High quality but economical gyrostabilized gimbal for Epic/Scarlet

[Lauri Kettunen]

For some time I've been processing on the technical details of a high quality gyrostabilized gimbal for Epic and Scarlet.

We all know that high quality gimbals are very expensive. I've learnt that they start from $150K - $250K, and one easily makes an assumption that the high cost is somehow forced on us by the nature. And yes, scientifically we know that there is no way to recognize gravity from other mechanical forces/accelerations acting on a moving object, so proper stabilization requires highly skillful engineering. The other side of the coin is, take a look at a $150K gimbal and compare it to a car which costs $150K --what is your conclusion on that?

Once I've put my mind on this I've become convinced that technically it is possible build a first class fully gyrostabilized gimbal for Epic/Scarlet with a reasonable cost. I have been developing high accuracy navigation systems etc. and have quite a lot of experience in developing challenging engineering products and thus feel not standing steadily 20cm above the ground level. Furthermore, I'm talking about using the gimbal when shooting from car, boat/ship, snow mobile, RC helicopters etc., but not about airplanes or helicopters as that's an regulated area.

If you've read this post up to this point and you are still interested in the subject, let me point one interesting detail making the challenge bit easier; Designing a gimbal to Epic or Scarlet is easier than designing a gimbal to Red One. For, with Epic one needs to stabilize only the brain and lens and consequently in both cases there is less mass to stabilize than that of Red One. (A word of warning; Still one should not expect the Epic brain will weigh almost nothing. Second, it is not yet clear how long cable one may attach between the brain and the rest of the camera.) Less mass imply smaller and cheaper motors, and the cost reduces much faster than linearly with respect to the mass.

Basically I'm just interested in getting a proper gimbal, but buying the existing products are not economically justified. Second, have neither indefinitely time to design a gimbal. For this reason I decided to make an attempt to find out how many other people are interested in such kind of gimbals and who could contribute something if such a system is developed. You know, this post is about probing whether a kind of Linux approach could work out. That is, eventually if everything worked out we posted on RedUserNet drawings, list of components, software etc. enabling anybody who knows how to use a screwdriver to construct a gimbal or some 3rd party takes over and finds the outcome as a business opportunity.

So, if you are still reading could you reply if you have an interest in a reasonable cost gimbal, and second, if you could tell whether you would be willing and able to contribute something on designing such a gimbal. If yes, to get an idea of feedback, could you rate yourself exploiting the following guidelines:

1) High experienced user: You are able to tell where the technical challenges lie in practice. Such as, you are able to tell which axis pan,tilt or roll is most critical one to stabilize.

2) Skillfull designer in engineering: You know say what is a PID-controller, what the moment of inertia is, or you know why modelling of the mechanics of the gimbal is important.

3) Electronics/electrical engineer: Your are able to distinct between embedded system, induction machine and stepper motor, and you know how motors are driven.

Have no idea whether this attempt will crash immediately or whether it will lead somewhere, but at least there's an attempt. Finally, if you hesitate to post in public, send me a pm. Those filled with pessimism, could you please let others to have a rational discussion. Thanks

 

[Evin Grant]

Well, I have operated a fair share of gyro-stabilized helicopter shots using the Red one and the Wescam system. So I guess I would fall into category one.
I only know a rudimentary amount about the electronics and mechanics but I do dedsign and build rigs for 360º surround photography so I do have a bit of experience in camera support and deployment in the field. It sounds like an interesting project for sure.

 

[Lauri Kettunen]

Well, I have operated a fair share of gyro-stabilized helicopter shots using the Red one and the Wescam system. So I guess I would fall into category one.
I only know a rudimentary amount about the electronics and mechanics but I do dedsign and build rigs for 360º surround photography so I do have a bit of experience in camera support and deployment in the field. It sounds like an interesting project for sure.

Evin, thanks for your comment. Perhaps I put my words too strickly. Any feedback and comments on experience is very wellcome. Then the other issue is, how many people in the RED community are interested and able to make technical contributions that are needed if something is to be made.

Once you mentioned Wescam, how stable is it? Is it one of those rock steady systems?

 

[Erik Widding]

This sounds like a fantastic idea. Given that the lens mounts we make and the still camera lenses allow for a drastic reduction in weight, we would be happy to throw some labor at this... probably in the control system end of things.

Get the mechanical design going, and a rough start with the sensor/motor selection, and we will jump in.

I would propose solidworks as the design environment for the mechanicals. We have a Dimension 768 SST printer at the office, which allows us to make parts from ABS plastic up to 200mm x 200mm x 300mm. We MIGHT be able to throw ina couple of prototype parts.

