148 replies to this topic

[dankers]

Quote

Maybe someone knows if there would be a way to turn off the outputs to those camera gimbal servos until the system has fully initialized.

Create a Jira ( http://progress.openpilot.org ) and assign it to Oleg, he might or might not do it. Might be worth sending him a PM also, his username is OSNWT.

[aerialcameraman]

FWIW I have had more success with the imu placed on the model and mounted on O ring supports. If you place the imu on the camera mount anywhere then you are relying on the PID loop tuning to self stabilize the mount. This introduces all kinds of problems because each mount has different inertia and you will pick up the opposite reaction force from the servo. I used direct drive magnetic encoder servos and then it becomes very simple, If the model tilts ten degrees then the servo moves ten degrees. When the imu is islolated you are then able to crank up the gyro gain without seeing the usual oscillations. You can start to correct the small initial movements which are a problem with most of the stuff out there.

[aerialcameraman]

The way that I do this is to power the 7.4v servos from a 2 cell lipo. The last thing that I turn on is the camera mount power supply.

[Patscherpofel]

I remember I read this somewhere but cant find it anymore...turning off and on camera stabilization via Rc control

this would be super important for me...since I dont really like to have it on while flying fpv...but once locked in on a moving object it would be nice to switch it on for a while
actually the sickest thing would be to lock the position and angle when you press the button...that would be off the hook : )
similar having axis lock  for a moment, but on the camera mount

[lspphoto]

Just wondering if its actually possible to have camera stabilization working with only 6 channels.
I have a dx6i and now realise the usefulness of having more channels...

I followed the instructions to set up a gimbal but if I have the flight mode on channel 5, it only leaves channel 6 so cant manage pan as well as tilt...?

[osnwt]

Probably you can use channels 5 and 6 for camera control and any of them for flight mode. Just set all 3 positions of FM switch to the same mode (say, Stabilized 1), and you should be able to control both axes.

[viper522]

I just wanted to mention a few things that may or may not contribute to this thread.  I have worked on camera roll stabilizer and I stopped when I realized that all inexpensive options that allowed for lightweight/small packaging were susceptible to 'gyro drift' as they have no point of reference.  I used a GS-1 gyro servo and a helo HH gyro plus standard mini servo - both piezo gyros I believe - with a modicum of success.  Just a few moments of steady roll off center and the gyro causes the servo to drift back to top dead center - the point at which the gyro was at when it initially received power and calibrated itself.

12vdc battery -> UBEC -> 5v HH gyro -> Servo tester -> 5v mini servo: 808 HD #11 keyfob camera on L-bracket, velcro to MC tail



I would be ecstatic if someone developed a small board that could use standard RC parts to light this fire again.

MotoGP/Dorna had this figured out, including HD wireless transmission, back in 2010 and have used it every race since.  I also have a detailed rendering of the camera pod design.

http://www.motogp.co...+OnBoard+camera

[dankers]

Looking at the video, I see no reason why CC could not do exactly that, single axis like that is not too hard. All gyros drift so you need a point of reference to correct them, for roll and pitch you can use an accelerometer. I seriously see no reason why CC could not do exactly what you want, am I missing something?

[viper522]

dankers, on 10 May 2012 - 01:18 AM, said:

Looking at the video, I see no reason why CC could not do exactly that, single axis like that is not too hard. All gyros drift so you need a point of reference to correct them, for roll and pitch you can use an accelerometer. I seriously see no reason why CC could not do exactly what you want, am I missing something?

For one, there's no stock of CC   The discussion had turned to the feasibility of a secondary board which would be mounted to the camera, moving the camera stabilization duties off of the flight controller board.  Or at least the secondary board would contain sensors which would then feed data back to the FC for processing.