pic: Conveyer belt tracking

Having issues with belts not tracking quite right and running off of rollers. Conveyer is driven via chain/motor from back roller and that drives whole conveyer. Any ideas how we may build this better so as to have the belts track truer? Belts are rubberized and we have tried just putting orings to separate each belt but they still run walk. Thanks!

Two years ago we used rubber belts and just placed a sticky piece of rubber about a cm wide in the center of where we wanted the belts to ride. They would want to center on this piece of rubber.

In this picture you can see the place rubber on the rollers where we have belts prior to removing them.

If this is flat belt, build up the area under where you want the belt to stay with tape.

Flat belts will track to stay on top of a RAISED area of your roller.

It sounds very counter-intuitive, but try it and it will keep your belts in place.

Thanks for the ideas folks!!

We have done this in the past with great success. Electrical tape works well for this purpose.

I concur, look up information on “crown roller” for the “why” in how it works. For your axles you’ll have multiple crowns.

Any chance the ‘crown’ method works for timing belts? We’re adjusting our approach to one of our designs, and pulley alignment may not be perfect afterwards.

Crowned Pulleys, make them your friend. My first robot I made I discovered this phenomena by accident, it was a very happy accident.

Flanged pulleys, for us, have always been adequate in cases of slight mis-alignment.

The physics works for flat belts on a smooth shaft. Teeth are going to complicate things a little.

Timing belts usually run in sheaves (V-belt grooves) rather than on crowns. The flanges around the sheaves maintain alignment. Note that sometimes only one pulley will have these flanges, and the others are wide enough to work where the guided one leads. The drive belts on this year’s KOP drive are a good example.

Here’s a better example showing one flanged pulley and one unflanged pulley.