Thorny hotrod 3.2 stumble / hesitation issue

[thetorch]

My friend has hot rod 3.2 that has a persistent stumble / hesitation, most noticeable under load, that starts right around 2800 RPM, with a 100 RPM +/- stumble before smoothing out. It sometimes stumbles at 3200 or 3800. Happens under light, medium and full throttle.

Once past the stumble, car drives fantastic: AFRs looks good in all ranges, no pops, stumbles.

Picture of a typical stumble log readout at the moment of hesitation is below. I am not an EFI expert but learning. We drafted a longtime Porsche hotrod tuner who has extensive EFI dyno tuning experience to help figure it out, and have tried many different means of narrowing the issue down. No luck so far.

Car is new to the owner in the past year, and it's had the hesitation since he bought it.

Out of ideas at the moment. Feels like a transient fueling issue, but open to ideas.

From logs:

• TPS, Spark advance, VE values, battery voltage, MAP, injector PW all seem to show consistent, expected readings during the throttle ramp.
• RPM stumble is visible in the log, along with a brief AFR spike.

The engine:

• Twin plug 3.2.
• Megasquirt 3 Pro ECU
• 6 dual Electromotive coils / EDIS
• Triumph ITBs w rain hats
• Non stock cams, Dougherty Racing cams, GT2 EVO spec: http://www.drcamshafts.com/911profiles.htm
• SSIs with 2 in 2 out (M&K?)

What's been checked / tested.

• Fuel pressure under load on a dyno. No problems.
• Crank position sensor: reset the gap to trigger wheel. No change.
• Balanced and sync'd ITBs at idle and at 3K RPM.
• Intake leaks, none found with carb cleaner method.
• Plugs. None fouled.
• Wideband PLX: recalibrated to ECU. Seems fine and reads consistent.
• Fuel map is ITB (MAP / TPS hybrid). Tested adding and removing fuel to the VE table in the area of the stumble, just to see if it's a poor/rich/lean map. No change.
• Tested smoothing/flattening the ignition map in the area of the stumble. No change.
• Changed Ignition map from Speed Density (MAP based) to Alpha-N to rule out MAP fluctuations causing ignition fluctuations. No change.
• TPS sensor replaced and recalibrated to pedal. No change.
• Coils tested by removing one plug wire from each dual coil at a time. Coils are paired cylinders so as you would expect, no change.
• Tested with fuel overrun cutoff on and off as it feels like a quick fuel cutoff event. No change.

Blue vertical line is start of the stumble, white line is RPM, upper green is TPS, yellow shows small AFR spike, upper red line is MAP. Some other data included. Hope it's readable:

[thetorch]

Posting the solution on this, in the event it might help someone else in the future...

Friend took it to a Megasquirt knowledgeable expert. Apparently the firmware was the problem. He updated the MS firmware and the problem went away.

The mechanic suggested the components degraded and the firmware could not handle some increase in signal noise, and the new firmware improved signal noise filtering.

Not clear to me what noise specifically (crank sensor, tps, ??) or whether it was the ECU itself degrading. We never found anything the ECU was logging to indicate the source of the issue.