are getting misfires.

Just so I'm reading it right, the car is still NA?

The plugs, are they a grade higher (usually colder) or a grade hotter (usually a lower number)

Could you grab a datalog and post it up?

Whats your mat curve like, what are your iats on average? Have any logs?

I have found that the MS MAT correction table is VERY stupid, and had to completely redo it to make it work properly. Thats for the climate here tho. And it never showed up as lean spikes, just random richness, lean, etc. Go out and do a few datalogs, and then post them and ill do my best to help you with them. Im really trying to get good at megasquirt. Im finding it to be the most troublesome ecu to tune. Haltech, VIPEC, Link, ive had no issues with. Megasquirt....very tricky.

Talking about the plugs the difference is ik16 stock to ik22 in there at the moment. So multiple heat ranges in spacing and now that I check quite a significantly different gap too...

Will post logs when I can get home.

Easiest way to view it is to have AFR and AFR error loaded. You'll quite quickly see exactly where it's gone off-track.

Can't read the log on mobile but check your injector duty cycle when it happens , sounds like it might be an ignition fault to me

To follow up with the (suspected) solution, it was electrical.

I believe it was induction from the ignition wiring interfering with the wideband signal.

Thats good you got some resolve. Wiring gremlins are a pet hate