BTW if you see data on forced landing, it's probably actioned by me to get the drone to land on my hand.
Mavic Air 1 - Low Altitude hold issue
- Thread starter Atomwild
- Start date
OK ... so the first .TXT log you posted before was the second flight on the same power on, you started to fly & landed (motors shut down) & you took off again.
When doing like this you should get a new HP recorded when you take off again ... & the barometric sensor should reset to 0m.
Looking into the DAT log event stream confirms this ... the orange highlight is the barometric reset.
All good there ... but if we look into the data after the take off (from IMU 1 which was used) the barometric height isn't reset at all. Have placed the marker in the chart where the Barometric sensor should have been reset according to the DAT log event stream ... according to the red graph the relative height is -13,6m when it should be 0m.
(Click on the chart to make it larger ... all data values from the marker position in the legend under the chart)
Have also red-circled the event where an uncommanded change in VPS & Barometric height is together with a vertical speed change to -0,39m/s (ascent).
Can't see any logical reason for the uncommanded vertical change ... but it's really strange that the Barometric sensor didn't reset as it should have.
I would test this further ... & this time in one single flight without any landing in the middle & see if it repeats.
Thanks for the analysis. Will do the run again as per your advice in the next couple of days.
By "barometric sensor" I assume you are referring to IMU:relativeHeight:C. It isn't actually the sensor, it's the result of a computation that records the actual barometric pressure at launch to offset the pressure in flight. For most, if not all, of the DJI platforms relativeHeight is computed by the AC and included in the .txt log file. Some of the platforms, like the Mavic Air, don't include relativeHeight in the .DAT. To cover those cases DatCon attempts to compute relativeHeight - it just doesn't always get it right. That's what happened here.
There is something quite odd about the actual barometric pressure altitude data which can be seen in the flight before the flight presented in the .txt. There are three separate instances where pressure altitude seems to "drop out" in an non-physical way. Here's one
These happen when the pressure altitude slips below about -7.8 meters. Note, this condition will only happen close to sea level and on the days where the ambient barometric pressure is high - consistent with the OP's observation that the problem doesn't always present itself.
There is also something odd about IMU_ATTI:alti(1)
. I think this is probably the altitude data used by the FC and is the result of fusing the several altitude data. It should, and usually does, track those other altitude data. In this case it doesn't
The green plot shows the alti data not tracking either the GPS:height or pressure data which do track each other.
If you made it this far you're probably asking yourself is this guy ever going to provide an answer. Sorry, the best I can do is suppose that the barometric sensor is faulty in a way that confuses the FC. Maybe it's time for a trip to DJI repair.
Yep ... my thought was to keep it to 2 different expressions for the height measurement to make it easier to follow, VPS & Barometric.
Agreed ... came to that conclusion too.
? No worries, I'm a fast reader & still a learner so feel free to take the long explanation. But yeah ... if this behavior continues it might be worth while letting DJI look at it.
Going a bit further with this it can be seen that the reason the AC drops is that the FC thinks the AC is beginning to rise but it actually isn't.
The green plot shows the alti data starting to rise. The response is the FC, not the pilot, commands the AC lower as can be seen by the red pressure data.
The green plot shows the alti data starting to rise. The response is the FC, not the pilot, commands the AC lower as can be seen by the red pressure data.
@BudWalker @slup cool thanks for the all the info that was my gut feeling when I looked at the 1st data, something was wonky in the internals.
There was also a odd battery voltage event. Voltage was dropping steadily as would be expected but then went back up, it was about .1 of a volt in the same area as the altitude differences.
My question is would that be "normal" if engine load dropped or possibly just a reporting artifact ?
There was also a odd battery voltage event. Voltage was dropping steadily as would be expected but then went back up, it was about .1 of a volt in the same area as the altitude differences.
My question is would that be "normal" if engine load dropped or possibly just a reporting artifact ?
