Ideas for rotating head

I've taken a first stab at a rotating head for my bot, and am looking for ideas for improvements. First, some pictures:
http://www.westnet.com/~chris/Robots/RobotHead-1.jpg
http://www.westnet.com/~chris/Robots/RobotHead-2.jpg
http://www.westnet.com/~chris/Robots/RobotHead-3.jpg
The center shaft is made from two extensions for old wheel-style electrical boxes. This gives plenty of room for wires to thread down and not get tangles. What you can't see under the rotating top is a hardware store lazy susan bearing. The square plate on the top screws directly to the bearing. I used this plate to line up the shaft on the center of the bearing. This way the shaft would at least be centered, even if the table itself was slightly off.
Picture three show how I motorized it. I essentially made a digital servo of the whole thing with a motor and three-turn pot connected to the center shaft via a cog belt and pulleys. A quick hack on the positional PID code that drives my bot runs the motor through one of the small TI h-bridge chips. The control works quite well.
The problem is belt tension. The motor and pot are mounted on a U shaped bracket (actually the cover from an dead CD-ROM drive). I set the belt tension with over-sized holes in the mounting bracket, and adjusting it's position relative to the center wheel.
However, this is enough error in centering that as the head turns, the tension changes on the belt so it either goes slack, or becomes too tight for the motor to turn.
So before tinkering further I thought I would get some input. One idea I had was to mount the pot on a spring-loaded lever and use that for tension. Increasing precision is always an option, but I thought I did the best I could this time, so I'm not quite sure how to go about that. Or, do something else entirely.
--
==========================================================
Chris Candreva -- snipped-for-privacy@westnet.com -- (914) 967-7816
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
Christopher X. Candreva wrote:

Where does the error come from? Is the shaft wobbly, thus making the rotating platform unstable? Is the drive mount hardware, specificaly what looks like thin metal holding the motor in place, too flexable?
While there are a number of issues with the design, if you can code around the inaccuracies, you should focus on correcting the mechanical slack. Basically correct the mechanics so that it doesn't get too tight or too loose when it turns.

I don't know what you are using to support the rotating platform, so it is hard to really say.

Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
: I don't know what you are using to support the rotating platform, so it is : hard to really say.
A hardware store lazy susan bearing -- this: http://www.herbach.com/Merchant2/merchant.mv?Screen=PROD&Store_Code=HAR&Product_Code=TM92MEC2204
Rather than try to center the shaft on the whole platform, I decided to motorize the lazy susan as a separate problem. I reasoned that the shaft need only be centered on the bearing to work. The platform itself could be slightly off-center and give only a visual problem, so that was less critical.
I cut that square plate you see to the same size as the lazy susan, found the center of that (standard, low-tech, draw an X corner to corner), and mounted the shaft to the plate. The plate screws through the platform to the bearing in the corners.
: Where does the error come from? Is the shaft wobbly, thus making the : rotating platform unstable? Is the drive mount hardware, specificaly what : looks like thin metal holding the motor in place, too flexable?
I think it's a combination of any error in my centering of the shaft, any slight angle the shaft as to the plate, and the cog-gear has on the shaft itself. The metal holding the motor is actually steel (it was an OLD CD-ROM drive) and doesn't give at all.
: While there are a number of issues with the design, if you can code around : the inaccuracies, you should focus on correcting the mechanical slack.
If I put the tension to the point where it doesn't go slack at any point, at the tightest point it won't turn with the motor hooked directly to power, so I didn't see a way to code out of it.
The other solution I didn't mention was a bigger motor, but brute forcing it didn't seem like a great idea.
--
==========================================================
Chris Candreva -- snipped-for-privacy@westnet.com -- (914) 967-7816
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
On Fri, 20 May 2005 14:15:48 GMT, "Christopher X. Candreva"

In most situations like this a spring loaded idler wheel is run against the belt between the motor and the pulley to keep ~ the same tension on the belt while allowing it to account for off center pullies and such. The motor can also be put on a hinged mount with a spring keeping it tensioned on the belt allowing the motor to move some to account for the pulley being off center. If you are using a cog belt, I'd look into using one of the big quarter scale servos to turn the lazy susan (but wouldn't have the power of the motor setup).
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
Christopher X. Candreva wrote:

