Cannot Get all 4 Rollers to Engage all 5 Bearings

I am part of a small group building our first MPCNC, and while we thought things were going well, we’ve run into an issue where not all of the 4 Roller parts, have all their 5 bearings engaged which provides slop between the left and right rollers on an axis.

I’ve tightened up all combinations of the 5 bolts and nuts (actually all 7 when the motor piece is on it). Now we have cracked (3) some of the Rollers starting on the side with the two lower bearing positions, usually closer to the middle but sometimes a bit more to the left or right. The cracks must have been caused by some stress in the triangle of bearings, but even when glued in a clamp, there is a gap on one or more bearings on the conduit. I’m really at a loss as to what is going on.

I thought the gantry X&Y beams might be pulling the rollers and causing the issue, but that doesn’t make sense, when all 5 bearings are engaged nothing is going to knock it off the rails.

Lastly, I was able to make a part with it yesterday, which came out pretty good, but then I had seen the cracks in the rollers and noticed some jerkiness on the X axis so took out the 3 cracked rollers, glued them with CA clamped over night, tested this this morning with a short amount of conduit, they were all fine, all bearings engaged, but then when putting them back on the machine, I have rollers with loose bearings.

Lastly, today it started binding badly, steps were missed (on the Y axis), which is a new symptom, I moved the rails by hand and experienced the binding, but could not find the cause of it. Took the belts off the Y axis… rolls very freely and smooth, no binding, still not 100% engagement of the bearings… but it felt very good.

Specs of our machine;

3’ x 3’

6.5" spindle collet to table Z Axis

3/4Inch Canadian Conduit

X and Y Dual steppers are wired in series

Any thing else I’ve missed to provide please ask, looking for pointers here… before I start taking the whole thing apart.

Oh… and I’m printing a PETG roller right now… just to see…




You have too much going on here. Sort your problems in order of the instructions.

Rollers not touching- Take everything off of them and if they all do not touch there is a huge gigantic issue with your prints.

The center has nothing to do with the rollers. You cracked parts meaning you went at least 3-4X too tight. Try a torque wrench to calibrate your buff arms.

A 6.5" Z axis is kinda ridiculous if you plan on using a spindle.

Canadian conduit, what are the OD measurements?

Pictures are worth a thousand words for the rest of the issues.


Thanks for the reply.

Canadian Conduit = American Conduit = 23.35mm to 23.48mm OD

6.5" Z, yes that’s big for this type of machine, but it’s for doing things like;

Which is what I ran on it yesterday before I pulled the roller pieces.

Thank you for confirming the centre has nothing to do with the rollers, agree I did tighten the nuts and bolts too much, I now have a torque wrench on site with the machine for future work.

I’m beginning to think I got a dodgy piece of conduit… I bought 3 x 10’ lengths at different times. The small 6" piece I have here at home with me, I tested on the rollers before I went on the site to install them, all the rollers were doing their thing properly. I’ll measure all the conduit and make sure it’s all the right diameter.

Hoping I don’t have to strip it all down, the wiring with the cable chains was not an enjoyable task.

Thanks again!


All that machine height adding to the instability to add in a vise is better achieved by making the table itself lower. 3" should be more than enough for any length of 1/8" endmill you can get plus full retraction and a little extra room. Not that this is causing any of your issues. Even with glued parts and a wobbly bearing that carve looks pretty good. Maybe just run with it as is for a while?


Pictures of your actual build is what I was hoping for to look for obvious issues.

Hi Ryan;

This is our first build, as the dimension is completely based on the length of the conduits we decided to try for a 6" Z 24" X 24" Y build, if it didn’t work out, make it smaller, not that hard. I have seen examples of people lowering the table, might go down that road. Our group really wanted something very flexible, to do milling, lasering, cutting, extruding, etc.

Here’s a picture of one of the rollers with the problem, the bearing isn’t touching the conduit… might be hard to see. I replaced this roller with one I printed last night in petg, it has all 5 bearings fully engaged and is doing what I expected. The other roller on the other side is not perfect, much better than the first, i’m currently printing a roller-m to replace it.

I also noticed that the belt tension has a very very small window of working, goes from too loose to too tight very quickly, did some playing with that today and it also helped smooth things out.

I think I’m on the right track. Here’s a picture or two of our build;

I had to go to the makerspace to take the photos today, sorry about the delay.




This is not a problem I have ever heard of on a roller before. Those parts are actually an interference fit. With no tension on the tension bolts for most they actually seem too tight. One of the most common questions is “is this supposed to roll with gravity (or something similar)?” So something seem pretty off with those prints.


Again, another odd thing. You should have a very large window of tension. This could just be a matter of our perspectives on tension but it should actually work fine all the way up until you actually stall the stepper (which is probably 5x too tight).

I can’t zoom on those pictures or see what board you are using but maybe you just did not set the drivers correctly, but at the same time if you went tight enough to crack the parts then the steppers are probably not strong enough to overcome that. The machine should be pretty good with zero tension on any of the tension bolts.