Could you be a bit more specific? At what times? And, which voltage signal are you looking at?
@BudWalker sorry if I am asking a lot of obvious questions, but I am big of a dog with bone when it come to troubleshooting. l like to understand processes.
In post 13 the 1st log file he posted starting at 6m 50 Secs, I have cut the basics out of the log, it looks odd to me.
This may be normal, but I would expect voltage to decrease or stay stable.
In post 13 the 1st log file he posted starting at 6m 50 Secs, I have cut the basics out of the log, it looks odd to me.
This may be normal, but I would expect voltage to decrease or stay stable.
| 6m 52.0s | 11.048 | FALSE | 3.695 | 3.648 | 3.678 | 0.053 |
| 6m 52.1s | 10.972 | FALSE | 3.671 | 3.6 | 3.618 | 0.071 |
| 6m 52.2s | 10.972 | FALSE | 3.671 | 3.6 | 3.618 | 0.071 |
| 6m 52.3s | 10.972 | FALSE | 3.671 | 3.6 | 3.618 | 0.071 |
| 6m 52.4s | 10.972 | FALSE | 3.671 | 3.6 | 3.618 | 0.071 |
| 6m 52.5s | 10.972 | FALSE | 3.671 | 3.6 | 3.618 | 0.071 |
| 6m 52.6s | 10.972 | FALSE | 3.671 | 3.6 | 3.618 | 0.071 |
| 6m 52.7s | 10.972 | FALSE | 3.671 | 3.6 | 3.618 | 0.071 |
| 6m 52.8s | 10.972 | FALSE | 3.671 | 3.6 | 3.618 | 0.071 |
| 6m 52.9s | 10.972 | FALSE | 3.671 | 3.6 | 3.618 | 0.071 |
| 6m 53.0s | 10.972 | FALSE | 3.671 | 3.6 | 3.618 | 0.071 |
| 6m 53.1s | 10.846 | FALSE | 3.671 | 3.6 | 3.618 | 0.071 |
| 6m 53.2s | 10.846 | FALSE | 3.716 | 3.67 | 3.696 | 0.071 |
| 6m 53.3s | 10.846 | FALSE | 3.716 | 3.67 | 3.696 | 0.071 |
| 6m 53.4s | 10.846 | FALSE | 3.716 | 3.67 | 3.696 | 0.071 |
| 6m 53.5s | 10.846 | FALSE | 3.716 | 3.67 | 3.696 | 0.071 |
| 6m 53.6s | 10.846 | FALSE | 3.716 | 3.67 | 3.696 | 0.071 |
| 6m 53.7s | 10.846 | FALSE | 3.716 | 3.67 | 3.696 | 0.071 |
| 6m 53.8s | 10.846 | FALSE | 3.716 | 3.67 | 3.696 | 0.071 |
| 6m 53.9s | 10.846 | FALSE | 3.716 | 3.67 | 3.696 | 0.071 |
| 6m 54.0s | 10.846 | FALSE | 3.716 | 3.67 | 3.696 | 0.071 |
| 6m 54.1s | 11.063 | FALSE | 3.716 | 3.67 | 3.696 | 0.071 |
| 6m 54.2s | 11.063 | FALSE | 3.677 | 3.629 | 3.657 | 0.071 |
| 6m 54.3s | 11.063 | FALSE | 3.677 | 3.629 | 3.657 | 0.071 |
| 6m 54.4s | 11.063 | FALSE | 3.677 | 3.629 | 3.657 | 0.071 |
| 6m 54.5s | 11.063 | FALSE | 3.677 | 3.629 | 3.657 | 0.071 |
| 6m 54.6s | 11.063 | FALSE | 3.677 | 3.629 | 3.657 | 0.071 |
| 6m 54.7s | 11.063 | FALSE | 3.677 | 3.629 | 3.657 | 0.071 |
| 6m 54.8s | 11.063 | FALSE | 3.677 | 3.629 | 3.657 | 0.071 |
| 6m 54.9s | 11.063 | FALSE | 3.677 | 3.629 | 3.657 | 0.071 |
| 6m 55.0s | 11.063 | FALSE | 3.677 | 3.629 | 3.657 | 0.071 |
