Iris rocket - Sudden vibrations triggered from battery failsafe (Audio resonance?)

Hello guys,

I was flying my IRIS quad on Loiter today when a battery failsafe triggered and the quad went into RTL mode as configured. Instead of going back to base, however, it started climbing like crazy with no signs of stopping. I tried going into stabilize but the quad still climbed. I ended up pulsing the throttle from zero to the middle, dropping the quad a few meters each pulse, until it reached safe crashing altitude - And crash landed (just a broken prop - woo!)

The iris was flying with a cell phone attached to a camera - about 270 grams. The phone was on flight mode (ha ha) so I don’t think it was a phone call or anything like that.

Here’s the log of the flight:

The log shows that right after the battery failsafe started the Z vibrations peaked for constant periods of time that fit well with the failsafe beeps - Could it be the failsafe beeps causing the Z accelerometer vibrations? Maybe it’s a resonance issue?
If not, what could have caused this?

The Pixhawk is a 3DR original Pixhawk, and it’s closed in its case. I haven’t tried looking on the Pixhawk board yet.

Thanks
Nitay

Could well be the piezo buzzer - is it close to the pixhawk? Note in the wiring wiki page it tells you to move it away from the pixhawk for this reason, it’s known to cause high Z-accel which results in the behaviour you’re describing:
http://ardupilot.org/copter/docs/common-pixhawk-wiring-and-quick-start.html#connect_buzzer_and_safety_switch
Here’s a similar problem:
http://diydrones.com/forum/topics/upgrade-to-3-3-2-alt-hold-rising-full-trottle-up?id=705844%3ATopic%3A2161709&page=2#comments

Yes, I review your log and as soon as battery failsafe kicked in there were high Z axis vibrations that sent the copter into full throttle up.

Mike

Alright! Not what I’d expect from a 3DR assembled drone!

I’ll move the piezo.

Thanks for the help and links!
Nitay