 

[Lauri Kettunen]

Given that the lens mounts we make and the still camera lenses allow for a drastic reduction in weight

Erik, thanks a lot for your input. If Birger is in the project, that makes all the difference.

Yes, the Birger mount and the light still camera wide angle lenses together allow a significant reduction in weight. Furthermore, the wireless control of the Birger mount exists already and together with the existing 10ft LCD cable this means that in a car, boat, or snow mobile one did not need a wireless video link at all. All this imply a drastic reduction in costs as well.

In fact, I think such a gimbal should be designed with the same principle of modularity as what RED has adopted. This means, that if the manual/camera-user pan & tilt (perhaps also roll) control interface is implemented say with Bluetooth, then one could choose whether to use a mobile phone, a laptop, (perhaps) the Impero or whatever selfmade system to operate the camera rotations. As good gyroscopes, especially the optic ones, are bit expensive, it would be nice if the system allowed all kinds of gyros starting from the cheap MEMS sensors. Then one could choose the level of stabilization and balance with how much money one is willing to invest in a gimbal. BTW, a nice thing there is that Nintendo Wii has opened the markets for gyroscope mass production. In the near future we may expect to see much improved gyros with a reasonable price tag.

Solidworks environment sounds like a good idea and just fine.

 

[Larry Gebhardt]

Hey Lauri....Great Idea!

I would guess I am in between 1 and 2, having graduated college with an engineering degree, and then out of boredom with the lack of creativity allowed in several jobs after, got into camerawork. I spend about 1/2 of my professional time shooting or planning/thinking about shooting, and the other half in machine shops or planning / thinking about devices, rigs, etc to use in the first half. It is an all consuming passion..."what could you shoot if you had an XXX". The shop I hang out in mostly these days is a very high precsion CNC shop (Haas machines), small enough to do prototype work reasonably and capable enough to be able to hold several tenths on, for example, lens mount work.

I am also experienced with most of the major gimbal mounts, Wescam, Gyrosphere, Spacecam, etc. and recently shot part of a 3D film for Coke using a Scorpio stabilizing gimbal mounted on an ATV, chasing a car through some narrow cobblestone streets in France...amazing device (used side by side Photosonic 35mm cameras) .

So yes, I would be interested in contributing my time on such a project. My reel at www.velocity.tv/ourwork has a lot of high energy mount stuff and although I shot a lot of it handheld (among others, I built a gyro stablized shoulder mount), there is gimbal work throughout...mostly chasing things, not beauty shots.

And to answer your question above...I think that the Roll Axis is the most important and hardest to stabilize... less inertia about the lens axis (usually), and the most noticeable when not stabilized.

Thanks for the challenge.

Larry

 

[Lauri Kettunen]

Hi Larry, great feedback and footages!

Since you have experience with gimbal mounts couple questions:
- Do you know what kind of motors they use? Axial flux motors -they are disc type of motors- have high torque. A stepper motor, especially with a wormwheel, is very precise in positioning the angle, but my guess is that these major gimbals have some sort of disc type motors.
- Second, do these major palyers have two motors for the roll axis, where the other is especially for fine tuning?
- How did you build the stabilizer in your shoulder mount -what kind of motors did you use, and how do you control the system- and are you still satisfied with the solutions you made?

 

[Larry Gebhardt]

Hey Lauri. Thanks very much for the compliments. But first off, I need to clarify my background and the kind of stuff I build for myself...it is mainly mechanically and structurally engineered rigs and devices, with a smattering of electronics (very little) thrown in. For instance, the shoulder gyro mount I mentioned is a "platform" that holds the camera along with one or two Kenyon KS8 Gyros (two for heavier cameras such as the RED...one for the EX1/3 Sonys). Through experimentation, I determined the best placement of the Gyro(s) for stabilization and balance. I didn't build any control mechanisms (for example, PID types), as my stabilization is gyro based, not position sensing and error feedback with motors. Using off the shelf stuff that is easily available and relatively cheap, the beauty of this becomes the availability...nothing like it exists (for Me) for the kind of shooting I do. I satisfied that thing I mentioned before..."If i had XXXX I could shoot YYYY".

I feel my contribution to the process can be the experience I have had shooting a lot of things that move, out of or from all kinds of moving platforms like you mentioned...Cars (back of pickup trucks), boats, snowmobiles, ATV's etc. etc. I have done a lot of this stuff...then applying that experience to equipment design, which I also feel I am pretty good at. With a consortium of people with different experience, and especially if Eric is in the mix with his brilliant electronic design, it could be a really worthwile pursuit.

Now, to answer your questions...

I have never been "inside" of any of the Gimbal mounts, but I would imagine they use Pancake motors as they need quick positioning and torque at the same time (and they are very compact), and I am sure that they don't use stepper motors...too slow and way to little torque.

