Hi All,
Running a little survey last thursday looking for some Giant Hogweed and seemingly all went well with the flights (running a Vulcan Octocopter using Pixhawk with arducopter 3.2.1). However on checking the log for the third and final flight of the day, the below error was marked up with the auto log analysis:
Test: Brownout = FAIL - Truncated Log? Ends while armed at altitude 3.80m
This is a first for me of this type of error and on checking the logs further it does indeed seem that the barometer and GPS think that it has landed whilst still ~3.8 m in the air. However as far as I was concerned during the flight (an auto flight at 115m AGL, folowed by a couple of guided mode points at an altitude of 60m AGL), all went very well and the landing was fine, prety well much bang on where it took off from (which was a flat track). On checking the telemetry logs, it appears to be reporting the same discrepancy.
The GPS says that it landed at an altitude lower than it took off from, although of course GPS is very inaccurate with height and so that why the Baro is used, and I cant see anything obviousely wrong with any of the Baro log options, except that the temperature at the start of the flight seems to be some 8 degress higher than the end of the flight (I am not sure what the baro temp units are in, see image below). Could this be the cause of this discrepancy?
It would be good to know what has caused this issue (or if its just a glitch that has been seen before), especially as I use the dataflash log to geotag the images (which are then processed in photoscan), so height errors could cause me further post processing issues, as well of course flight navigation issues.
If you have seen this before then it would be great to know what might have caused it. I have attached the log within a zip file (7zip) as it was 15 MB big, so if you anyone has the time to help diagnose the issue that would be great.
Many thanks.
S.
Replies
"Hog Weed" huh.... ;-) ... me too ;-)
Looking at your logs, everything seems fine : no errors, good voltages, good HDOP, perfectly stable altitude. The barometer chip used on Pixhawk is known not to be very reliable in absolute measures over time (it drifts). You can experiment the same on the bench if you let your barometer measure for a few minutes, you will see a few meters of drift. This worry has been discussed many times on this forum but conclusion is that there are no software solutions for this baro drift issue; a solution must involve extra and/or modified hardware and/or external reference points, but no one has come with a solution to this problem yet.
Hi Hugues, thanks for taking the time to have a look and yeah from what I had picked up in the forum is does seem to be a niggling issue. I will just have to keep an eye out for it to see if it continues and/or gets any worse, in case that signals a potential hardware issue.
Regards,
S.
Hi Simon,
the cause is most likely a drift in the baro height estimate. Although it should only occur after longer flight times. Some discussion can be found at:
https://github.com/diydrones/ardupilot/pull/2310
Best regards,
Thorsten
Thanks Thorsten, the flight was ~14 minutes long and I will check that link out.
Regards,
S.