Please select Into the mobile phone version | Continue to access the computer ver.
Attn DJI: Why does gimbal jump when tilted straight down?
2575 10 2020-6-5
Uploading and Loding Picture ...(0/1)
o(^-^)o
Zbip57
Second Officer

Canada
Offline

The user manual says the Mini's gimbal has a Controllable Range of motion between -90° to 0° (default), or -90° to +20 (extended).
The manual also says that the gimbal's Mechanical Range in Tilt is -110° to +35°.

With the gimbal controlled to tilt straight down, this means the Mini should be able to pitch another 20° rearward when flying backward, or when braking to a halt while flying forward, without exceeding the mechanical limits of the gimbal (see diagram).

This can be demonstrated as working properly, also not working properly, in a simple test (see video below).

Power up your app, controller, and Mini, but do NOT start the motors!! Pick up the Mini in your hand and, using the app, tilt the gimbal down to the 90° straight down position. Now manually tip the Mini to different angles while observing how the gimbal moves.

As you tip the Mini forward, the gimbal automatically remains stabilized pointing straight down, as expected.  When tipping the Mini rearward, the gimbal also remains stabilized pointing straight down, also as expected.

However, as you tip the Mini more than 20° rearward, the gimbal hits the physical limit of its Mechanical Range of motion and obviously the camera is dragged upward as you tilt the Mini further rearward.  But, note carefully, as you bring the Mini back down to level and as the Mini's pitch attitude comes back to less than 20° rearward pitch, the gimbal smoothly resumes holding the camera stabilized pointing straight down.

The point is, even if dragged outside of the gimbal's Mechanical Range with extreme attitude angles, the gimbal does continue to work properly and resumes its initial stabilized position once the attitude angles return within the limits of the gimbal's Mechanical Range of Motion.  But it only does that when the motors are not running!

Repeat the same test, but this time with the motors running.  Hold on tight to the Mini because it will fight you when you tip it away from level.  (It would be much safer to do this test with the props removed, but I didn't want to disturb the loctite on the prop mounting screws.)

The gimbal behaves very differently when the motors are running.  As the Mini is tipped backwards and the gimbal nears the downward limit of its Mechanical Range of Motion, rather than just dragging the camera forward along as the Mini is tipped further rearward (like it does when the motors are not running), the gimbal instead suddenly jumps forward by a small angle.  Watch the video carefully and you can see the gimbal steps forward.  And, once the gimbal has been pushed forward, it remains at that displaced forward angle rather than returning to vertically straight down.

Of course the gimbal cannot be forced to operate beyond its Mechanical Range of tilt.  But the firmware seems capable of making the gimbal work properly when the motors are not running.  Why is the gimbal not working properly when the motors are running?



This issue was raised in a previous topic thread:  Annoying Firmware Update: Auto pan up
That thread included this video demonstrating how the gimbal jumps.


Gimbal Mechanical Limits.jpg
2020-6-5
Use props
Suren
Captain
Flight distance : 9026447 ft
  • >>>
New Zealand
Online

Maybe it has something to do with the strong wind warning you getting on your screen
2020-6-5
Use props
m80116
Captain
Flight distance : 3264131 ft
Italy
Offline

For a moment I was expecting someone fighting the motors... good analytical eye, it's uncanny how you managed to isolate the problem as only w/ props running.

Can I suggest you spend a little bit of your time to compile a report within the Fly app ? If you PM me some txt (I have limited time) I can do the same. Perhaps at some point DJI might consider this.

I've never been able to discern this whim but apparently it should be consider as such, bug. Something that can and should be fixed.
2020-6-5
Use props
Zbip57
Second Officer

Canada
Offline

Other than the obvious jump in the recorded onboard video, it's difficult to shoot external video showing this happening while the Mini is in flight.  So I thought I could just record it while holding the Mini in my hand without the motors running.  I was then very surprised to see the gimbal operates differently compared to when the motors are running.