Then there are the two basic system designs...large gyros to stabilize the mass and small motors for positioning (Wescam, SpaceCam, etc.) and the other using small gyros for position sensing with large motors to correct position (Scorpio, Flighthead, etc.), but you already know all of this.

My sense of what you are suggesting, is something similar to the new small Scorpio head that I saw at Cinegear Expo last June...looked about right for what I think the Epic will be...and you are right, We need to know just how big (small) the Epic and whatever is necessary to be attached to it for it to function, and what can be cabled to it a ways off, to be able to be most efficient in the design of the stabilizing head. Right now I am designing a portable Jib with complete mechanical pan and tilt that will fly a small camera, hoping the Epic can be made small enough to be able to fly it on the Jib. If it isn't, then I will have to scale everything up...and then it may be too big to fly on airplanes as baggage, etc. We have the same issue...don't really know much about the payload.

Yes, I am very satisfied with the shoulder gyro mount. I used it on a Jet Ski commercial last year and had a lot of very complimentary comments...There is an advantage to having your eye to the finder, and the other eye open as a huge "see-around" device. Try a swish pan into a shot using a remote joystick or such...

Lauri, Where are you located?

Larry

 

[Michael Hastings]

Lauri:

Interesting idea. I make underwater cases for cameras. I'm a self-trained machinist (with the scars to prove it) Our shop has typical small and mid-size lathes and mill as well as a Haas VF-1 CNC. I don't claim to be too expert on the CNC side - normally do pretty basic stuff - but may be able to contribute in making prototypes. I've always been interested in stabilization and I like the idea of a compact Epic with Birger mount- so willing to throw some time and materials at it.

 

[Lauri Kettunen]

Interesting idea. I make underwater cases for cameras. I'm a self-trained machinist (with the scars to prove it) Our shop has typical small and mid-size lathes and mill as well as a Haas VF-1 CNC. I don't claim to be too expert on the CNC side - normally do pretty basic stuff - but may be able to contribute in making prototypes. I've always been interested in stabilization and I like the idea of a compact Epic with Birger mount- so willing to throw some time and materials at it.

Michael, thanks for chiming in. I'm still thinking of buying one of your underwater housings, but have postponed this until Epic-X and Scarlet is available. Erik, Larry, now you ... this starts to sound much better than what I expected earlier today when writing the first post.

Larry: I'm located in Finland, so not that close to Northern America. (Still, my record is 8 hours from O'hare to home. Tail wind, early landing and catching the earlier flight in five minutes after landing.) I've never had a chance to take a good look in the best gimbals so everything you tell is very wellcome and confirms quite a lot of what I've concluded.

It seems we have to think of at least so called five-axis stabilization (what a misleading term!), perhaps pan-axis manages with a single motor. I checked the Scorpio pages, but my gut feeling is that the basic mechanical principle (forget the belt drive etc., just focus on the meachanical support) employed e.g. here

http://www.copterworks.com/Products/ThreeAxisCameraMount/tabid/60/Default.aspx

results in a more simple control system and thus higher quality stabilization. For, such an inner ring for the most critical roll-axis makes results in a symmetric setting; Whatever is the roll angle, the system appears the same from the controller point of view. In addition, centralizing mass to the point where the three axes intersect should be easy with this kind of setting. Furthermore, creating a dynamic mechanical model of such a symmetric system for the controller algorithms (including friction etc.) seems to be possible. What do you think?

Some year ago I designed a pan-tilt head for a crane and used Maxon motors. Have found them pretty good piece of equipments. Something like these

http://www.maxonmotor.com/product_overview_maxon_flat_motor.html

might be the way to go on.

I've been seacrhing for gyros for some time, and I think the best ones I've found so far is DSP-1500 or DSP-3000 from KVH:

DSP-1500: http://www.kvh.com/Products/product.asp?id=165
DSP-3000: http://www.kvh.com/Products/product.asp?id=83

DSP-1500 is very compact and light whereas DSP-3000 has a very small drift. Three gyros are needed and these cost around $3500 each. There are MEMS gyros that are in the range of $200-$1000, but not sure how well they work in this kind of system. That needs to be tested later on.

 

[Gilles Debord]

Hi Lauri

For the same application that i develop i use a 3DM-GX2 Microstrain IMU 3 axis XYZ, 3 accelerometers, 3 gyro, 3 magnetometers (your Suomi dealer is (www.lisab.se) Sweden), For the time i develop it with servo motors drived in PWM (not the best) just to verify the kinematic on pan/tilt axis. My software is Delphi 6 based and run on an Eee PC I have to finalize the system this month and correct the bugs.

