Need help with a project (AC motors and programming)

Started by
8 comments, last by capn_midnight 9 years, 6 months ago

Sorry if this is the wrong forum....

Im trying to get more information on a device Im trying to put together. But I dont want to start throwing money at this idea until I know what needs to be done. So I figured this would be the best place to address this question..

Im looking to build a motion control rig. Basically an electronic arm with 4 or 5 degree's of freedom, its purpose is to hold a camera and it would have the ability to keep repeating the same exact movements as many times as needed. Its a way to composite multiple things into the same moving shot without having to rely so much on cgi.

But im also kind of in the dark about a controller. What I want to do is use bezier splines (just like the fcurve editor in blender) to control the path for each ac motor. The shaft turns one of two ways so having the motor communicate with a computer somehow by means of splines is what Im trying to do. (there are a few spline editors available online that are stand alone apps, so I dont think using blender would be necessary if it can all be done independently).

But heres the problem, im not skilled or that knowledgeable in electronics or programming (I know, it sounds silly that Im pursuiing something so unrealistic). But thats why Im just going on research Ive already done.

So far the things I THINK I will need are,...

- AC motors

- VFD (variable frequency driver) controls the speed and direction of the motor shaft.

- A pc that connects to the vfd to communicate.

Now Im not sure how to go about something like this. I would imagine that a vfd would have to be ripped apart to get access to its circuit board. But thats why Im asking all the questions I can before I jump head first into this. And its obvious at this point that I will have to hire an engineer or 2.

But if someone could point me in the right direction or give me some advice it would be greatly appreciated.

Thanks.

Advertisement

You may want to go looking for a robotics forum rather than gamedev. I'm sure there is tons of overlap of people who are interested in both fields, but robotics of that level is a very tricky bit of technical issues to deal with.

Some of the questions you need to ask yourself is what kind of speeds are you aiming for, what is the acceptable minimum accuracy of the device, just how large of a camera are you going to deal with, and how large are the movements you're going to make?

I suggest minimizing the 'reaching arm' effect. Keep your forces as in line and stable as possible. The more the robot vibrates, the more your accuracy and precision will tank from shot to shot. The heavier the camera gear is and the faster you move, the lower your overall stability and harder it is to control.

Running the camera up and down a vertical shaft that is then moved back and forth, side to side, and rotated around its vertical axis may be your best bet.

Old Username: Talroth
If your signature on a web forum takes up more space than your average post, then you are doing things wrong.

You may want to go looking for a robotics forum rather than gamedev. I'm sure there is tons of overlap of people who are interested in both fields, but robotics of that level is a very tricky bit of technical issues to deal with.

Some of the questions you need to ask yourself is what kind of speeds are you aiming for, what is the acceptable minimum accuracy of the device, just how large of a camera are you going to deal with, and how large are the movements you're going to make?

I suggest minimizing the 'reaching arm' effect. Keep your forces as in line and stable as possible. The more the robot vibrates, the more your accuracy and precision will tank from shot to shot. The heavier the camera gear is and the faster you move, the lower your overall stability and harder it is to control.

Running the camera up and down a vertical shaft that is then moved back and forth, side to side, and rotated around its vertical axis may be your best bet.

Thanks luckless. I looked into robots before but wasnt entirely sure. But I guess the robot forums are the place to ask around.

I appreciate your help, thanks.

Reads like a typical industrial motion control application. When I was doing similar things with optical equipment (lasers, crystals, mirrors, gratings, diodes), I used dc or stepper motors w/ gear box. The controllers were usually commercial products (Newport,Compumotor,Galil...) though sometimes I used boards built in house. Don't know why you think you would need to modify the boards.


having the motor communicate with a computer somehow by means of splines

I don't know what that means. I think you refer to some kind of position or velocity feedback.

Might be helpful for you to look at commercial products to get a feel for what you can get off the shelf. Most of those controllers have a way to compose complex motion with multiple axes. Also, look at sites for homebrew robotics, lots of info there.

The Four Horsemen of Happiness have left.

Reads like a typical industrial motion control application. When I was doing similar things with optical equipment (lasers, crystals, mirrors, gratings, diodes), I used dc or stepper motors w/ gear box. The controllers were usually commercial products (Newport,Compumotor,Galil...) though sometimes I used boards built in house. Don't know why you think you would need to modify the boards.


having the motor communicate with a computer somehow by means of splines

I don't know what that means. I think you refer to some kind of position or velocity feedback.

Might be helpful for you to look at commercial products to get a feel for what you can get off the shelf. Most of those controllers have a way to compose complex motion with multiple axes. Also, look at sites for homebrew robotics, lots of info there.

