Tgran, First of all it's very important to point out that the forces working on the rig during a roll with a mikrokopter are different than those with roll in a KAP rig. With a KAP rig, it's the sideways motion of the rig and suspension that rotate the rig (with camera) around the mounting point (kite line) of the suspension. With The Mikrokopter (MK) this is different; it's the MK that starts the roll. If the camera mount is fixed to the MK, it rolls the camera. If the camera mount is mounted on a freely rotating axis, it simply stays horizontal, with the MK rolling above it.
The difference is, that with a KAP rig, you have to force the camera to a horizontal position if there is roll. Where the mount under the MK is kept in horizontal position by gravity. So, apart from the force straight down, not a lot of force is woking on the actual axis for roll on the MK-camera mount. I tested it today just stabilizing roll using gravity. That did not really do the job. there's a little to much wiggeling. So I plan to stabalize this with a servo, just like tilt is stabilized.
The current MK-camera mount is looking very much like a KAP rig. Bottom and top u-shaped frame. Rol-axis is above C.O.G. in the upper frame part (I'll photograph it later. now it's time for bed.)
Being above the center of gravity helps a lot in terms of getting out of the way for a horizontal shot. If you want to be at the COG (i.e. for minimizing battery use for active stabilization on a KAP rig) then it seems you need to use half a U for the roll support (with unusual stresses on the bearing) or find an unconventional way to define the rotation (for example, David Wheeler's second rig). Maybe a clever use of battery weight would keep a COG axis out of the field of view.
Here's my planned design for my next stabilized rig. It puts the roll outside the tilt. I refer to as as "PRT" - Pan/Roll/Tilt, as opposed to the more common (and easier to build) "PTR", used for HoVer rigs.
With KAP in mind, this is the best, if not only way to go: When properly balanced, this puts hardly any strain on the servo's Do you have a good solution for constructiong the roll axis? I've seen something designed specially for this purpose, but that looked rather heavy. (will find the link to it sooner or later, and update)
Small question: with the cam pointing down, and the rig rolling it looks like the lower parts of the rig could get into the picture?
That is neat. The pan servo is more-or-less attached to one axis of a folding picavet? Gets it out of the way and I guess I'd also sacrifice endless pan circles for active stabilization.
Keeping the roll axis short and the camera light relieves some of my concerns about torque on the roll bearing (and reminds me that it can be just a little bit longer than the axis of a HoVer). If your camera was at all heavy I guess you'd want the roll bearing to have extended contact with the shaft. Maybe two roller bearings a centimeter or two apart. But a heavy camera probably destroys batteries pretty fast for your purposes.
I think it is the right balance. The vertical elements could be behind the camera and don't need to extend down. But this doesn't add much weight, protects the camera and the rest of the rig, and gives you room to attach batteries and boards.
Couldn't you eliminate the top frame all together and just come straight down off the pan axis to the roll axis then put your landing legs on the tilt axis frame? That would keep the legs out of the FOV and eliminate a lot of weight. Maybe mount the batteries & electronics on the back of the bracket between pan and roll to counter balance the camera fore and aft?
Small question: with the cam pointing down, and the rig rolling it looks like the lower parts of the rig could get into the picture?
Good point. There is a tradeoff between protecting the camera with something hanging below it and allowing a wider field of view. In my design, there is nothing structural below the crossbar for the roll support so I could just cut off the bottom if it's in the way.
You certainly can get the rig out of the way but if roll oscillations get larger than a certain angle I'd probably want the electronics to just turn the active stabilization off and wait it out.
Moving the landing gear to the tilt frame is an option. Personally, I always try to keep moving parts as light and compact as possible. Looking at the last image Scott uploaded, it doesn't seem to be getting in the way. (if the rig is swinging at those angles, I would stop any way.)
Tgran, With a stabilized system, it's always wise to limit the software's servo reach. If you don't do that, the software will try to push the servo beyond it's limits at some point in time. Simply put: there's an absolute turning point to which the servo can be rotated. If roll exeeds that angle, stabilizing stops until the system returns to within the max angles.
