Rack & Pinion Brainstorming

Taking this into a different discussion as per request. Let me start by saying that I am not any kind of mechanical engineer. I studied computer science, so there’s plenty I don’t know about building things. Just trying to have a dialogue, and hopefully learn stuff.

True, but don’t belts stretch as well? There are always going to be sources of inaccuracy in a CNC. There’s no such thing as a truly rigid body, nothing is perfect. However, I know there are R&P based machines out there that do amazing stuff. One example is Joe’s Hybrid CNC:


GT belts are fiber-reinforced (or something similar) so the the stretch is minimal and not an issue.

Belts do stretch under tension, like any other material in our world. But like Antoine said, they are fiber reinforced so they’re actually incredibly stiff. Plus, GT style belt is known to be among one of the best choices for applications where registry matters (cnc).

This came up on a different forum I go on. Belts stretch a tiny tiny bit under tension, but don’t stretch overtime like some people believe them to do.


The cnc tables at Esko are rack and pinion. Though for the most part, they’re used for sign making and corrugated(don’t call it cardboard!) cutting industries, so not much flying dust. I believe the cnc router parts tables are also rack and pinion. Belts are cheaper, and easier to install.

Rack and pinion are like leadscrews and ball screws, cheap ones are only good for specific uses, good ones are awesome but should be paid for with a business account. The gears need to be spot on, and they wear overtime so they need maintenance, lube, just like Balls screws. Bang for the buck, you can’t get close to toothed belt and physically it is the most efficient transfer of power (not sloshing around lube). Remember we are using reinforced toothed belts, essentially gear train that is misalignment tolerant that needs no lube. A vbelt is a whole different story.

But… I think a 3D printed rack and pinion is totally viable. I doubt there would be any loss of accuracy if it is designed right and for the price…super cool. As long as it isn’t mounted on the top surface (or covered well) I think it would be freaking cool to try.


Like this?

[attachment file=62618]

I can’t argue the technicalities of printed r&p but can attest to the FUN of it… and it’s suitability for some purposes…

– David



Heck yes!

Why not doing both at the same time?

Print a flat support for a belt, tension the belt and then glue it to the support. You could then use it like a rack and the pinion would be a normal pulley.

I wouldn’t recommend to use the tiny belts we are using for the mpcnc (they might work though), but with GT2 style belts it should work pretty well.

I think it is a very cheap and easy solution to make a relatively accurate and silent rack/pinion system, while limiting the backlash. Plus it could be made up to any lenghts without any gap, very easily.

1 Like

I was thinking the same thing.


May be a solution no print but twice more belt higher strength ans precision everman patent

But not the initial project rack and pinion


I tried something like that on my LowRider. It was effective, but unless the lower belt was held in place, glue or something, I think just using a larger belt is a easier option. If you did glue in the lower belt that would have to be freaking amazingly accurate, pretty easy to tension as well.