Position and speed control is what I meant by spline control. Sorry I should have explained that better. I asked some robot programmer girl on fiverr about this stuff about a month ago (needless to say she took my money and hauled ass), but I made this video for her that explains how I wanted to use a spline editor to for position control.

(if the link doesnt work let me know, I have it locked.)

BTW, I wasnt talking about modifying a board. Just hacking it so to speak so I can repurpose it to be used with software. But im just spitting out idea's, I dont know how or if any of that would even be possible.

Is there a difference between dc, steppers, and servo motors?

The video doesn't work for me -- says it is private.

Servo implies that there is an encoder on the shaft. Motor can be any kind.

Steppers have distinct coils, so with appropriate driving signal there are discrete stable points along a single rotation - steps. Not the smoothest motion but cheap, easy and can bear a load -- with these you move things into position, then let everything settle down before doing something. And you don't need an encoder because you can keep count of the steps, assuming the mechanism is well made.

If you want to do something while in motion then you probably want a dc motor. Smooth motion but more work and expense typically. Harder to hold still under load. Needs an encoder on it if you want to know position.

I've remembered since my first post that once programmed a multiaxis controller made by a company called Newport that would coordinate along a defined path. It would interpolate as needed, you just defined points and time. You don't need any custom software or hardware -- I bet you could do it yourself if you had the manual.

The harder part might be the mechanics of the device you have in mind.

The Four Horsemen of Happiness have left.

Reads like a typical industrial motion control application. When I was doing similar things with optical equipment (lasers, crystals, mirrors, gratings, diodes), I used dc or stepper motors w/ gear box. The controllers were usually commercial products (Newport,Compumotor,Galil...) though sometimes I used boards built in house. Don't know why you think you would need to modify the boards.


having the motor communicate with a computer somehow by means of splines

I don't know what that means. I think you refer to some kind of position or velocity feedback.

Might be helpful for you to look at commercial products to get a feel for what you can get off the shelf. Most of those controllers have a way to compose complex motion with multiple axes. Also, look at sites for homebrew robotics, lots of info there.

Position and speed control is what I meant by spline control. Sorry I should have explained that better. I asked some robot programmer girl on fiverr about this stuff about a month ago (needless to say she took my money and hauled ass), but I made this video for her that explains how I wanted to use a spline editor to for position control.

(if the link doesnt work let me know, I have it locked.)

BTW, I wasnt talking about modifying a board. Just hacking it so to speak so I can repurpose it to be used with software. But im just spitting out idea's, I dont know how or if any of that would even be possible.

Is there a difference between dc, steppers, and servo motors?

sorry, the vid should play now.

Cool, I wasnt sure what the difference was between AC's and DC's either. I just assumed one ran off battery while the other used an electrical outlet. Do you think it would be in my best interest to go ardruino for a start? From the looks of it its designed to be programmable. It wouldnt be for a final product because im looking to build something more heavy duty, but it seems like a nice learning device.

But thanks jms,

And thanx to all the gamedev community too. you guys are great.

How much are you willing to spend on this project? And what is your goal? Are you thinking you're going to save money by doing this yourself, or are you trying to learn? Because if it's just about the money, you're going to be extremely hard pressed to make something cheaper than you can buy, and that's assuming you're going in already knowing what you need to do. Something like this probably costs in about the $5000 range. And it will probably cost you 50% more to try to build it yourself.

Because it's just full of really expensive parts. And those parts are expensive because they're the only ones that are good for the job. When you get into precision control systems, you need a physically large machine (the weight reduces shaking and vibrations) with very powerful motors (to counter the momentum of the payload) and precision milled components (to reduce friction and try to reduce backlash). You'd have to get all of that at retail, and custom made. It gets expensive. And hopefully, you get it right the first time and don't have to redesign, or don't drop and break your camera in the process.

Now, if all you want to do is make something that can control a small webcam, then we can talk. If you use stepper motors as you said, you'll need a lot of geering to divide their movement down to small enough steps to matter for your application. With the right driver circuit (a phased sine wave) you can drive stepper motors smoothly, but their strongest locked position will be limited to only a few, discrete positions.

It's easier to use regular DC motors and optical encoders. It's the "optical" in optical mouse, the mouse ball rotates two clear, plastic disks inside the mouse that have black lines printed on them, and an optical coupler to count the number of lines that interrupt its beam. Also, many inkjet printers use a similar system, with a linear strip of clear plastic instead of a disk. The stepping precision is as good as your gearing (backlash again, but significantly less gearing than with steppers), your optocouplers, and you encoder strip. This way, you can keep all of the control circuitry digital, rather than having to muss about with analog currents. While analog is fun, it's not easily controllable from a digital context without increasingly more expensive hardware.

You'd want high-torque, low-speed DC motors. Use gears to adjust the speed as necessary. You might want to look at a variety of geometries for arranging the parts. Delta robots are rather interesting.