Very cool! I like the idea of losing the u-Frame and building straight to the roll servo. I'd also be concerned about stress on the joint between the roll servo and tilt frame when a heavy camera is used, but that may be less of a concern as compacts continue to improve.
I like to keep everything as light as possible. I am not saying move the servo to the Picovet. Just changing the upper yoke to an L bracket. All that big yoke does is hold the roll axis and the landing gear. Reduce it down to a simple L bracket will save a lot of weight and adding four 3mm 4 CF tubes to the roll yoke will not add that much mass. There would be a problem of the roll axis rotating when the power is off but by balancing the equipment on the L bracket around the roll axis the friction of the gears would normally hold it in place. It would also lower the COG giving more righting moment to the Picovet.
Gorgeous design. Man, I can't wait to see you build that thing and fly it!
Hey, if you're looking for a fairly light (FAIRLY light...) and inexpensive bearing system for the roll axis, I recently gutted a stack of hard drives and CDROM drives, and recovered some nice parts. One was a really pretty double-row sealed bearing pack. It's flanged on one side, and had a slot for a clip on the other (and of course it came with a clip). I haven't tested it for run-out, but my guess is it would be negligible, given that it was the pivot for the hard drive's arm. I've used similarly sized bearings in the single-row variety, and I'm pretty sure it would support the weight of your camera and tilt axis without any problems.
Pretty sure you've got dead hard drives to play with, but if you want one let me know. I'll get it in the mail to you.
Scott, wonderful design. Kite Aerial Video (KAV) may actually begin to take off with a well stabilized rig. Been shooting KAV for some time and the gyro stability combined with de-shaker software could be the sweet spot! Keep up the excellent work.
That axis arrangement is a thing of beauty, Scott. And with the way everything is structured, it looks to me like you could almost eliminate one of the side brackets and clear the camera's field of view for portrait orientation shots.
Thanks for the great ideas. I've taken another swag at the design, shortening the back and putting the supports on the tilt brackets, which I am growing more fond of the more I look it at. I am not as concerned about the rig falling over when the power is off as I am about putting stress on the bearings and servos when it touches the ground.
For the roll bearing, I planned to put two flanged ball bearings on the wood or plastic blocks that support the roll shaft, one on each side. The roll shaft would be 1/8" steel, which can be threaded to mate with a 6-32 nut.
The reason for the strange pan servo mounting is to put the pan servo and pan gyro outside of the pan stabilization loop. It's easier to sense unwanted pan rotations if the gyro does not sense intentional pan movements. I always use proportional pan servo control rather than continuous rotation pan, so I should be able to keep the wires from getting wrapped up.
I don't have any problem with the proportional pan servo. I think I like it even better than a CRS.
The new landing gear will protect the camera and it is now no problem even if the roll servo turns to a side.
Your video showed what this rig can do brilliantly. But will the extra weight a costs gives us better pictures? Or, to put the question in another perspective, can we think of a similar rig for video cameras?
Your video showed what this rig can do brilliantly. But will the extra weight a costs gives us better pictures? Or, to put the question in another perspective, can we think of a similar rig for video cameras? I guess this answers your question regarding video rigs. Image quality won't improve with a stabilized rig. The percentage of keepers out of a serie of pictures will likely be bigger. Reduced numbers of motion blurred pictures. Only drawback would be more change of vibration with the servo motor and gears constantly running.
Although Scott's rig design is beautiful, we don't have to wait for it. A Brooxes PRT rig is almost ready (thanks to the clever design of the Brooxes frame parts):
But will the extra weight a costs gives us better pictures? Or, to put the question in another perspective, can we think of a similar rig for video cameras?
In general, added weight will give better pictures, for the same reason a heavier car gives a smoother ride :)
My intent was for this to be a video rig - there is no provision for a shutter trigger. I think a stabilized rig will help to compose pictures since the rig is not swinging, but I am not convinced it is worth the effort if you are only trying to reduce motion blur for still images. Then again, other than size and weight, the added cost is only about $50 per axis so it's not too expensive relative to the cost of a typical KAP system.
