Broken Glowforge Rollers - A trip to the hardware store

I noticed that the original wheel on my machine is slightly worn, and it has only ever seen hobby-level use. I suspect the spring loading of the carriage puts a decent amount of pressure on these wheels.

Durability will depend a lot on the material used, but they should last awhile if the fit is good. I use PLA for almost everything, but for this I might use ABS or perhaps PETG. If I had a carbon printer, that would be the way to go.

One problem with this part is that the groove around the wheel comes within a few millimeters of the inside bearing race. I suspect this is why the OP’s wheel split the way it did. It might be possible to tweak that a little, but I would worry about alignment of the print head if the diameter changed too much.


Annealing is an excellent suggestion.

I think printing this standing vertically would definitely present challenges, mainly that it would probably need supports, which will affect the surface finish. But you’re right - the wheel is going to want to split at that center groove.

That gives me an idea, though. Instead of making the groove a V shape, I wonder if it would have less stress if the channel was the same shape as the rail?


If I was close by I would turn one on my lathe for you…


Or just buy a metal or urethane one online. It’s a track or guide roller, possibly an off the shelf one.


They are only a couple of dollars, but finding the exact one is not so easy…


Thanks to everyone for their replies and feedback. Things got a little crazy this afternoon so I didn’t have the opportunity to try to print it. I only have PLA so hopefully it will work for the short term anyway. Ill follow up once printed. I agree that having the wheel groove shaped differently would definitely help. As it is now, It’s like running a knife across plastic over and over all day long… bound to split


I have searched online but have not found anything that is the exact or
reasonably close size

1 Like

This is a standard mechanism for precision motion control, used for many years on many devices, including the Glowforge since the beginning.

The issue is with the quality of a recent batch of the wheels, not the design.


The problem with this replacement is that all the force on the printed part is going to be where the print is the weakest. 3D prints are much less strong along the layer lines. This may split fairly quickly, parting at the bottom of the groove where all the forces are going to be working to pry the two sides apart along that line. You should consider printing it with supports (even though you’ll have to clean it up some afterwards), standing on edge so the layers are at right angles to the forces that’ll be applied to them…


Also, GF can’t be using a custom part for this. I’ve got similar looking wheels (metal, though) for a large 3-axis gantry CNC machine. Do a Google Search for “Linear Rail Guide Wheel” and you will get plenty of hits for things that look similar. You just need to size the replacement to be able to order one.


Most adhesives won’t work with nylon or Delrin. They don’t have any chemical or physical availability for the adhesives to use.

I would not try to resolve this on your own. I would wait for GF to get your replacement parts. You are looking at a pretty exact part and TinkerCad is anything but.

I wouldn’t use ABS, PLA, or PETG for this part. I would use PMMA or nylon due to the strength requirements. You have to seat a bearing into that wheel and have it properly pressed into place. I don’t think you are going to get good wear and also the strength you need in those other materials.


If a hobby level printer can print ABS it can print nylon and PMMA.

If this was my machine I would not be trying to fix it because the possible error induced from the printed part would not be worth the risk to my customer orders. I would certainly not be doing anything to risk damage to the ret of the machine. I would be looking at my backup plan to fill those orders while I wait on replacement parts from GlowForge.


It’s all been said in the multiple threads they started. They are determined to prove everyone wrong.





Yes, thanks for all your help. I see at least 7 new Problems and Support posts in the last week of people with wheels that disintegrated on their machines, but I’m sure trying a 3D printed wheel will have disastrous consequences.

Nylon filament is tricky to work with, and pushes the printer quite hard. It requires relatively high extruder and bed temperatures, and an all-metal extruder, preferably direct drive. So sure, a lot of hobby level machines can handle it, and a lot will break down trying.

Feel free to offer to print up some wheels for folks, if you want to help.

Be well.

1 Like

OK, you can McGuyver a wheel on your 3D printer BUT there are some real challenges with that.

