What size servos will I need?

[wannaBinventor]

I've got a 3 axis accelerometer going that I'd like to combine with three servos to make a mount for a camera that will move in relation to the angle of orientation to gravity.

Ultimately I'd like to mount it on my motorcycle so that the camera will keep itself centered on the horizon no matter what angle I'm leaning at or wheelie-ing at.

The camera weighs about 11 ounces. What do I need for the servos in terms of ounce/inches (kg/cm). Depending on the power, it looks like servos can be up to 2 or 3 ounces in weight alone, so please consider that as well.

Also, I'm wondering if anyone has advice on making the system ignore the G forces associated with the centripetal force of going around turns at speed. That's where I'm really puzzled.

 

[RMMM]

It would figure a steadi-cam jig would work better than an electronic approach.

I had a friend who was involved in a documentary filming for high-speed go-kart racing and the filmers used steadi-cam jigs mounted on different areas of the go-kart frame.

 

[wannaBinventor]

While those looks really nice for the application, I have two problems.

First, they are expensive. I'm thinking minus the camera I could have this rig set up for less than $100. Secondly, I want to "cool factor" of having done it myself.

Thanks for the suggestion though. I'm sure it is really the best for what I intend to do.

 

[Mr RB]

Ultimately I'd like to mount it on my motorcycle so that the camera will keep itself centered on the horizon no matter what angle I'm leaning at or wheelie-ing at.

You need to add gyros for that job, as an accelerometer-only solution will lean with the bike for the same reason the bike leans.

Check this Pololu product page;
Pololu Gyros and Accelerometers

Since you have the accelerometer you could get the $19 dual axis gyro kit, and you will need a bit of processing power to handle 3 accel + 2 gyro in real time.

Or for $125 they have a full 6d inertial measurement unit that has a built in 32bit processor and spits out nice simple TTL serial data saying what it happening in terms of tilt etc.

But either way this is not a trivial project to throw together, especially since you are coding in assembler (you said in your other thread) and haven't used accels or gyros before.

 

[wannaBinventor]

Okay, thanks. I was starting to get the feeling I'd need a gyro to cancel out the centripetal force. I have a three axis gyro waiting to be put on a PCB that I could fire up this weekend, perhaps.

To avoid having to learn a beefier uC (I have only ever used 16F Pics), I may just use dual PICs for the setup -- some of the enhanced midranges that is that can run at 32MHz.

I think I'll use one for all the sensor measurements. I'll only use pitch and roll on the gyro, because if you are yawing on a bike something has gone terribly wrong. I'll do most of the math on one and send data over SPI at a quick clock to the other PIC that will control the servos, because I've gathered that precise servo control of multiple servos can be a fairly processor intensive undertaking. For that matter, I think I'll only need 2 servos to adjust the camera for pitch and roll.

 

[Mosaic]

U tryin' to build a Steadicam?

 

[wannaBinventor]

I didn't even know those things existed until this thread. The prices are much too high for something that I'll just play with on my motorcycle.

It seems that those things are mostly mechanical mechanisms, which is something that I have no knowledge of beyond one college physics class. Plus, you can hit some pretty extreme angles on a motorcycle. I'm not sure but I'd imagine that you can hit 65 or 70 degrees of lean in a fast turn, and about the same degree of inclination on a big wheelie. It seems like some of the cheaper (read 800 bucks) steadi-cam units wouldn't be able to handle that.

Any advice on servo torque requirements?

 

[RMMM]

Any advice on servo torque requirements?

It all depends on the weight of the load (read: camera and mount connected to motors(servos))

Then you can calculate the general maximum force that it will experience. Say 5g's for instance.

You can then calculate the newton forces required to move the load AGAINST the counteracting force.

As for the steadi-cam, a few you tube videos may change your mind. For under $20 in parts, you can make a pretty good steadi-cam rig.

The "professional" rigs connected to the go-karts were plywood, plumbing parts, and some bricks with holes in them for counter weight.

Worked, REAL well.

For a motorcycle, you are dealing with wheelies and such that a go-kart need not deal with.. But still, look into it. You may have an "Ah-HA!" moment and come up with the "perfect" rig for you, and other motocross/cycle filmers.