PVC Rollers?

Do you have any pictures of this belt comb? I think I know what you’re describing, but I’d like to see it to be sure.

The few times we’ve used beating, we always had it ride between larger pvc pipe sections that were covered in wheel tread to overcome the slippery nature of the plastic.

What is a belt comb? I can’t find it on McMaster-Carr. Is there a different name they are called?

I looked for a good belt comb pic.

Here’s a cropped picture of a belt comb on our practice robot. It’s thin polycarbonate rolled around the roller and riveted, we neglected to take the film off of the polycarb so it’s white.

Our belt combs inside the robot (where both rollers need to contact the ball) usually cover only the top (non-contacting) side of the belt and don’t go around the roller.

These belts were under a lot of side load, so we needed a lot of combing to keep the belts on track. We actually didn’t rivet the spacers on the other side so the belts can walk as they please under load and move back into shape when the ball has passed.

Basically what Andrew said… The belt comb in his picture is pretty unconventional as it goes all the way around the roller.

More commonly one would use something like my attachment…

We usually use thin polycarb for ours but I have seen them made out of lots of different things.

The reason that we call them belt combs is because the applied piece we make looks like a comb before it is bolted on.

Regards, Bryan

Edit, In the second above picture by sportzkrazzy towards the top there is a clear piece that doesn’t go all the way around the belt but still looks like its serving the same function.

Belt Comb.png

Belt Comb.png

This past year FRC2168 used some aluminum tube with threaded inserts and pop riveted tread to make our rollers. You can read more about it here: Lift & Hopper

Part numbers included: McMaster 60945K16, 1968T576, 5994K711

Here is the best picture that I have of the belt comb that my team used. It is just a piece of plastic with notches that the belt runs in.
(sorry, unzipped is too large of a file)

P1010283.zip (4.69 MB)

P1010283.zip (4.69 MB)

We never really had a name for it but yeah that’s pretty much the function that is served. We just used a piece of Lexan that we bent in order for it to keep its shape and maintain strength rather than going all the way around the cord. Two rivets a small piece of lexan and 30 sec on the brake and your done. Simple and lightweight whats not to love.

We used simple PVC tubing from the store, and added a couple of “combs” to keep the belts straight. The combs were just pieces of polycarb with notches cut in them for the tubing. We placed them relatively close to the rollers, and didn’t have a single issue with our belts all year (2 regionals, 2 off season competitions, a competition at the State Fair, and a few demo’s that saw the robot running almost continuously for an entire day).

We had a similar conveyor system for Lunacy, but back then we used some solid ABS rollers (really heavy) and machined some grooves in them. I don’t know what we did wrong with the grooves, but we had to go in after every match and reset the belts, as they kept jumping out of the grooves and wandering. Making the combs this year proved both easier and more reliable.

If you’re just looking to drive the rollers rather than using the belts to react against a game piece, I’d recommend simply using lengths of PVC and round Nylon or Acetal Resin (DuPont’s is Delrin - it’s what most of the belted pulleys I designed were made of) to “plug” the ends of the rollers, into which you can machine grooves (assuming you can turn the material on a lathe). This gives you a solid piece of stock into which you can sink deep grooves to help minimize belt walk. As an added bonus, your belts can then be flush (or even recessed) in relation to the surface of the roller.

Of course, this is a very purpose-specific application. It doesn’t work quite as well if your belts are meant for moving a game piece rather than simply driving a roller. The advantage is that it allows for a very light, strong roller with a uniform surface while still providing a solid mounting point and great freedom in the depth of grooves. It also means that you can machine the plugs to either accommodate or fit through a bushing or bearing.

I would recommend checking out our website. We have CAD drawings up for rollers that we made to suit pneumatic tubing bands. If you have any trouble please ask, we are happy to share.


The links are near the bottom of the page.

Good luck!!!

We’ve used shelf-liner to both fill that gap and help increase friction between the polycord and the roller.

http://www.chiefdelphi.com/media/photos/32753 (maroon in that photo, the outer PVC portions are wrapped in black tape)

I wonder whether people have had different results with different guiding sytems when using polycord vs. urethane belting. Can anyone who has used both present an argument for one or the other, or list pros and cons. I know that one thing that the urethane belting has as an advantage is that it has a smaller pulley minimum diameter. Besides this though, what is the difference?