I've submitted the info to <support@dji.com> and am waiting on a response.

2020-6-5
Use props
Zbip57
Second Officer

Canada
Offline

It's weird.  I get it that the gimbal has a mechanical limit to how far  it can rotate in tilt.  If you exceed that range the gimbal cannot  physically rotate further to hold its stabilized position.  That all  makes obvious sense.

But I don't understand why it would need to behave differently with or  without the motors running.  The way it works without the motors running  makes more sense to me.

With the motors running, it's as though the gimbal senses the mechanical  limit is getting too close.  So the gimbal automatically jumps away  from that stop limit and re-positions itself a few degrees ahead and  clear of the bump stop.  That way it doesn't end up jammed tight against  the stop unable to move.  But then it won't automatically return  to the straight down position after the Mini's rearward tilt resumes a  less harmful more level angle.

But, as shown without the motors running, it doesn't seem to harm the gimbal  in any way even if it is trapped against its range limit stop when the  Mini is tipped too far rearward.

The difference between motors running or not really does seem like a  firmware bug because this way the gimbal is not making use of its full  mechanical range while in flight.                
2020-6-6
Use props
The Duck
Second Officer
United States
Offline

I just recreated your experiment and I had similar results. However, I could only get gimbal to auto rise when in sport mode.  It behaved in both Cine and Position modes.   But I didn't have strong winds.

Here's something else I noticed, when the gimbal was at full 90 degrees down, the Mini would not exceed 8.8MPH in P mode.  I didn't know about this limitation and I think slow speed assured the the Mini kept a level attitude and thus no auto gimbal rise.

While in sport mode, I could fly forward quickly without the gimbal rising.  But when the Mini braked or flew backwards the gimbal automatically rose.  I suspect this is related to the gimbal dropping when flying forward in  sport mode as some have said happens with the Mavic Air 2.

Interesting stuff.  I'm sure there's a reason as to why the gimbal is programmed this way.  Maybe to reduce strain on the gimbal motors?  I have no idea.
2020-6-6
Use props
hallmark007
Captain
Flight distance : 7009278 ft
  • >>>
Ireland
Online

The Duck Posted at 6-6 10:25
I just recreated your experiment and I had similar results. However, I could only get gimbal to auto rise when in sport mode.  It behaved in both Cine and Position modes.   But I didn't have strong winds.

Here's something else I noticed, when the gimbal was at full 90 degrees down, the Mini would not exceed 8.8MPH in P mode.  I didn't know about this limitation and I think slow speed assured the the Mini kept a level attitude and thus no auto gimbal rise.

This is dji explanation, the exact same thing occurs in all dji Mavic drones except the orignal Mavic, which incidentally could not change its attitude because of OA to cope with strong winds, many orignal mavics lost to the wind in Rth mode. So this was devised to enable craft to cope with strong winds.

2020-6-6
Use props
Zbip57
Second Officer

Canada
Offline

The Duck Posted at 6-6 10:25
I just recreated your experiment and I had similar results. However, I could only get gimbal to auto rise when in sport mode.  It behaved in both Cine and Position modes.   But I didn't have strong winds.

Here's something else I noticed, when the gimbal was at full 90 degrees down, the Mini would not exceed 8.8MPH in P mode.  I didn't know about this limitation and I think slow speed assured the the Mini kept a level attitude and thus no auto gimbal rise.

It makes sense if it only happens in Sport-mode because in that mode the Mini's maximum tilt (pitch) angle is 30°.   The gimbal can only tilt 20° beyond straight down before reaching its mechanical limit.  In P-mode and C-mode the Mini's maximum tilt angle is 20°.  That should mean, if the gimbal is pointed straight down, you should be able to give full back stick in those modes pitching the Mini rearward at a maximum of 20° tilt without exceeding that extra 20° of mechanical range in the gimbal.
With the gimbal pointed straight down, flying forward quickly in S-mode won't affect the gimbal at all because the Mini is pitched forward (nose-down).  In this position the gimbal is sitting near midway in its mechanical range of motiion and is unaffected by either end stops.

