Gimbal thrashing

[pjakobs]

I bought a used cheap P3S the other day, it was advertized with a gimbal that had been broken in a crash, fixed mechanically but would not come good.
I'm no stranger to electronics and found many places where thrashing gimbals were described and could be fixed - mostly by aligning the horizon potentiometer.
So I set out trying my luck.
This is what the gimbal looks like:

To view this content we will need your consent to set third party cookies.
For more detailed information, see our cookies page.

Accept third party cookies

As can be seen, the yaw axis always goes to the end stop and the others do so to varying degrees.
The rol Axis sometimes stabilizes almost horizontal, while the tilt axis mostly goes to it's bottom most stop.
I have taken apart the whole assembly and tried to make sure I get the alignment right when putting it back together.
The Yaw arm is fixed with the worm screw on the flattened side of the axis. The roll arm is fixed such that the flattened side of the axis points up and is parallel with the top of the camera.
The camera is fixed such that the dji logo on the lens is upright when the drone stands on the landing gear.
Interestingly, I have had some success calibrating two of three gimbal axis when the drone is sitting upside down (roll was ok, tilt was off ~25deg, yaw was at an end stop)
I have looked at the horizon poti and found that it shows strange values. Throughout the motion range of the gimbal, the resistance jumps from ~5kOhm to ~40kOhm - can someone confirm that this is right? Is it broken? Is there a replacement part? (I'm currently trawling for mechanically broken Gimbals to slavage electronics)
See video on the resistance:

To view this content we will need your consent to set third party cookies.
For more detailed information, see our cookies page.

Accept third party cookies

I have already updated to the latest firmware - which had no effect.
I'll keep digging, but unless I have some gross oversight, I don't quite know where to go next, it seems all the basics are covered.
Any suggestions?

thanks & kind regards

pj

 

[WV. Rootman]

The max should be 4k at horizontal. At 90 degrees 2k and 6k. Never 40k0m.

 

[pjakobs]

so you're saying it's a 8 or 10k poti then?
I might have measured on the "far" pin, but that should not make a difference other than for the direction of increasing value.
I'll try to get some old, thrashed gimbal then and replace either the part or the whole module.

thanks

pj

 

[WV. Rootman]

I learned that from utube videos from a guy that repairs gimbal. Make sure the camera is on right. The flat shaft of the tilt motor to the back of the camera. Make sure the stop is not broken off where the camera spins on the shaft. If the roll motor potentiometer is not horizontal to the camera, make it so.

 

[pjakobs]

yeah, I had those points covered but that didn't fix the issues, hence me digging into the actual components.
Thanks

 

[crane-man]

Hi im having the same drama , the only thing i see is the camera will roll past the straight down position ?? Maybe thats the drama with mine ,,, i will strip the camera down today and see if i can find a fault ,, i think mine may have been put together wrong ,
Are he any diagrams around with the correct alignment ,, Cheers all

 

[120CCPM]

I'm not really helping you here, but how can you measure a potentiometer's resistance, if it's still connected to the rest of the circuit? You're measuring the total resistance between those two points, meaning the potentiometer + whatever else is connected to it...

 

[pjakobs]

@120CCPM, That is correct. However, I will assume that most of what connects to the potentiometer will be active components, I havn't really dug in, but, apart from some biasing, passives won't do you much good. In many cases (again, not checked here) the value from the potentiometer will be more or less directly read by a micro.
The worst thing that can happen is the voltage from the multimeter could activate a diode junction at some point - which would show as a drop in resistance.
What I'm seeing here is a sudden increase in resistance from 5kOhm to 40kOhm - I don't think that's a diode activating. But you are right, I will re-test with a 2nd multimeter that has a lower measurement current and thus won't activate diodes.

pj

 

[pjakobs]

and an update: I have now re-checked the horizon potentiometer with a different DMM - and I have checked the azimuth potentiometer. The latter behaves as WV. Rootman describes it, but the horizon one does not. I'm thinking I might try to get just a replacement poti to test this out.
It also shows that the horizon poti is pressure sensitive, that is: it reads (mostly) in the sub 8kOhm range when I apply slight pressure on it. If I don't, it will sometimes go to double digit MOhms.
This all implies to me that, when the aircraft crashed, the poti got a good whack which probably disloged the wiper and maybe cracked the resistive coating. Certainly not a happy puppy.

update 2:
I have opened the potentiometer and lo and behold, it has cracked roughly at the 11O'clock position. No wondern this didn't work. Optical inspection is consistent with my measurements. Now off to try to find a replacement part.
This picture isn't uber great, but if you look at the 11O'Clock position (with 6O'Clock being the center pin on the flat side), you can see where the substrate is cracked. This one has met it's maker.

pj

 

[pjakobs]

ok, Mouser delivered the most expensive potentiometers I have ever bought, I have replaced the broken one and.... fail
The Gimbal is still flapping wildly. Back to troubleshooting.

 

[WV. Rootman]

We know the potentiometer was bad. There is a good chance that the board is cracked. The board only cost around $30 us. Maybe isolate the problem by disconnecting the ribbon from one motor at a time?

 

[pjakobs]

hmm.. good Idea, but since I don't have a good gimbal for comparison, I'm not sure what should happen if I disconnect one axis / motor at a time.
The general description is that a broken / misaligned roll axis makes the whole gimbal flap around all three axis. So I could either
- disconnect the whole roll board and see what happens
- disconnect the roll motor and see what happens.
I have just looked at the yaw axis, but that one doesn't have a poti at all and my potis for both the roll and nick axis now measure exactly the same resistance at same positions.

 

[pjakobs]

so I have completely re-assembled the gimbal from ground up and left the two axis disconnected.
Nothing happened with the gimbal
Then I connected the nick axis (up/down)
Nothing happened
Then I connected the roll axis (horizohn)
and the gimbal went wild again.

One thing that I noticed is that I'm getting lots of "signal lost" and "weak video signal" messages and no or a very bad video image. Can someone check for me which of the antennas goes where?

kidn regards

pj

 

[pjakobs]

yes! I fixed it!
I have gone through every single component again and in the end, the only thing that was conceivably left was the replacement flat flex that I had put it. It had come with the drone as I bought it (the previous owner had attempted a repair and bought a number of spare parts).
Put the original (slightly battered) flat flex back in, put all the screws in, switched it on - happy dance, calibration, boom! Everything came to life. Just did my first test flight, albeit in the dark, but the craft seems to work just fine now.

The satisfaction of having something that just works is superseded by the satisfaction of getting something to work

 

[pjakobs]

One of the first shots with the repaired drone. Happy with the quality (it was a murky morning, there will be some softness due to the foggy air).
The thing I have not yet figured out:
while I can control the gimbal using the blue circle on the phone, the gimbal dial on the controller does nothing. I'm sure that's a settings issue somewhere
But that's certainly just me being a novice.

Anyway, happy to now own an almost new P3s for under 300US$

pj

 

[120CCPM]

First of all, great job fixing the gimbal! Regarding the dial on the RC, do an RC sticks calibration from the app, and make sure to move both sticks AND gimbal dial to their extremes. See if that helps...

 

[pjakobs]

thanks 120. Yes, I did that, the gibal wheel action shows correct on the app screen and is calibrated to +/- 100%. When I use the blue ring on the screen to tilt the camera, that works just fine, but the wheel still does not. I suppose that's a setting somewhere, but could not find it.

 

[120CCPM]

Go in the Advanced Gimbal Settings and do a reset, see if that helps...