  1. The dimensions required to function well with the others need to be +/- .001" or better, especially for the bearing press fit.
  2. 3D printers do not produce circles, but a regular polygon shape that probably deviates from the circle by several thousandths. That wheel will drive really bumpy on the track, which will wobble the head and if nothing else create poor cuts & scores
  3. It is not the knife edge that actually splits the roller, but the wedging forces of the vee. The wheel makes flank contact with faces of the rail and the splitting force is significant. I agree that printed part with the bore perpendicular to the bed of the printer would be very challenged with regard to layer adhesion. If you print the part with support in the on edge orientation the out of round will be significant.

This wheel is a precision part and must be replaced with a precision part. Buy the one from Glowforge. I have the tools and equipment to machine one, but it’s not worth the time unless you are really stuck.

1 Like

I print nylon all the time on my Robox hobby level machines. I find that I really only need about 60 C bed temperatures and about 235-250 hot end temperatures, depending on the formulation. The Robox doesn’t have a direct drive extruder; it is a Bowden style.

My Prusa machines also handle nylon without an issue. It really isn’t that hard to print in my experience, and I have a lot of it.

Nylon is only tricky to work with if you don’t know what you are doing. It warps, so a heated bed is required and an enclosed bed is preferred. If not, a good adhesive surface like GeckoTek or DymaFix usually handles it without much issue, especially on something as small as a wheel like this. You also have to keep it dry, or it will fall apart under load. But with the proper precautions and a little knowledge you should be able to print nylon on any quality hobby level 3D printer.


So I printed these parts, and they are actually very strong, even in PLA, and even printed in the flat orientation shown. Printing a set of 4 wheels took about 20 minutes on my printer.

Since the original wheels on my Glowforge have not failed, I am not keen to destroy them just to get the bearings out to test this temporary replacement. I have other bearings around, but at least two dimensions are critical: the inner bore of the bearing must be 6mm to match the post, and the width of the bearing must be about 6mm to avoid vertical slop. The outer diameter of the bearing used is 10mm, but this part could be modified slightly to accommodate a very slightly different diameter. I did some searching to find an identical replacement, but haven’t found one with all dimensions exact. (Possibly this one would work, if you stacked two of them in each wheel: )

I did a quick knife-edge test. Using a very sharp kitchen knife, I put the blade into the groove, put some downward pressure, and rolled the wheel back and forth rapidly on a smooth counter. The wheel split within a second of this.

I then tested one with the other side of the blade, which is about 1mm thick, square edge. The part did not separate at all.

The wheels are very smooth. (Knowing that these are rotary contact surfaces, I had exported the original model with a relatively high angular resolution.) They roll like marbles.

If I had a way to mount these on my Glowforge, I would already be cutting with these to see how they hold up in real life. But since I have taken this as far as I can for now, I will offer this:

Anyone who needs a set of replacement wheels, and who is not able to print their own, send me a PM and I will print a set for you to try. Be warned that these are not tested, and may fail again in short order. I cannot speak to whether trying these might affect your warranty, but I know the company is trying to resolve this issue for us, and hopefully you will not need to use these for very long.

You will need to extract the bearings from your original (presumably damaged) wheels, and press-fit them into these parts. I haven’t been able to test this, but I expect the fit will be very tight, so you would need to press them carefully and evenly, and might need to run a 10mm drill through the opening to enlarge it slightly, or perhaps heat the part with a hair dryer/heat gun.

All I ask in return is that you share your results back here so I can make any necessary improvements to the design.


I’m actually interested to see how well these work. They might be a good stop-gap measure for those who are too desperate to wait for the Glowforge replacement.

Necessary word of warning…

And I hate to have to bring this up, but you all need to be aware of it. This quite likely does qualify as a modification of the machine and as such, if you have a problem with the machine while using it, the repair costs will not be covered under warranty. So if the wheel breaks and the laser arm gets stuck and starts a fire in your machine, it’s going to cost you thousands of dollars to replace. It will not be covered.

So make your own call on using them, but be aware of the danger and never take your eyes off of it while it’s working if you try this. You might be able to stop the print and save your machine if you see it happen. You might burn the machine up if you don’t. (And there are plenty of pictures in the forum on laser fires. It doesn’t take long.)

Good luck, and let us know the results.