Discussions

I am posting this for information purposes, as I had only discovered I was running 2.9 yesterday and not 2.9.1 as I thought, but these incidents seem similar to some other unexplained crashes of late.

This hex was being used to trial camera isolation mounts of various designs, and as such had probably clocked up 20 to 30 flights on FW2.9.

The mission is simple, a 4 point route around the back paddock at 20m using the fence line to check tracking of a downward pointing camera taking video so I could see the shake rattle n roll going on.

Day 1. 

Placed hex on launch pad and take off looked fine until it reached height, then started towards WP 1 at clearly too higher angle and started the toilet bowl manoeuvre. Switched to Stabilise and then back to auto and the mission went fine, including the usual lovely auto land.

These missions are repeated 3 times.

Number 2 started the same way and shot past WP1 at high speed and excessive angle that was going to end in a crash. Back to stabilise and I brought it back towards WP1 and switched to auto, where it did a strange circle but completed the mission and another fine auto landing (just love those auto landings).

Number 3 takeoff was normal as were the previous 2 but the hex immediately went to 40deg WOT at 90deg to the correct heading, so I did not let it get far before switching to stabilise. I swung it around back towards the launch point and switched back to auto, expecting it to continue the mission. It didn't, just pitching hard and headed back in the other direction, the one it took after launch. So back to stabilise and bring it back, then back to auto to try again. At this point I am starting to wonder what's going on, but I had changed nothing from the previous 20 odd flights.

Once again it headed off in the wrong direction, a quick change to stabilise and started to recover. This is where things went south. 

It did not want to respond to commands, or rather it did, but then fought against the input, almost like it was auto stabilise mode (I have all that switched off).

The hex then went to full power and did a big toilet bowl manoeuvre into the ground at full throttle.

I had to change modes to get it to shut off the motors - out of stabilise and back in to stabilise, throttle was off before it hit the deck.

For some strange reason it had switched to 'Land' in the final part of the mission. No valid reason.

1 bent arm, 1 buggered motor, 2 props, 2 legs, afternoon to rebuild.

I am using a suspended plate for vibration dampening and I found one of the 'O' rings had broken. So I put the errant behaviour down to this, but was still a bit puzzled by the uncommanded Land.

Day 2.

A short test fly to check everything, stabilise good, alt hold good, loiter good.

So the next test was a mission. This time without a camera just to be on the safe side.

Take off was normal, as usual, and the hex headed off 'sort of' in the direction of WP1 but at a much too great an angle and speed (I have the mission speed set to 3m/s).

By the time a switched to stabilise the hex was getting away a bit and took a bit of distance to get it turned around, but once it was heading back I decided to use RTL (tried and tested many times in the past) as it was on the limits of distance for orientation.

The hex immediately swung around and headed off in the wrong direction, away from me. Back to stabilise and I knew I was in trouble as I was fighting for orientation and control and was loosing while the hex kept getting smaller and harder to ascertain what my inputs were doing.

If you check the kml file you will see how long I fought until I was positive I could not regain visual control.

I had nothing left but to try RTL and wait and see where it thought home was, and plan the search while I waited.

But RTL did bring it back to within visual orientation distance, and so, when it started to circle off in the wrong direction again, but it had reached the adjoining paddock, I took control with stabilise and started bringing it in for a landing.

As I brought it in thought I would try RTL again, as it was heading in the general direction, and see if all went well. It didn't, and shot off in the opposite direction, so once again to stabilise and bringing back to land manually.

Then it happened, again.

Another uncommanded Land, full tilt, full noise, toilet bowl into the ground. It seemed to respond to inputs at first, but like previously, overrode those inputs with even greater roll or pitch, almost like it was in auto trim mode.

Throttle was off, mode was stabilise but I could hear the motors squealing on the ground. I had to switch out and back in to stabilise to stop the power to the motors.

Same damage as day 1, back to the workshop.

It was while downloading the logs that I noticed I was running 2.9 and not 2.9.1 as I thought, so as I have seen other strange happenings reported on the forum, I thought I would add my tail to hopefully add to the collective knowledge.

I was not running telemetry at the time but I am including the log and kml files, APM 2.5+, Futaba TX and Rx, updated ArduPPM, virtually stock params.

APM on suspended board with Futaba Gyro tape, copper mass plate, and velcro. Z axis vibration <+/- 2

If anyone can spot anything in the logs and tell me 'you had that wrong' I would be grateful.

2013-02-20 08-37 2.zip

2013-02-21 15-48 3.zip