Heli tips while trying to hover after upgrade to 3.3

I’ve looked at the log, but can’t see why the heli would pitch up as it did while trying to establish a hover in alt hold. It’s a really short log! Anybody?

[attachment=0]2015-10-30 10-16-47 12.bin[/attachment]

Here are the plots of desPitch, C2 (RC TX pitch input), pitch, and throttle in. I think desPitch means “desired pitch” and is produced by the autopilot… why would the autopilot set this attitude? I thought perhaps the autopilot might be trying (incorrectly) to stop motion in the forward direction by pitching up abruptly, but IMU telemetry in the log doesn’t support that. Other ideas?

vibrations seem “normal”. RPY follow closely to desired RPY, so I don’t think there were any mechanical problems. IMU & IMU2 seem to match closely. It seems the Pixhawk deliberately tipped the copter up… but why?

[attachment=0]telemetry.JPG[/attachment]

Solved – the elevator channel was apparently missed during radio calibration. The max value was slightly above 1500 and the min slightly below, no doubt caused by a slight perturbation when calibrating the aileron channel. This made the heli hyper-sensitive to pitch input – and apparently resulted in A REVERSAL in forward/back behavior – which made the logs a bit more difficult to interpret. After re-calibration of the radio, the heli flies fine.