Hi All.. first post
Just tested my newly built quad at my mates place - because he has a large garden surrounded by high trees.. Maiden flight for both me and it - and a great place to test it ........... or so I thought
All was well with my intended low flying test (with a couple of bumpy landings that a little trimming of the controls sorted out), then while (I think) pitching back.. all of a sudden it shot up into the air and seemed to have a mind of it's own not responding to my control - fortunately it landed back on earth after switching into RTL. The controls seemed to work as expected otherwise (yaw, pitch etc.).
Result was a broken leg, gps stick, 3xrotors, 1 motor cable to resolder, 1 GPS power cable snapped at connector, Repairable break to frame.
Still - I was so lucky! and I know it! I was in shock after lol - I've had visions of it going wrong racing through my head picturing what I didn't see Next time I'll appreciate the power of these things and get somewhere more open!
Anyway, I have downloaded the logs - could anyone interpret them for me to give me some sort of clue about what went wrong? They're attached as a .rar
Also got this out of the APM.....
just got this out of mission planner - still a little unsure of what looks like the problem here...
Log File C:/Program Files (x86)/Mission Planner/logs/QUADROTOR/1/2015-01-17 16-34-29.log
Size (kb) 485.2548828125
No of lines 7186
Duration 0:00:00
Vehicletype ArduCopter
Firmware Version V3.2
Firmware Hash c8e0f3e1
Hardware Type
Free Mem 0
Skipped Lines 0
Test: Autotune = UNKNOWN - No ATUN log data
Test: Balance/Twist = GOOD -
Test: Brownout = GOOD -
Test: Compass = GOOD - No MAG data, unable to test mag_field interference
Test: Dupe Log Data = GOOD -
Test: Empty = GOOD -
Test: Event/Failsafe = FAIL - ERRs found: FS_BATT GPS_GLITCH GPS FS_THR
Test: GPS = FAIL - GPS glitch errors found (1)
Min satellites: 0, Max HDop: 3.68
Test: IMU Mismatch = UNKNOWN - No IMU log data
Test: Parameters = GOOD -
Test: PM = GOOD -
Test: Pitch/Roll = FAIL - Roll (-168.16, line 6975) > maximum lean angle (30.00)
Test: Thrust = GOOD -
Test: VCC = GOOD -
Thanks!
Replies
You need your test pilots license revoked and your grounded till further notice.
First off when testing your first build rely on only the basics. Any automated modes need to be tested before you say that should work.
Second turn up the logging so we have good data to look at.
Ok, the battery voltage went below 10.8 volts which caused two RTL's to occur. The copter landed and you switch to Stabilize mode and then took off again. After a short moment the battery voltage went below 10.8 again and that caused the second RTL which was your fly away.
The yaw was acting funny when it went into RTL so it my be that the compass is not calibrated correctly which is always bad when using RTL.
Also ware you took off from and ware the first RTL landed were not the same place. HDOP was always bad for the hole flight so the GPS plot may not be correct.
Mike
Fortunately for all there is no risk of me flying today!
I have tested my motors and they're all noisier than before - I think it's the bearings.
Is this a fix I can attempt or do I need to buy new bearings?
i'd be surprised if it was bearings on all motors at the same time from a crash. Have you troed bench spinning one by one? Check for bent shafts or checks missing from props etc that might cause vibrations.
Just tried - all are the same - fine. Couldn't say the same about the props - 3 in the bin! it did end up upside down on concrete mind :-/
2 of the motors aren't too bad to be fair but one is putting some fairly bad vibrations through the frame. The 4th hasn't been tested yet though as one of the wires broke from the windings and needs soldering back on!.
I should have added that these are noisy motors without the props on.....
hmmm ok. my bet is a bent shaft without seeing it.
Hope you get her sorted
You mention controls trimming. don't use the trims on the radio with arducopter. Use the radio calibration and auto trim features if your copter is wandering in stabilise if its not the wind moving it.
Looks to me you had a couple of battery voltage fail safe induced RTLs. Your battery voltage dips below the 10.8V limit you have set, this is very slightly conservative for 3s, but not stupid low. Assuming yours is 3s.
Was the battery charged properly? Do you have any wiring between the battery and power module that could be high resistance?
Have you calibrated the current sense? Looks like you only used about 550mAh assuming it is calibrated. What capacity is you battery? looks like 3800 from the log data?
Your flight started without a GPS fix, if you are going to use RTL you should wait for a good fix before arming. Normally home is set to the position where you arm, but if you fly before a fix the home gets set to where it thinks it gets a good fix (if you were over a tree or something when this happened then hom is over that tree! :)), also potentially this could be wrong if the gps got a bad position, and if RTL occurs it will try and go to this erroneous home position. You can take control back from an RTL, but you need to spot it and change flight modes and back to the mode you want (ideally stabilise). Doing a pilot induced RTL in this case won't help as home is wrong. If you are paranoid you could do a pilot induced rtl and see if it goes where you expect before doing anything else.
Not sure this has definitely happened in this case, but seems very likely to me.
There are a few ways of seeing if you have a good fix before taking off: Using the LED outputs, easier if you use external LEDs rather than the ones on the brain itself, telemetry to a ground station, or if using a Frysky Taranis using a teensy (progammed, have a search for frsky mavlink teensy) to allow mavlink to be monitored and sent back to the taranis (gps fix info is included in the data and the taranis would need to be configured to show it).
My preference on a new build is to not have any automatic RTL enabled and have the radio receive failsafe set to kill it, and then not fly too far from you for the first flights. I will then go though the alt hold and check it looks ok, and the loiter being ready to switch back to stab. If loiter looks ok with some manouvering I then try some pilot induced RTL to check it does what I want. All of this checks that the compass, gps and loiter settings are good before relying on RTL. If the compass is bad this would also affect RTL (and other gps based modes), as it might go in the wrong direction for home even if home position is good.
Your log settings are too low to check compass though.
Hope you get it rebuilt. Some things to check carefully before flying too far from your self.
Thanks for the great advice...
The battery is a 3800 3s and the way it took off did look like it was intentional by the quad so the RTL issue sounds plausible.
The first time RTL engaged it landed pretty close to where it came from - I think it was turned on for longer before flying than it was the second time I launched and RTL engaged. That also fits with the RTL point being bad before take off.
I only intended to fly a few feet off the ground to get a feel of it - I got the shock of my life when it just disappeared over the trees. Massive field next time :)
I've been into the parameters and changed the logging to include IMU data - I think that is what you're suggesting the settings are that are too low!
IMU is good for vibration checks. Mag data relates to the compass.
Michael's response is harsh, but he does kind of have a point....if put rather abruptly.
have you calibrated your accels, compass and done a compass mot? if not add it the list before next flight. check the compass point where it should on the ground.
it says access denied when I click on your log.
I've attached them not zipped up.... thanks for looking (or trying!)
2015-01-17 16-34-29.log
2015-01-17 16-34-29.log.gpx
2015-01-17 16-34-29.kmz