I put a GoPro on the front of my F450 as a test to see what the video would look like. I noticed that the quad wasn’t flying as stable as it normally does so I quickly switched it over to loiter mode while I assessed the situation. It didn’t feel near as stable as it normally does and had a lot of drift in all directions. Was very difficult to control. Then all of a sudden it took a huge pitch turn away and crashed against a house. Fortunately not much damage, however I am trying to figure out if it is an issue with having the gopro so close to pixhawk and GPS or something else.
BTW - the wifi was turned OFF on the gopro.
There is one weird entry in the log that leads me to believe that the issue is with my GPS. However before I replace my GPS I would love to narrow it down to a gopro issue or not.
I see that up until entry 17813 I had an hdop of 2.45 on average with 9 satellites. At 17813 hdop went to -298 and #sats went up to 27.
GPS, 1, 0, 64707, 27, -298.90, -197.5621317, -112.8533956, 7171827.35, 796004.59, 27473421.78, 12179557.33, 3.747091E-29, 132857430
I think this tells me a GPS issue? Is this what caused the crash?
What caused the instability before the crash?
Any help is appreicated!
Here is an Auto Analysis of your data flash log using MP 1.3.23:
Size (kb) 1445.6875
No of lines 19526
Firmware Version V3.2.1
Firmware Hash 36b405fb
Free Mem 0
Skipped Lines 0
Test: Autotune = UNKNOWN - No ATUN log data
Test: Balance/Twist = GOOD -
Test: Brownout = GOOD -
Test: Compass = GOOD - mag_field interference within limits (24.11%)
Max mag field length (624.74) > recommended (550.00)
Test: Dupe Log Data = GOOD -
Test: Empty = GOOD -
Test: Event/Failsafe = FAIL - ERR found: CRASH
Test: GPS = WARN - Min satellites: 6, Max HDop: 3.5
Test: IMU Mismatch = WARN - Check vibration or accelerometer calibration. (Mismatch: 1.15, WARN: 0.75, FAIL: 1.50)
Test: Parameters = FAIL - Bad parameters found:
THR_MIN set to 300.0, expecting less than 200
Test: PM = FAIL - 8 slow loop lines found, max 13.10% on line 4616
Test: Pitch/Roll = FAIL - Roll (147.39, line 18954) > maximum lean angle (45.00)
Test: Thrust = GOOD -
Test: VCC = GOOD -
Your HDOP was very high
THR_MIN is above min range
Pitch/Roll way beyond max allowed
Thanks much for the analysis. My assumption was that the Roll issue was from just before the crash where the copter rolled over to the side and headed right into the house.
Does anyone know what the THR_MIN is? What does this setting do? What are the ramifications of it being too high?
Is there anything that shows why there was instability before the crash? What caused the crash? Is it because of the high hdop?
Its a weird one, not sure whats going on…the gps number of sats shoots up to 25 momentarily and the hdop goes negative gps altitude spikes to 800 ish metres at the same time…very weird. this is a little while before the drop in altitude…so not sure if this is it or not.
vibes look quite good, motors look like they are loaded up equally.
There are confirmed issues with GoPro (measurable) emissions that can swamp the GPS. I now wrap the back and sides of the GP3B with foil. Some GPs are worse than others. So it’s something that is dealt with on a case by case basis. The emissions don’t have to be eliminated, but reduced to a “tolerable” level.
On another FC and before I had a GCS available I could verify the issue by being unable to obtain GPS lock.
I’m not sure how that value got so low. What does it do, what can it affect? Is it related?
Your issue is GPS Drift. You were too close to the house and the GPS signal bounced off the roof and caused the GPS to drift into the road way. Your copter then corrects for this and flies into the house not knowing it was not moving.
Look at the GPS plot of your flight and just before it crashes I see the GPS moved North West into the road way but you say it crashed into the house…
Don’t use GPS modes when in a closed in area, such as tall trees and houses. I have had a couple of close calls doing that.
Also if you fly in stabilize mode thinking your ok, make sure battery failsafe doesn’t kick in as it uses GPS and will also cause bad things to happen.
Even the Land mode has issue with bad GPS data.
I like to test everything on the ground under worst case scenarios.
Boot up the APM and connect telemetry open mission planner with flight data screen.
Blow really hard on the copter, look for alt drift.
Shine a bright florescent flight at the copter, look for alt drift.
Get a good GPS lock, put your stuff next to it, look for position drift.
Swear at it, make funny looks, etc, look for position drift.
Run up motors, check magnetometer.
Etc, etc, etc. All under controlled conditions first.
Thanks all for the help! I have some work to do it seems.
Two open questions:
- What does having THR_MIN so high do? Is it related to the crash or just something that should be tuned?
- The erroneous GPS line that was seen around the time of the crash - is this indicative of a bad GPS?