I've added a diagram showing the Mini flying forward at its maximum 20° tilt angle in P-mode or C-mode.  You can see there is no issue here if the gimbal is pointing straight down.  It's the opposite extreme that becomes an issue when flying forward.  If the gimbal is raised to point at the horizon, i.e. 0°, you'd still have a 15° buffer before the gimbal reaches the maximum upward tilt limit of its mechanical range.  But if you choose the extended option in the menu allowing you to rotate the gimbal higher than the horizon (+20°), that would exceed the gimbal's mechanical range when the Mini is pitched full nose down in forward flight.

In Sport-mode the Mini will pitch nosedown up to 30° in forward flight.


Gimbal Mechanical Limits.jpg
2020-6-6
Use props
DJI Stephen
Super Moderator

Online

Hello there Zbip57. I am sorry to read and to know that you are having issues with your DJI Mavic Mini. In addition is is good to know that you have contacted our DJI Support Team for further assessment with regards to this matter. Please keep us posted on the said issue for updates. Thank you.
2020-6-10
Use props
Dublinphantom
Second Officer
Flight distance : 1033222 ft
  • >>>
Ireland
Offline

Before anyone suggests this is an issue with one persons drone etc, it seems that it is in the design but must not have been spotted during manufacturing and testing. i too have this same issue and my belief is that every Mini has it. as shown the first post, it only happens while motors are on which would suggest a software update should fix it.
2020-7-9
Use props
Zbip57
Second Officer

Canada
Offline

To update this thread, here is the response I received from DJI Support;

June 10th,
"We are sorry for the inconvenience this has caused you. For this  concern, we would like to have this checked with our internal team and  see if this is an issue or if it is still normal. I will forward the  video to them as well for evaluation. Then, I will get back to you as  soon as I received any feedback from them."

June 17th
"We just received a feedback from our internal team.
They have  mentioned that there are 2 modes of the gimbal performance, one is for  the motor stop; the other is for the motor work. When the motor works,  the gimbal uses the performance with the motor stop, the gimbal will  become unstable, thus causing it to jump."


Ah, yup.  That explains everything, not.

I understand that in flight at extreme pitch angles it is sometimes possible to exceed the gimbal's mechanical range of motion.  The gimbal runs up against the bump stop and thus obviously is unable to maintain image stabilization.  That was a problem with one of the earlier Mavic models when, in fast forward flight with the gimbal pitched all the way up, the gimbal would come into contact with its uppermost mechanical limit.  If jammed against that upper stop, the gimbal is unable to damp out vibrations thereby creating jello effect in the video.  So the solution was to make the gimbal automatically jump a few degrees down to avoid contacting the mechanical bump stop.

The same thing happens when the gimbal is aimed straight down and the aircraft pitches too far rearward causing the gimbal to reach the maximum downward limit of its mechanical range of motion.  To prevent the gimbal being held against the bump stop, it automatically jumps several degrees upward to avoid touching the bump stop.

All of that makes sense and you can see the gimbal jump forward in the video I posted above.  But none of that answers my question as to why it should operate differently with motors running compared to motors not running.  I still think it makes more sense to leave it the way it works with the motors not running.

The gimbal will potentially hit its end stops in only rare extreme aircraft pitch angles, and even then only momentarily.  You would see that happening on the video, but it would make sense...  Oops, I'm being too aggressive with the stick inputs.  Just back off a bit on the stick inputs and the gimbal should resume it previous stabilized position (as it does when motors are not running).

The way it's operating now (with motors running) doesn't make sense to me because, rather than a brief nudge against the bump stop, we're instead seeing the gimbal take a very noticeable jump in pitch angle and sticking there.  The gimbal will not automatically resume pointing straight down by itself.  You need to actively command the gimbal to point straight down again.


2020-7-9
Use props
Advanced
You need to log in before you can reply Login | Register now

Credit Rules