Using timing belts horizontally

My team has been doing some prototyping this off-season and we have taken a great interest in trying to tackle belts. Although our team used belts in the drive system in 2008, the knowledge gained from that experience has been lost due to no need to use belts again.

With the general use of timing belts becoming more common, and the KOP drive transitioning to using belts, our team would love to have the ability to use belts instead of chain.

Which leads me to my question, how well do belts hold up when they are used horizontally. (IE running belt between two vertical shafts) Would they slowly sink over time due to gravity? I know some pulleys have a lip on either end to prevent the belt from slipping, but will it prevent the timing belts from sinking? Are there any special techniques to use when running belts horizontally? Has it ever even been done in a FRC setting?

Would attaching a guard to bottom side of a sprocket prevent slipping?

_ |||||_

___ |||||___
(Sprocket with guard on bottom)


Though it seems Gravity would pull it down to stretch the belt it is Extremely Hard and nearly impossible for them to stretch, little own stretch due to gravity. You will not have an issue with belts being horizontal.

You can plan to run them in any orientation you want as long as you have a means of tensioning them, either with movable shaft location(s) or a tensioner/ idler. As long as they are not slack and have reasonable tension (and of course the shafts are well parallel), the pulley flanges will keep the belt tracking. Only a loose belt would climb over the pulley flanges from gravity.

How tight is ‘reasonable tension’? How much force should we use when tensioning drivetrain belts? How much tension is too much? Will floating idlers work?

I’ve been running horizontal timing belts for a little bit, off a FRC robot. I don’t think you would need it quite this tight, but for the application I’ve been using them in, if the driving motor started to turn, the belt needed to start to move immediately. If there was delay, I’d hunt down and remove the slack.

For an FRC robot, you don’t have to have them that tight, but you still don’t want much slack.

You wouldn’t happen to have any pictures would you?

Not knowing the geometry of your situation its impossible to describe exactly how much tension is right. There are rules of thumb for the amount of deflection when pushing on the belt from the side, but this depends on the belt arrangement, width, torque, etc. The least amount of tension which transmits your load reliably and does not skip under max. load is the right amount. Any tension that is wrapping the pulleys securely and not slipping teeth will also keep the belt tracking. Use a method that is adjustable and just experiment with it. When possible I always prefer an adjustable shaft location with no tensioners vs. tensioner idlers, as idlers add moving parts and complexity and subtract from your efficiency. If you use an idler and are transmitting power in both directions, it should have a fixed location. Spring loaded or “floating” idlers are for power in one direction, and are located on the slack side of the belt path.

The manufacturer has a specification for tension. That would be the best choice. A floating idler may not provide enough tension for a severe application such as a drivetrain because it won’t let the belt be tight enough

I don’t think so. I’d have to get them the next time I’m in the shop (the best part of a week).

BUT! of the three belts on any one of the machines they’re on, 2 are horizontal (the third is on horizontal axles) and only one of those is a continuous belt. That one has a slidable tensioner–set and forget–and is used to drive a pair of leadscrews. I’ve seen minor stretch and occasional loosening. The non-continuous belts are tensioned by pulling more of the belt through the clamps that hold it in place.

For the application, I need them that tight–they’re being driven by a trio of stepper motors, and need accuracy as tight as possible. But with them a little bit slacker, I can still get acceptable results. An FRC robot could go slacker still, but not by much.

We used a horizontal high speed belt (direct to a cim, speeding up an impeller) without problems for Breakaway. 1 pulley had flanges, this kept the belt from wandering.

1717 didn’t seem to have any problems, but then again, they have access to magic.