There are a couple of changes I'd suggest.
The most critical is the radial load being applied to the potentiometer. Even timing belts require a certain tension, and the tension must be increased as torque increases. As torque increases, so does the overhund loads on all the shafts. I wouldn't worry about the motor, but the pot may die an early death if the load is too high. It'll start getting flakey, and then eventually go out altogether.
So I'd recommend doing two belts: a drive belt between the motor and turntable, and a separate one from the turntable to the pot. The separate belt can be a smaller cogged belt. Because you probably have more flexibility regarding the length of the turntable shaft, you might mount the sprocket for the pot belt there, instead of on the motor. You can add a spring-loaded tensioner, if you need to, on the motor-to-turntable belt, and it won't affect accuracy or load on the pot.
You might also want to think about scrapping the metal mounting thingie and create a mounting plate out of 1/8" plastic, like ABS or polycarbonate (PVC would probably be too soft unless it'sx 1/4", and acrylic cracks too easily). Mount some aluminum standoffs of the desired length. That ought to be a lot more rigid than the metal frame you have now. The flex in the sheetmetal might be causing some of your problems. You can also afford to remake the mounting plate a couple of times to improve accuracy. Cut out a couple squares, drill the mounting holes in all of them, and then try out a few until you get it right.
If the shaft to the turntable is slightly off center, you'll get a misalignment that may be what you're seeing. I think you might be able to mitigate some/all of this misalignment if you enlarge the holes on the turntable so that you can adjust the centering of it over the lazy susan flange. Loosen the screws (they should have large heads, or use fender washers), adjust the centering, retighten, and try again.
One way to check for alignment is to remove the center shaft, and shine a flashlight down the hole, and onto the floor. Spin the turntable. If the spot on the floor wobbles, the turntable is not centered on the lazy susan.
-- Gordon
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
    --Couldn't find a photo of Carl's commutators, but here's a drawing from his site that you might find useful.
http://www.carlpisaturo.com/diagrams/OctopusStudyB_Main.JPG
    
--
"Steamboat Ed" Haas : The other night I
Hacking the Trailing Edge! : dreamed about wasabi...
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

had
That would be a way to fix your problem, but as Gordon pointed out, it will eventually kill your pot. Modify your idea slightly and I think you'll have a winner...
Use your idea of a spring-loaded lever as a tension arm, but run the belt to an idler wheel, not the pot. Move the pot farther up the tension arm and run a second (obviously much smaller) belt (like a thick rubber band or a large O-ring?) between the pot and the idler wheel. This will allow the arm to keep a constant tension on the drive belt, but not subject your pot to the strain of that tension, which would ruin it. The pot and idler wheel will always be a fixed distance apart so you can make the tension between them just enough to work without putting undue stress on the pot.
Alternately, you can try an entirely different method - put a large encoder disk (maybe made from a large painted can lid?) on your shaft and read the position that way - eliminate the pot entirely. I'd still use the tension arm, but it would be a simpler design. It means redesigning some of your control circuits, but you may gain more functionality, so there are pro's and con's to the idea...
~WEC
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
Thanks to everyone for the ideas. I hadn't thought of the load on the pot being a problem, or at least didn't think of the load being that much of a problem.
In fact, I've already someone damaged this pot putting the gear on. It's such a tight fit, I pulled the shaft loose trying to pull the gear off again, so the shaft now slides up and down in the pot. I was surprised to find it still worked after that, but it does, and I've been using it since I figure I'm going to destroy it completely if I ever try to take it out now.
I'm going to try to set aside a night of tinkering, with a printout of all your suggestions, and see what I can come up with.
--
==========================================================
Chris Candreva -- snipped-for-privacy@westnet.com -- (914) 967-7816
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

Polytechforum.com is a website by engineers for engineers. It is not affiliated with any of manufacturers or vendors discussed here. All logos and trade names are the property of their respective owners.