Uncommanded altitude drop in v3.06

Joined
Mar 20, 2014
Messages
747
Reaction score
55
Location
Boston, MA
I noted this during the beta test period but I experienced the v3.0.6 altitude drop at the end of a forward run and stop in a pretty significant way today.

I lost about 20ft and it was pretty sobering, especially since I was out over water. Really hoping DJI can fix this. Similar data points or observations welcome.

Documented in Flytrex data here:
http://www.flytrex.com/mission/quadcopt ... per-59169/

Visual (note the sharp alt drop, the blue line):
 

Attachments

  • ImageUploadedByTapatalk1407817997.279908.jpg
    ImageUploadedByTapatalk1407817997.279908.jpg
    118.2 KB · Views: 837
I've noticed as well in 3.06 if you are flying fast and stop in a hurry it will decend a fair bit. Mine hasn't done 20ft but maybe 10 or so.

I dunno if it's software or aerodynamics.
 
ProfessorStein said:
So, how were you flying? On the stick, or waypoints? GPS, ATTI or manual? And what kind of bird is it?

Flying GPS using stick inputs. Phantom 2 non-vision. Flying straight and fast.
 
Soulnet said:
I've noticed as well in 3.06 if you are flying fast and stop in a hurry it will decend a fair bit. Mine hasn't done 20ft but maybe 10 or so.

I dunno if it's software or aerodynamics.

I think it's software (firmware) reacting (badly) to aerodynamics.
 
Just a wild guess that It may even be differences in barometric pressure making it think its at different altitudes. Did you see it drop 20 feet or just going by the flytrex data?

It could even be just some glitches in the gps data as its moving and or if its handing off to another satellite. I noticed my flytrex ata will some times show my altitudes change or show movement even what I just have the thing parked and just turned on and not flying or moving at. Some times as much as 10 or so feet in altitude and as much as 14 feet distance from the home point. all with out really moving any were. Just from the inaccuracy of GPS and global drift.
 
J.James said:
Just a wild guess that It may even be differences in barometric pressure making it think its at different altitudes. Did you see it drop 20 feet or just going by the flytrex data?

It could even be just some glitches in the gps data as its moving and or if its handing off to another satellite. I noticed my flytrex ata will some times show my altitudes change or show movement even what I just have the thing parked and just turned on and not flying or moving at. Some times as much as 10 or so feet in altitude and as much as 14 feet distance from the home point. all with out really moving any were. Just from the inaccuracy of GPS and global drift.

You can definitely can see it visually. Scary stuff at lower altitudes over water.

It's manageable like TBE if you know it's going to happen, but undesirable nonetheless.

I don't believe it's GPS hand off related as it is reproducible 100% of the time regardless of distance, and can be induced on demand by just letting go of the sticks after a full-throttle run.
 
GPS isn't used for altitude stability.The barometer is. It has been there for a while but something in the new release has made it more pronounced. I first noticed it on the beta after hitting the default gains button. Damon, what about easing to a stop? Does that help mitigate the impact?
 
ianwood said:
GPS isn't used for altitude stability.The barometer is. It has been there for a while but something in the new release has made it more pronounced. I first noticed it on the beta after hitting the default gains button. Damon, what about easing to a stop? Does that help mitigate the impact?

I'll try that and report back.
 
GPS isn't used for altitude stability.The barometer is

In that case it could maybe even be from aerodynamics as the air is flowing at different speeds over a surface it can cause pressure changes. Then the phantom might try and compensate by bringing it down some when you stop and the pressure changes. Some thing in the new software could be using some wrong Algorithms for what ever adjustments it needs to make for things like that.
 
Definitely something that needs fixing. I'd also argue for

- adding an ATTI gain for yaw to enable buttery smooth pans without buying a $500 3rd party controller or creating a franken- controller

- adding configureable tilt in Naza mode to a certain maximum for GPS and ATT modes to lift the speed governor if desired,

- having the gimbal "sleep" while compass calibration mode is active

- add imperial metrics for iOSD mini display

- (for ground station) add POI capability to circle smoothly for video
 
J.James said:
In that case it could maybe even be from aerodynamics as the air is flowing at different speeds over a surface it can cause pressure changes. Then the phantom might try and compensate by bringing it down some when you stop and the pressure changes. Some thing in the new software could be using some wrong Algorithms for what ever adjustments it needs to make for things like that.

The barometer is inside the Naza so it's in a box within a box. There is probably a slight pressure drop once you start the motors up but after that I would imagine the airflow over the surface doesn't do much to change the barometric pressure internally. I'd bet it's an algorithmic issue.
 
The barometer is inside the Naza so it's in a box within a box. There is probably a slight pressure drop once you start the motors up but after that I would imagine the airflow over the surface doesn't do much to change the barometric pressure internally. I'd bet it's an algorithmic issue.

Id think it was both. Cause it dont take much air flowing over any of the holes in the shell to cause tiny bit of venturi effect. and I would assume that the barometer has to be super sensitive to even be able to read the difference in just a few feet up or down based on pressure. And if the pressure dropped by even a tiny fraction along with some thing algorithmic. If its not doing it in other vershios of the software.

What sucks is if I was thinking about this yesterday some one was at my house who has a devise that could of been used to see if the internal pressure does change at all if air is just blowing accost the body.
 
I still think this may be a gains issue, with the FC attempting to stop the aircraft too quickly in GPS mode when the sticks are released, and applying so much rearward pitch that vertical lift is compromised.

If it were changes in measured pressure due to airspeed variations then I would expect the converse - altitude gain on acceleration - to occur. I've never seen that.
 
We've also seen that the update can wipe previous gains settings for some users, I believe... seems like that'd be an excellent place to start, and not too difficult to see if readjustment makes any difference.
 
ProfessorStein said:
We've also seen that the update can wipe previous gains settings for some users, I believe... seems like that'd be an excellent place to start, and not too difficult to see if readjustment makes any difference.

Personally, my gains haven't changed from 3.04 to 3.05beta to 3.06.

The issue is easily reproducible by anyone who cares to test it and has been documented by other 3.05beta testers.
 
damoncooper said:
ProfessorStein said:
We've also seen that the update can wipe previous gains settings for some users, I believe... seems like that'd be an excellent place to start, and not too difficult to see if readjustment makes any difference.

Personally, my gains haven't changed from 3.04 to 3.05beta to 3.06.

The issue is easily reproducible by anyone who cares to test it and has been documented by other 3.05beta testers.

However - just because the gains haven't changed doesn't mean that the FC interprets them identically, or behaves the same, in subsequent firmware versions.
 
sar104 said:
damoncooper said:
ProfessorStein said:
We've also seen that the update can wipe previous gains settings for some users, I believe... seems like that'd be an excellent place to start, and not too difficult to see if readjustment makes any difference.

Personally, my gains haven't changed from 3.04 to 3.05beta to 3.06.

The issue is easily reproducible by anyone who cares to test it and has been documented by other 3.05beta testers.

However - just because the gains haven't changed doesn't mean that the FC interprets them the identically, or behaves the same, in subsequent firmware versions.

Agreed.
 
You could check influence of vertical gain. Remember, within earlier formware versions the default gain was at 200. I am still using a gain of 180 for best maintaining of altitude.

Regards, Gerd
 

Members online

Forum statistics

Threads
143,066
Messages
1,467,358
Members
104,935
Latest member
Pauos31