| 6m 55.1s | 10.962 | FALSE | 3.677 | 3.629 | 3.657 | 0.071 |
| 6m 55.2s | 10.962 | FALSE | 3.71 | 3.665 | 3.692 | 0.071 |
| 6m 55.3s | 10.962 | FALSE | 3.71 | 3.665 | 3.692 | 0.071 |
| 6m 55.4s | 10.962 | FALSE | 3.71 | 3.665 | 3.692 | 0.071 |
| 6m 55.5s | 10.962 | FALSE | 3.71 | 3.665 | 3.692 | 0.071 |
| 6m 55.6s | 10.962 | FALSE | 3.71 | 3.665 | 3.692 | 0.071 |
| 6m 55.7s | 10.962 | FALSE | 3.71 | 3.665 | 3.692 | 0.071 |
| 6m 55.8s | 10.962 | FALSE | 3.71 | 3.665 | 3.692 | 0.071 |
| 6m 55.9s | 10.962 | FALSE | 3.71 | 3.665 | 3.692 | 0.071 |
| 6m 56.0s | 10.962 | FALSE | 3.71 | 3.665 | 3.692 | 0.071 |
| 6m 56.1s | 11.057 | FALSE | 3.71 | 3.665 | 3.692 | 0.071 |
| 6m 56.2s | 11.057 | FALSE | 3.683 | 3.63 | 3.661 | 0.071 |
| 6m 56.3s | 11.057 | FALSE | 3.683 | 3.63 | 3.661 | 0.071 |
| 6m 56.4s | 11.057 | FALSE | 3.683 | 3.63 | 3.661 | 0.071 |
| 6m 56.5s | 11.057 | FALSE | 3.683 | 3.63 | 3.661 | 0.071 |
| 6m 56.6s | 11.057 | FALSE | 3.683 | 3.63 | 3.661 | 0.071 |
| 6m 56.7s | 11.057 | FALSE | 3.683 | 3.63 | 3.661 | 0.071 |
| 6m 56.8s | 11.057 | FALSE | 3.683 | 3.63 | 3.661 | 0.071 |
| 6m 56.9s | 11.057 | FALSE | 3.683 | 3.63 | 3.661 | 0.071 |
| 6m 57.0s | 11.057 | FALSE | 3.683 | 3.63 | 3.661 | 0.071 |
| 6m 57.1s | 10.973 | FALSE | 3.683 | 3.63 | 3.661 | 0.071 |
| 6m 57.2s | 10.973 | FALSE | 3.676 | 3.628 | 3.659 | 0.071 |
| 6m 57.3s | 10.973 | FALSE | 3.676 | 3.628 | 3.659 | 0.071 |
| 6m 57.4s | 10.973 | FALSE | 3.676 | 3.628 | 3.659 | 0.071 |
| 6m 57.5s | 10.973 | FALSE | 3.676 | 3.628 | 3.659 | 0.071 |
| 6m 57.6s | 10.973 | FALSE | 3.676 | 3.628 | 3.659 | 0.071 |
| 6m 57.7s | 10.973 | FALSE | 3.676 | 3.628 | 3.659 | 0.071 |
| 6m 57.8s | 10.973 | FALSE | 3.676 | 3.628 | 3.659 | 0.071 |
| 6m 57.9s | 10.973 | FALSE | 3.676 | 3.628 | 3.659 | 0.071 |
| 6m 58.0s | 10.973 | FALSE | 3.676 | 3.628 | 3.659 | 0.071 |
| 6m 58.1s | 10.969 | FALSE | 3.676 | 3.628 | 3.659 | 0.071 |
| 6m 58.2s | 10.969 | FALSE | 3.676 | 3.628 | 3.659 | 0.071 |
| 6m 58.3s | 10.969 | FALSE | 3.675 | 3.627 | 3.658 | 0.071 |
| 6m 58.4s | 10.969 | FALSE | 3.675 | 3.627 | 3.658 | 0.071 |
| 6m 58.5s | 10.969 | FALSE | 3.675 | 3.627 | 3.658 | 0.071 |
| 6m 58.6s | 10.969 | FALSE | 3.675 | 3.627 | 3.658 | 0.071 |
| 6m 58.7s | 10.969 | FALSE | 3.675 | 3.627 | 3.658 | 0.071 |
| 6m 58.8s | 10.969 | FALSE | 3.675 | 3.627 | 3.658 | 0.071 |
| 6m 58.9s | 10.969 | FALSE | 3.675 | 3.627 | 3.658 | 0.071 |
In this case, perfectly normal ... that slight voltage sag was due to that the OP used the throttle for ascending.
