DJI-M and 2 Axis Camera Mount Servos
- Thread starter rwilabee
- Start date
I finished my mods to the Coptersky GoPro 2 axis camera mount. While adjusting the servos I accidentally hit the default value button which caused the servos to bind. I immediately disconnected the servo leads to the MU. I reset the gimbal values to where they were before and now I get only the slightest movement in roll and tilt and no response at all from the transmitter tilt switch. I have recycled the WKM and reset the controls again but nothing. I checked the servos attached to the receiver directly and they work. Any suggestions? I am traveling and brought my WKM Quad with me along with a UBEC. I have not connected it yet.
Make sure the Automatic Control Gain in the Gimbal screen is not set to the default of 2.00, you will get minimal to no servo movement if they are. For the same mount with Savox servos I have the gains set to around 20.00 for each, they need to be at least 10 to get the servos to move. Make sure to read the settings first, I've done changes and saved them before only to find later that for whatever reason they didn't save everything that I had changed.
Ken
I found the problems. On the Tilt settings, I forgot to put the negative sign on the Min and had the Max set to 50. When I increased it back to the default of 1000/-1000 it was fine. Unfortunately it appears that when the roll servo locked, it stripped the gears. Unfortunately I will not be able to get a replacement until Monday. At least the roll servo is holding level so I can fly. If the wind dies down tomorrow I will try some flying in Vegas. It should be a lot better than the snow in Tahoe right now. Viva Las Vegas! Thanks for your help.
I think I am finished. Installed metal gears on the Roll servo and added a CC BEC Pro set at 6v to power the Align DS510's that I took off my Trex 500. I am still seeing a little notchiness around center on the Roll axis. I would like to do some flying but we had snow and wind today and they are forecasting 5-8 inches tomorrow then again on Wednesday and Thursday. I may not be able to fly again until I get back to Las Vegas.
GGoodrum
Member
It looks like this is more or less a direct-drive, type of mount, like my custom ServoBlocks-based version, so you'll see quite a bit of of improvement with some faster, 12-bit minis, like the Savox SH-1357s, or the Hitec HS-7235MHs. The higher resolution especially.
For the gains, I think you will get the best result with them both set right at 20. What these gains control basically, is the gear reduction of the mount. At 20, the ratio is 1:1, or "direct-drive". With my reworked mkTR mount, which has an effective roll reduction of about 2:1, I got the best results with the roll gain set at 42. The tilt axis is still 1:1 (but now with a fast DS610, which has 12-bit resolution...), so a gain of 20 still works best.
-- Gary
For the gains, I think you will get the best result with them both set right at 20. What these gains control basically, is the gear reduction of the mount. At 20, the ratio is 1:1, or "direct-drive". With my reworked mkTR mount, which has an effective roll reduction of about 2:1, I got the best results with the roll gain set at 42. The tilt axis is still 1:1 (but now with a fast DS610, which has 12-bit resolution...), so a gain of 20 still works best.
-- Gary
GGoodrum
Member
Yes, it was me. Sydney PM'd me to say they will have someone look at this first thing Monday, their time.
Yes, I'd definitely try setting the gain to 20, for both settings. What I do is hold the platform up so that I can line up the top of the GoPro with something horizontal. As you move the platform, in the roll axis, the top of the GoPro should stay parallel with your horizontal reference. If you roll the platform clockwise (looking at the front...) and the GoPro top is tilted to the left (i.e. -- clockwise...), in relation to the horizontal reference, you need to increase the gain. If it ends up tilted the other way, it is overcorrecting, so reduce the gain. You can do the same sort of procedure in the tilt axis.
If you make a quick movement, and there's a lag in the camera platform catching up, you need faster servos. If you get a stairstepping movement for slow/small platform movements, you definitely need higher resolution servos.
-- Gary
GGoodrum
Member
BTW, what Sydney is going to have them look at is the small drifting that occurs right at the neutral point. This is a bit different than a steady "twitching", that can happen with low resolution servos in mounts with lots of backlash/play. This drifting is like it is walking in the same direction, two or three "clicks", and then back the other direction two or three clicks. Very annoying. I did another test flight today and really cranked my X8 around quite a bit on a very windy day. The only time I see evidence of the drifting is when I was trying to have it just hover. Also, it doesn't happen at all until the system finds at least enough sats that it changes to flashing red twice, or once. When it is still flashing three times, there's no drifting. Weird.
-- Gary
-- Gary
Gary there is not much you can do with that lag accept run high speed servos and at 7.4 volts. Set you response volume with a small spirit level and it should give a good degree of stabilisation. It is heavily dialed in to the accelerometer but it needs a little more gyro to pick up that initial movement. At the moment we cant access that but soon I expect we will. The other problem that you have sounds like a deadband problem. If they are programmable servos then try a little more.
Last edited by a moderator:
DucktileMedia
Drone Enthusiast
Great. My av130 works perfectly on my Xaircraft FC. I really hope switching over to the DJI doesnt screw things up. Does anyone know if i could use my Xaircraft FC solely to control the gimbal, like a picloc? Perhaps the answer to running an octo is to just have a separate controller if the DJI is having output issues.
GGoodrum
Member
I don't have any lag, as my setup is using Hitec HS-7940TH digitals, and they are running on 7.4V, via a programmable CC 10A BEC. These also have the 12-bit, hires feature, so the action, even in my direct drive setup, is actually quite smooth (see this post...). The only issue I'm having right now is the hunting/drifting of the WK-M's gimbal outputs right at the neutral point. On a higher-end mount, like an AV-130, this effect is less noticeable, mainly because of the larger reductions in the tilt and roll mechanizations, but it is still there.
The good news is that I received a PM from DJI, and they are going to have somebody take a look at this problem right away.
-- Gary
david.scarbrow
Member
Gary, you're servo block set up, its apparent robustness and the results you appear to be getting to-date lends itself perfectly to what I want to do. Hopefully now that DJI have shown this interest the 'glitch at neutral' problem will become resolved. I will be more than pleased if you could PM me the parts list/details for this set-up. Thanks in advance.
David.
David.
GGoodrum
Member
Skysurf
Member
I don't have any lag, as my setup is using Hitec HS-7940TH digitals, and they are running on 7.4V, via a programmable CC 10A BEC. These also have the 12-bit, hires feature, so the action, even in my direct drive setup, is actually quite smooth (see this post...). The only issue I'm having right now is the hunting/drifting of the WK-M's gimbal outputs right at the neutral point. On a higher-end mount, like an AV-130, this effect is less noticeable, mainly because of the larger reductions in the tilt and roll mechanizations, but it is still there.
The good news is that I received a PM from DJI, and they are going to have somebody take a look at this problem right away.
I am experiencing the same thing with my rig. I increased the deadband but it dosen't make any difference. I hope that DJI can find a fix for this soon.
/Jesper