Developer Comment by R_Lefebvre on November 26, 2012 at 12:35pm

Vince, you have really nice design skills, but I feel I should point out a few issues with your design.  I have been engrossed in gimbal design for a while and have seen a lot of various successes and failures.

1) Roll is usually a child of yaw.  You have it the other way. If you have the camera pointed to the side, and the airplane pitches down, you won't be able to level the image.

2)Direct mounting of roll axis to servo.  This isn't great.  This will result in a sloppy joint, which will allow for vibration/shaking.  Ideally you should use a rigid shaft with bearings, driven by a belt or gears.  Look at the Servo City Power Servo Gearbox for inspiration.  At the very least, use a Servo City Servo Block here, as the bearing will be much better than that in the servo case.

3) Gearing up for yaw.  This is not great.  Servos jitter, and you will magnify that jitter.  Normally people gear down to reduce the jitter.  You should either use a 180° 1:1, or gear it down and use an external pot.

4) Direct drive of pitch.  Similar problem as for the yaw, (and Roll, really).  This is doable if you have VERY good servos.  I've never seen a very good micro servo.

Fantastic aircraft by the way!

Comment by Vince Hogg on November 26, 2012 at 2:22pm

My initial sketches were for a belly mount with the pan gimbal first (working from plane to camera), then roll and tilt. I incorrectly transposed the bits for a front mount. Good spot.

Direct mounting servos – yes I agree its a bit of a tempting shortcut and was intending to use a serious 2 bearing digital for that one. I usually over engineer but am forcing myself to start simple, light and small, then fix things that need fixing.

I think I will go back to my first idea and start out with a belly mount. Dont really want to hack the nose off my plane!

 

Any advice/opinions gratefully received.

 

Developer Comment by R_Lefebvre on November 26, 2012 at 3:20pm

What would be great is if there were more choices in servos with a full metal case.  This would really help make a stiff direct mounting.  Something like this:

http://www.hobbyking.com/hobbyking/store/__22607__Turnigy_Integrate...

This:

http://www.hobbyking.com/hobbyking/store/__12451__Turnigy_MG959_All...

Or this in a Micro:

http://www.hobbyking.com/hobbyking/store/uh_viewItem.asp?idproduct=...

Comment by criro1999 on November 26, 2012 at 5:38pm

Take a look at my GoPro Hero2 simple gimbal...see the video also on the link below.

http://diydrones.com/forum/topics/h-frame-quad-for-10-weight-1kg-wi...

Distributor Comment by Dany Thivierge on November 26, 2012 at 7:17pm

hard to beat these ones... micro and metal gears, digital... and price is good too... 

EMAX-ES08MD

I will soon do a video of how smooth they are. 

Comment by Firefly on November 26, 2012 at 9:24pm

Great project! Will you be making the design open source?

I agree with  R_Lefebvre, I would also like to add that direct mounting puts a very large load on your servos. You may consider incorporating some of the concepts used by ServoCity for their mounts. Either way please consider using a seperate power supply/battery for the gimbal while testing. I have seen a few expensive vehicles lost due to binding servos browning out the receiver/AP.

I know your design if for carbon fibre and aluminum but if you want to prototype something in ABS to prove the design feel free to send me a message.

Firefly

 

 

Comment by Vince Hogg on November 27, 2012 at 3:22am

Firefly, thanks for offer. Good point re- power supply. I am currently using an ELYQ from a 2 cell lipo, feeding 6v to most of the servos and regulated 5 to the APM. The RX is being fed direct from the lipo (optima9 SPC slot). However I have run out of slots on the Elyq. Perhaps I should consider using some HV servos?

 

Am also wondering about butchering a servo and remote mounting the pot on reduction gears to get 360deg+ pan but some servos only have half gears I think.

 

 Dany –wow seems like a bargin if they work. Hows the Canadian $?

Comment by healthyfatboy on November 27, 2012 at 7:02am

Using something like a Servo Block from ServoCity would help stiffen things up considerably. I use one on my pan servo on my r/c truck and it works great.

Comment by Vince Hogg on November 28, 2012 at 6:55am

Back to a belly mount design. Much happier with that arrangement.

 

Thanks all for advice on potential issues. I am not expecting perfect results first time but will tackle problems as they arise.

I know direct drive servos are not ideal but will see how it goes. The one driving the tilt is a HS-5245. It seems very smooth in the hand.

The roll servo as noted before, is carrying the load of the camera and tilt servo, which comes out at 153g at 3.5cm. This gives a torque of 0.53kg/cm, well inside the 12.4Kg/cm quoted for this (HS-7985) servo, albeit axial load rather than radial. The pan servo will also be a 7985, giving 280deg with belt gearing. It will be soft mounted in the fuselage.

 

Will probably mould a dome cover and face the camera away from the nosewheel for takeoff else it will get splattered.

 

 

Time to start making.

 

Developer Comment by R_Lefebvre on November 28, 2012 at 7:35am

That design looks OK, the only thing is again, the 2:1 ratio on the yaw.  I would suggest in that case, you won't want the servo to *stabilize* the yaw, you only want it to be useful for actually moving the aiming point.  ie: use it to change the aiming point in 5° increments, but don't let it try and correct 0.1° errors in the yaw.  Really dampen it down so it's not moving around a lot.

Another option, possibly better, is just to use a 180° servo, which are available.  Many Hitec digitals can be programmed for 180.  There are a few cheap Chinese options too.  HiModel has some "robot servos" for like $30.

Comment

• ‹ Previous
• 1
• 2
• Next ›
• Page