In my sytem all is geared without belts (module 1 gear)

I can reverse the movement and limit the "angled movement" speed limit is 180°/sec torque depending , i think to add the possibility to inhibit axis.
The head reproduce in speed and displacement the movement of the IMU (inertial measurement unit) if you work with this kind of produce don't use the "Inertialink" it is without magnetometers and you cannot control the yaw axis.

Steppers motors is a good way but you can loose steps and you have to add an angular feedback on the axis, there are good one with 4096 increment or more.

Have a look on the Microstrain.com website and read the white papers.

Best regards

Gilles

 

[Tom Hamilton]

Greetings Lauri, All,

I've been looking at various potential stabilization solutions for Scarlet, with a view to doing some aerial work with our Cessna 182 aircraft for an upcoming nature doc.

Judging from the considerable talent already posting on this thread, I'd be happy to contribute the use of our aircraft for eventual testing, liasing with our FAA contacts for necessary permits. It alternates between Florida, California and Baja, depending on the time of year, but can be flown anywhere in North America for this purpose.

Best,
Tom

 

[Constantine Tyrintzis]

Hello. If you talk gyrostabilized head is a different story , but if you need something that is light can be shoulder mounted and configured with Kenyon gyros which are proven to be durable and excellent , have a look at my new product, the ACROBAT.
Made for Epic and Scarlet. Parts are being made at the moment, I had to apply for patent , so it took me a while to release. you can just about view a render with a loan photo of Epic so you can get what it will look like.
The price is very friendly as well and you wont break the bank.
coming up 2010.

 

[Lauri Kettunen]

Gilles, Could you please send me an email or pm.

Tom, thanks for the offer. I'll keep it in mind when we get to the point of testing.

 

[Gilles Debord]

Hi Lauri


Send me an e-mail with your questions here: info@wwvideo.tv



Regards

Gilles debord

 

[Tom Hamilton]

Red DSMC Oct 30 Announcement

Red DSMC Oct 30 Announcement

Relevant to this thread:

3 axis internal motion sensor
Dimensions- Approx. 4”x4”x5.5”
Weight (Brain only)- Approx. 6 lbs (2.72kg)

Any thoughts?

 

[Constantine Tyrintzis]

And this is what I mean, just to get a picture of it with the Epic.
All renders of course. And there are many more places you can place at list one of the KS6 , for more effective pan stability, like in the place of the shoulder mount which in this case is not needed..Also 2 handle bars instead of one cam be added on the rods..
A system like this would cost you around 5.500 -6000Euros together with 2 KS6 gyros. And though its designed for Epic Scarlet, with different configuration and placement it can work for the red one as well as other cams...

 

[Cliff Gill]

Lauri,

Your idea of an economical gyro stabilized mount is excellent, sort of an open source gyro stabilization project?
I have stumbled across this thread in search of something very similar, although for a different application, mine is motion stabilization for vertical aerial photography, rather than cinematography, but the main points are very similar.
I am seeking a mount very similar to the following:
http://www.aerotopol.de/html/aerostab.html
but that would have fewer of the additional features and be less expensive.

I am not an electrical engineer, but I might be able to contribute other services. I do have a microstrain 3DMG gyro as Gilles has speced as a possible gyro, and I also have a Novatel SPAN system which is a very high end gyro (around $40,000) which might be useful as a reference. I also have a few pan-tilt systems, similar to the copterworks one mentioned previously but from servocity.com
The primary element that I am missing for my own project is programming talent to integrate the output from the gyro into a controller for the servo's or motors in the eventual gimbal. I am happy to contribute some equipment and possibly even some funding to get a gyro controller suitable for my application.
Some useful information about different gyro systems and design is at:
http://www.camerasystems.com/gyrostabilization.htm
Are you thinking specifically about an "A" system or a "B" system in reference to the link and description above?

 

[Tom Lowe]

It's funny because Larry Gebhardt and I had a similar conversation to this, what, two weeks ago?

There is a huge need for lower-cost gyro heads for cameras like Red is building.

The members of our Timescapes timelapse forum are working on a somewhat similar project called "OpenMoco" that will hopefully open up advanced motion-control rigs to noobies and the general public. All the software and code is being programmed in the open, while various types of hardware (mostly put together from off-the-shelf parts) are being tested as platforms.

 

[Djordje]

Hi Lauri, this is a great topic,and I will follow it,although I have no Knowlege of this,so I can help in some other way.
I have a question..Are you thinking about stabilizing camera only, or there is room for tracking data from camera and using it for virtual camera also.. It would be good to have that option too...or is it more complicated. ..

Once again, This is a great idea and I hope a lot of redusers will jump in ...