Trying to figure out what happened

A couple of nights ago I was doing a long test flight of my Tarot 680 stretch (to about 810) to test a replacement of a faulty Taranis. I didn’t fly it aggressively, and about 11 minutes into the flight, I was at a point about 200 yards out from my location, and was hovering at about 18 meters, then I noticed an uncommanded descent. Right into the ground.

There was little damage, luckily. Actually, nothing that realigning motor mounts didn’t help, as it didn’t flip over until it hit the ground, it just acted like a fast descent.

That’s where you guys come in. I’ve pored over the logs and can’t find anything that would indicate what happened here. The auto-analysis provided no clue (no brownouts reported) and the only thing I can think is maybe a burst of 2.4ghz made the signal go goofy without actually triggering my radio failsafe.

Now a caveat…the power module doesn’t report correctly (I’ve used the old 3DR power module, but recently put an Attopilot 180 on it, and tapped the mains to provide the FC power via a Castle UBEC), so you’ll see some low numbers for a 4S battery. I use the Taranis telemetry with a voltage module that reads the battery voltage off the balance plug and use that for my voltage checking.

At the point of the ‘descent’, I was showing about 3.75v per cell and was preparing to return her home when this happened.

Can I get some of you log analyzing gurus to have a look at both of these (I have a bin and a tlog) and let me know if you can glean anything from them?

Thanks alot guys!

The bin file was too large to upload; here’s a link to it on Gdrive:

drive.google.com/file/d/0B1qWW8 … sp=sharing

The tlog is attached.

I guess I should probably add my setup:

Esc: 6X 30A Castle Quadpack
Motors: 6X Sunnysky V3508 580KV outrunner
Radio/Receiver: Taranis X9D Plus/X8R in 16channel mode Sbus with a voltage module.
Propellers 13"x5.5 Carbon Fiber
Power:
4S, battery HK Zippy 8000mah
Pixhawk is being driven by a Castle UBEC configured at the default of 5.1V
Camera gimbal being driven by a 12V/3A BEC.

The fault doesn’t seem to lie with the Pixhawk. See the attached plot of the period of your descent.

The red is altitude, the blue your throttle request to the Pixhawk and the pink is the Pixhawk’s throttle out request to the motors. As you can see, the Pixhawk was faster than you. As soon as the altitude starts decreasing the Pixhawk starts ramping up the throttle from 65% all the way 100% power and holds a 100% throttle request until it reaches the ground. Eventually you notice the descent and also request a full throttle output. But the Pixhawk was way ahead of you. The Pixhawk was trying to stop the descent but something in your power system wasn’t cooperating. So the culprit might be a bad motor connection, a failing LiPo cell or a flaky ESC.

OK, a quick followup. I think I see your problem.

The attached plot shows what the Pixhawk was requesting of each motor. At the time your excitement began the Pixhawk started asking for 100% of motor 4 (pink) and zero percent of motor 3 (blue). This suggests a failure in your motor 4 power system (motor, ESC or connectors). Because it was a hex it was still able to maintain some stability but you no longer had enough power left in the remaining motors to hover and down it came…but slowly., That would be my guess, anyway.

That’s some good analyzing there! I agree, that’s what it looks like. :slight_smile:

Thanks for the responses; that certainly breaks it down.

And the descent, while slower than a free fall, wasn’t real slow. It landed pretty hard, but luckily it landed upright and then the force of impact made it bounce and flip over.

Guess I need to look into motor 4 then. Funny is that I’ve flown it twice since then for about 15 minutes per flight with no apparent issues, but now it makes me nervous. :slight_smile:

Once again, thanks for the analysis.