[Formerly "capn_midnight". See some of my projects. Find me on twitter tumblr G+ Github.]

How much are you willing to spend on this project? And what is your goal? Are you thinking you're going to save money by doing this yourself, or are you trying to learn? Because if it's just about the money, you're going to be extremely hard pressed to make something cheaper than you can buy, and that's assuming you're going in already knowing what you need to do. Something like this probably costs in about the $5000 range. And it will probably cost you 50% more to try to build it yourself.

Because it's just full of really expensive parts. And those parts are expensive because they're the only ones that are good for the job. When you get into precision control systems, you need a physically large machine (the weight reduces shaking and vibrations) with very powerful motors (to counter the momentum of the payload) and precision milled components (to reduce friction and try to reduce backlash). You'd have to get all of that at retail, and custom made. It gets expensive. And hopefully, you get it right the first time and don't have to redesign, or don't drop and break your camera in the process.

Now, if all you want to do is make something that can control a small webcam, then we can talk. If you use stepper motors as you said, you'll need a lot of geering to divide their movement down to small enough steps to matter for your application. With the right driver circuit (a phased sine wave) you can drive stepper motors smoothly, but their strongest locked position will be limited to only a few, discrete positions.

It's easier to use regular DC motors and optical encoders. It's the "optical" in optical mouse, the mouse ball rotates two clear, plastic disks inside the mouse that have black lines printed on them, and an optical coupler to count the number of lines that interrupt its beam. Also, many inkjet printers use a similar system, with a linear strip of clear plastic instead of a disk. The stepping precision is as good as your gearing (backlash again, but significantly less gearing than with steppers), your optocouplers, and you encoder strip. This way, you can keep all of the control circuitry digital, rather than having to muss about with analog currents. While analog is fun, it's not easily controllable from a digital context without increasingly more expensive hardware.

You'd want high-torque, low-speed DC motors. Use gears to adjust the speed as necessary. You might want to look at a variety of geometries for arranging the parts. Delta robots are rather interesting.

well I guess I was under the impression that building a homemade system would save me money and to be honest its more of a hobby experiment right now. But I do intend to use this for some film projects.

Im not really all that versed when it comes to the mechanics either. I mean I had the mindset that anything can be rigged up and if it works it works right? Ive checked out so many of these electronic camera rigs (especially the one you linked me to) and I have to admit the set up looks intimidating. I can only assume that stuff was designed by qualified engineers. and each piece serves a purpose.

.....Going back to what you said about something heavy duty being more stable, thats why I mentioned the use of AC motors. Im assuming bench grinders use ac motors. Not sure but if they do thats the type of stuff im referring to. Ive seen how much HP they can generate. Pretty strong motor.

Nope, not webcams. The particular type of cameras I had in mind are dslr's. But also built to support the added weight of other lenses like anamorphic.

Now heres another question..... AC vs DC. Will both these motor types give me the same amount of HP?

Theres the devices jms mentioned, the newport multiaxis controllers. I looked into those a bit. Could be wrong but as far as I can tell they are built for everything but AC's.

Theres so much I dont know about all these motors and the digital/optic/encoder stuff you mentioned too. So I guess I have my homework cut out for me.

BTW,, Just to get a human response on this without having to figure out a bunch of technical info,... what are encoders?

and your advice is def much appreciated. thank you man,

There is nothing special about AC for making motors. For example, the Tesla Model S uses a DC motor. The only reason certain common devices like floor fans or kitchen appliances use AC motors is because the electricity in the wall is AC (and even that was an arbitrary choice, there are many reasons why DC would have been better), making for smaller devices that don't need a lot of circuitry for power rectification. But since you want a control system, you'll need rectified (converted from AC to DC) power anyway.

You probably won't need extremely powerful motors. In fact, I bet even a quarter of a horsepower is way too much. It will be far more important to make sure you get high-torque, low-speed motors. Two motors could have equal horsepower, but one achieves it with low torque and high speed, and the other achieves it with high torque and low speed. Either one can be made to work, but the high speed motor will require a lot of extra gearing, which is difficult to get right and is energy inefficient, adding friction to the system. Think more like an electric screwdriver than an RC car.

Encoders are basically just counters. You have a small, black box that has a slot in it, almost like a mini credit card reader. In the box, a beam of light passes through the slot, to a detector. Putting something in the slot and breaking the light beam triggers the detector. If you make that thing a clear piece of plastic with black bars printed on it at regular intervals, you can count the number of ticks that go by and correlate that to a distance that has traveled, or a certain amount of rotation, or whatever else you're trying to measure.

[Formerly "capn_midnight". See some of my projects. Find me on twitter tumblr G+ Github.]

This topic is closed to new replies.

Advertisement