Today I did a test flight before doing a mission. After switching to poshold, it began behaving erratically, twitching and drifting. I was able to land, but puzzled because there were no errors from the GCS.
Running auto analysis, it gave these results?
Size (kb) 8554.380859375
No of lines 114684
Duration 6 days, 22:11:00
Firmware Version V3.2
Firmware Hash c8e0f3e1
Free Mem 0
Skipped Lines 1
Test: Autotune = UNKNOWN - No ATUN log data
Test: Balance/Twist = GOOD -
Test: Brownout = GOOD -
Test: Compass = GOOD - mag_field interference within limits (8.12%)
Test: Dupe Log Data = GOOD -
Test: Empty = GOOD -
Test: Event/Failsafe = GOOD -
Test: GPS = FAIL - Min satellites: 0, Max HDop: 99.99
Test: IMU Mismatch = GOOD - (Mismatch: 0.62, WARN: 0.75, FAIL: 1.50)
Test: Parameters = GOOD -
Test: PM = FAIL - 17 slow loop lines found, max 7.20% on line 108374
Test: Pitch/Roll = GOOD -
Test: Thrust = GOOD -
Test: VCC = GOOD -
The logs show no GPS problems (Nsat and hdop) so I don't get it. Is this a communication problem with the GPS unit? Possible wiring issues?
Should have noted:
What does it show for SAT and Hdop when hooked to mission planner at home?
23 satellites and 1.1 hdop. No indication of a problem. No warning from the GCS. It was 30 degF today.....no big deal; I wait 5-10 minutes to acclimate, fire it up, wait a minute or so for GPS after lock, and fly fine. Heck, I flew in sub zero with no problems last week, several times in the teens.
It just started freaking out in poshold. That's why I always leave stab mode without Simple/Super Simple checked so at least my pilot skills have a chance of getting it landed, as low as that may be.
If you have logging set to all plus disarmed, it might take some time to get a gps fix, check the sat count your self manually in the log. The auto check picks up a low sat count even if its at the start of a log.
Not sure what the PM fails is caused by.
I have the same unit with a PH and couldn't be more satisfied.
How about posting the actual log file?
Are you running this solo or in conjunction with the LEA-6H GPS? If in combination, how do you have it setup? Is it primary and LEA as secondary or the other way around?
As you know I have mine under the clam shell as primary but not using the compass and my LEA left up on the mast as secondary using the it's compass.
When you run it in LOITER do you get the same results?
The first part makes sense, and yes I had it set to 'all plus disarmed' Generally on a clear day it locks within 10 seconds. On a very cloudy day it may take 30 seconds. In either case I've yet to have conditions that don't allow flying.
I believe I now know what the problem is, and wonder of countless others are unknowingly causing grief by enabling too much logging.
which lead to this:
and installing the analyzer. I would highly recommend the use of this app. It backed up my suspicion that the cpu is being overloaded and the GPS was the main victim.
Now to wait for this snow storm to subside, go above 15 deg and see if enabling the bare minimum in logging makes the problem go away.
Here's the message from 'APM Flight Log Analysis':
I did not use the 3DR GPS yet. There's a reason for this thread being a question rather than a statement.
Shortly after installing 3.2 and working with Leonard on tuning, I enabled 'All' logging thinking more is better for analyzing tuning issues. I now believe that is a mistake and is asking for trouble especially with 3.2's added workload on the cpu (ex. EKF enabled). A few threads I read suggest too much logging and may overload the cpu.
See post above. Virtually nothing in the logs sticks out as being responsible for the occasional twitching and drifting I've been experiencing only in modes using GPS and/or compass, unless you know what to look for.
I could be totally off base as my experience is with pc's, but it makes sense. What do you think?
I have mine set to log almost everything and have had no problems with the M8N.
I've read to keep logging to minimum necessary, and found accounts of pilots having twitching issues similar to mine. The suggestion was to limit the logging. I would say based on my experience on pc's, it doesn't take much to upset the data flow of an OS/application and slow it down such as video gaming even on very fast platforms. For something like a FC, I would think it entirely plausible overloading the cpu even for a few ms could glitch the system. I have no proof of this and just assumed there was a problem with the tuning. Now I wonder.
Do you have EKF enabled?
There is very little information on INSErr and INAVErr. I didn't have these errors in 3.1.5 with my old GPS and minimum logging. Only since installing 3.2 and enabling 'All'. I don't know, and it could be entirely coincidental my copter had a seizure and the Auto Analysis threw up a flag for this speed error. So if it's a speed error, that tells me there's a bottleneck of some sort. No?
If its APM, it is probably worth toning down the logs keeping key ones that help. Also if APM EKF is not used due to the additional processing required (I think :)...pretty sure that is the case unless its changed).
For pixhawk full logging should not cause issue.