Belt drive rotary axis.

[Art Ransom]

I want to increse the ratio on a rotary axis to range between 20:1 and 50:1. I will be using stock up to 100 lbs at RPM up to 50 RPM in lathe mode and 5 RPM in indexer mode. Using a jack shaft and 3/4" belt what minute of arc tolerances could I expect?

[keebler303]

If you have the belts tight, there is almost zero backlash at zero load. As you load it, the belts stretch and you get backlash. How much, not sure, that would depend on the type of belts you are using and the loads your setup will see. Probably a question for an application engineer who sells the belts. Also certain belt profiles are more prone to backlash than others, might check into that as well.

Have you considered a low backlash worm drive? They will easily give you the ratio you need and can have very low backlash. Might also check out ebay for planetary drives. About 6 months ago, I picked up a big nema 42 planetary for something like $30. Its about a $2,000 piece and in good condition. Anything with gears will give you more rigidity as long as the backlash is low.

Good Luck
Matt

[Mike Everman]

Belts are zero backlash, but only if you define backlash as a "zero stiffness deadband". I have 4:1 reductions here using Gates GT2-3mm (the style that would be best for indexing accuracy, according to Gates). People do not normally encounter the error in a belt reduction. Simply put, there is tangential tooth clearance that "walks around" as the pulley turns, such that teeth march from contact with a forward face to a rearward face. It's not a small amount, even on this one.

The reduction I have is on a large camera pan-tilt, where indexing accuracy had to be less than .12 deg TIR. Plain Jane out of the box, standard clearance pulleys are going to get you that or more, so call it 7 arc minutes!

The way to get rid of that is to order zero clearance pulleys (not widely available). In my case, the pinions can be found low or zero clearance, (and it's far less an issue at low tooth count), but the bigger pulley is hobbed to order.

Engineering at Gates stated the obvious: the pulley needs to be hobbed to a larger OD, effectively thickening the teeth to reduce clearance. They gave it a value of .004" over nominal on the diameter. The first 200 or so were hobbed to nominal and were blowing my spec. The quick fix, and you may get away with this generally in a DIY project is to cut down the OD, which buries the belt teeth in the bottom of the pulley grooves. Not an ideal mesh for maximum life, but it works. The next best way is to have the pulley plated with something like .002" build-up to eliminate the clearance.

I do agree that a worm drive would be ideal. I'd buy a cheap manual rotary indexer and retrofit a motor to it, and spend some time adjusting out the backlash occasionally.

[Robert M]

Hi Art,
In your quest to make this 20 :1 and 50 :1 ratio, can you share some data / specs on the material your focusing on ? ( belt & pulleys type, brand, size…etc) !
I’m asking, as you may remember, I too will make this MM with an indexer and been searching my head off for some of those !
Amicalement, Robert M
PS your web site is off ??