Ironically, this thread was created because we were getting off topic on a previous thread and I guess we are doing that again. Sorry, Wicherd.
No excuses needed Scott, as far as I'm concerned this is a fine thread for everything regarding (Roll) stabilization. Including cost and functionality. Personally I prefer to keep things together in one thread.
Are you planning to mount all electronics on the picavet? (mostly the gyro's) In my opinion this would be the best place. It would also eliminate most of the disturbance rig and servo movement could have on the gyro's. Only challenge would be to keep the wiring to a minimum at this location.
That ServoCity DDP125 looks interesting. I had assumed a 1:1 gearing for roll to take the stress off the roll servo, but adding an extra ball bearing support to the roll servo shaft may work better, since there would be less chance of play in the gears. ServoCity makes another bearing mount (the DDP155), which I have used with their tilting mount for time lapse motion control. It works pretty well and seems strong enough for a roll axis. Back to SketchUp to check this out !
I was thinking of mounting the electronics and battery on the roll support crossbar. That way, I would have the minimum number of wires going through the pan pivot.
That would be the best place, I agree. It does place the pan gyro on the object it's controlling. I could imagine that complicates the programming and control over pan. Am I reading that you prefer the 155 over the 125? (me not native english speaking) If so, than I see no technical difference that's making the 155 stronger than the 125.. only more heavy.
Servocity has a page suggesting the upcoming availablity of light and medium duty roll systems. Probably pretty similar, but maybe it involves a bearing specifically designed for radial loads (roller bearing, etc.). Maybe someone can communicate with them.
Am I reading that you prefer the 155 over the 125? (me not native english speaking) If so, than I see no technical difference that's making the 155 stronger than the 125.. only more heavy.
I probably would not use either one, but rather I might borrow the idea of using a shaft directly from the servo being supported by another ball bearing.
I think ServoCity mounts are well engineered, but the ones I seen were a little heavy for KAP, so their new ones may be interesting.
Pretty clean looking frame Broox. I especially like the modification for the roll frame. It looks about perfect. I also appreciate the wires and green (velcro?) tape ; ) I sometimes put random modifications on with scotch tape or whatever is available.
I finally got over a stumbling block on my own roll axis support for a free-form carbon fiber PRT rig. Tempting to go your direction for a back-up rig.
Scott, have you had any luck with active stabilization of the pan axis (i.e. all three axes for straight down shots)?
I think you said somewhere that typically KAP rigs were too flexible to easily damp the pan axis. And I saw your work on coupled pulleys picavet pulleys for passive damping. I guess I'm wondering whether it's worth the effort to try to make a rigid rig for active damping.
I'm about to engineer my pan axis and I'm wondering how much effort I should put into it.
One of the problems with stabilizing pan is that, unlike tilt and roll, the pan gyro feels the intentional pan motion. So you have to tell the software how to recognize and not correct for the motion that is the result of intentional panning. In theory, this is possible, but in a KAP rig, the dynamics of how the rig moves depends on many factors that vary from one rig to another. In addition, the stability of Picavet and pendulum suspensions changes with kite line angle changes, so the stability can change from one moment to another. In theory, this could be handled with an adaptive software algorithm, but that's something I have not pursued.
That will help, but I have tried that and in my case I found it to be an awkward way to mount the pan servo. But I don't mean to discourage you from investigating this - these are problems in need of a better solution by somebody.
Just finished my first Roll Stabilization rig. Got a box from Brooks today, and spent the remainder of the day until just now in the basement tinkering. Although I had seen the video demos on Scott's site many times, it's really cool once you build one and it works. I used mostly Brooxes parts and obviously the GS-1 servos, and cannibalized some parts off my main rig which I also just turned into an AutoKAP rig. Looking to do some High Def video and see how it turns out.