How do we make grooves to retain our polycord?

We don’t. We use flat belting.

We use McMaster-Carr p/n: 6075K19 flat belting on a simple PVC pipe. To control the belt tracking we put 3-4 wraps of hockey tape on the center of where we want the belt to be. If the belt gets pushed, it tracks back on to the tape auto-magically. The belting is a bit expensive, at $4/foot for the 2" wide stuff we use, but it requires fewer lineal feet than round belting needs to cover the same span. Additionally the flat belt typically has a smaller minimum roller diameter than round belt which can help packaging. Because we do not need to cut grooves in the roller, the wall thickness can be thinner thus saving weight and cost in the rollers. The big advantage is time saved by simply not needing to make parts.

Here is an example from our 2012 conveyor.

We’ve used pvc rollers for several years now. You can buy plastic tubing from mcmaster to cut and put over the pvc to act as guides that are pretty much a perfect fit. Superglue in place.

The biggest difference I noticed this year for teams that used round belting was tension. If you put a LOT of tension on the round pvc belts, they don’t deflect nearly as much and will track very straight. Of course, you’ll want bearings and not bushings on your rollers to minimize friction if you’re going this route.

Our shafts are all mounted in a slot and terminate in small block of aluminum, and we use a bolt to push on the block of aluminum to tension them. (Call it a jack screw if you’d like). That allows us to account for any variability in our manufacturing, because we don’t have CNC ability and everything is fabbed by hand, so nothing ever comes out perfectly straight.

How did you customize the belts to the correct length? Did you weld them? Also why does it require fewer lineal feet than round belting needs to cover the same span.

We buy the belt in bulk so we mount the rollers and wrap the large piece of belting in place and mark it with a sharpie. We pull it snug, but do not attempt to tension the belt at this point. Then we pull the belt off and measure the length we marked out above. We subtract 1% off the length for tension then cut the belt at the new mark and weld it up. I think only once have we had to increase the belt tension, but I think we only had 2% cut off even then. We try to keep the tensions down to minimize frictional losses in the system.

In the past we have welded these belts with just a heat gun and some blocks of wood which is is not a great experience. This year we sprung for the heating tool, McMaster-Carr p/n: 62065K11 which greatly simplifies the welding process.

The span I am referring to is the width of the conveyor. Poor word choice on my end. As can be seen in the photo above we only used two flat belts to manage the 8" ball width. Using 1/4" round belt would have taken 2-3x as many belt runs to keep the ball from pushing past the belts. This cuts the cost premium for flat belts at least in half.

I believe I understand what you are saying, but anyway we can see a picture of this set up? Just curious to see the relationship you have between your mounting blocks and your frame members in order to faciltate this “jack screw.”

Absolutely. Here is our 2009 setup. The bolt is in front of the wheel. The small block that the bolt runs through is welded to the frame was tapped. It is backed all the way out in the picture so it is not pushing the shaft end.

CD won’t let me embed a photo from flickr, so here is the link:

Here, you can see it on the left side and you can also see a bit of the slot that the shaft rides in:

We improved upon that this year. Instead of welding to the frame, we used 1/8" thick aluminum angle and rivited to the frame, then used a couple jam nuts on the bolt. That got rid of the welding and the tapping. We also put a small aluminum square on the shaft and drilled a hole for the bolt end to rest in, so it didn’t try to walk off the shaft under high load.

We used regular PVC pipe cut to length. We machined end caps that fit to the inside diameter of the pipe. We hex broached the end caps and ran a 1/2" hexed aluminum rod through the pipe with the end cap screwed into the pipe. We used .125" thick rectangular aluminum tube to create a frame. Then we ran timing belt inside the tubing to spin the rollers. We wanted to run the intake real fast so we used timing belt instead of poly cord. We can run our rollers at 20’ per second to load the balls into the tower. Timing belt is superior in strength.

We had one of the best ball intakes last season. Poly cord is kaput to us.


Did you run center to center on the timing belt pulley distances or did you build in a tensioner?

In the picture it looks like the PVC was uncovered. That provided enough “grab” to move the balls with any covering?

  • T