Compare the green dotted graph which is the throttle command with the red which shows the voltage drop. As soon as he let go of the stick the voltage recovered. And as seen by the purple graph dropping (height) relative the green (throttle in neutral) the event came later.
I did 2 low-altitude flight runs and the flight records are attached. The 1st was with the battery that I thot was the culprit for the low-altitude flight issue but it was OK (as you had indicated). In both flights, there was an incident when the drone descended close to the ground but recovered back to the same height. Could be caused by the wind. Also noticed that while I set the height to about 10 ft and flying horizontally up and down the field, the drone altitude drops (after a while) to about 5 ft above the found but the display was still showing an altitude of 10 ft. The drone was still very stable at that height and as mentioned above, at one incident (both flights), the drone came down lower to the ground but recovered and came back up to the same height. Let me know if you need the DAT file.
If you still experience the same behavior with uncommanded descents followed by equal uncommanded ascents the DAT logs will show the same info & we can't reach further. Time to send it in ...
Today. it's a lot better, the drone did not descend until it hit the ground and ascend. It descends a bit and ascends back to the same height. Could be the slight wind and it behaves kind of OK. Noted and thanks for your input.
Looks like there is still a problem. Looking at FLY061 at time 513 secs the Flight Controller incorrectly thinks the height is increasing as shown by IMU:alti (blue plot). The FC responds by descending as shown by pressure, GPS:heightMSL and ultrasonic height. For the next 6 secs the FC thinks the MA is ascending and continues to command MA to descend.
I had stated previously that the pressure sensor seems to dropout when the AC descends below a pressure altitude of -7.8 meters. But, the dropouts occurred in this flight at a different pressure altitude (+7.8 meters). I took a look at some other MA .DATs and found more examples of these dropouts and at different pressure altitude heights. Additionally, those flights did not exhibit incorrect IMU:alti data. From this it's obvious the dropouts are not the cause of the problems seen in @Atomwild 's flights.
The common denominator in these flights is that the ultrasonic height data shows the dropout happening as the MA is descending below 0.5 meters above the ground. My current theory is that the drop in pressure altitude is due to a momentary pressure increase between the downward prop thrust and the ground. Just a theory.
I had stated previously that the pressure sensor seems to dropout when the AC descends below a pressure altitude of -7.8 meters. But, the dropouts occurred in this flight at a different pressure altitude (+7.8 meters). I took a look at some other MA .DATs and found more examples of these dropouts and at different pressure altitude heights. Additionally, those flights did not exhibit incorrect IMU:alti data. From this it's obvious the dropouts are not the cause of the problems seen in @Atomwild 's flights.
The common denominator in these flights is that the ultrasonic height data shows the dropout happening as the MA is descending below 0.5 meters above the ground. My current theory is that the drop in pressure altitude is due to a momentary pressure increase between the downward prop thrust and the ground. Just a theory.
