Cutting extruded ali

Just to let any body building a printer know,that you can cut it with a hack saw,it does not need to be done by machine.if you take your time to cut it as square as possible that will be fine,when building it just make sure that it is square. Ok it might not look as good, but for people that don't have the funds it is fine

Another low cost method (albeit not as cheap as a hacksaw) is to use a non-ferrous fine-tooth blade on a chop saw. I'm going to give this one a try:
Freud 84T Non Ferrous

Make sure you clamp the material, and use safety glasses or face shield.

Anyone have a better solution for cutting Al extrusion? I believe a circular saw would give a better finish than a band-saw. How about a dry-saw vs. reg chop saw (i.e. 1300rpm vs 4000-5000rpm)?

Almost any saw that can cut metal (avoid abrasive saws) will cut aluminum. It isn't difficult. What is difficult is getting square cuts. If the cuts aren't square, you'll have to use metal plates or printed plastic parts to hold the extrusion pieces together, usually giving up a lot of the rigidity of the material.

A lot of the places that sell the extrusions sell it with square cuts at specified length for a little more $. It's well worth it to pay the extra for square cutting if you don't have tools available to square up your cuts because you can then screw the pieces directly to each other and end up with a square frame without giving up the rigidity you wanted when you decided to spend the money on t-slot extrusion.

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

Even if you do have the tools to square up cuts, it can still be good to order pre-cut because you also want multiples of precise, and usually different but related, lengths. For example, if you are making a frame from 2020, you might have some elements that are x mm long, and others that are x-20 or x-40 mm long. If you have the ability to do precision measurements of around x mm, great, but that's beyond typical home or makerspace shop. (For example, on the ubiquitous harbor freight mini mill, you have a max of 9 inches X travel, minus edge finder radius, minus tool radius, for milling a precision length.) In which case, you would want to face square against a fence for the longest pieces, and then stack segments of the extrusion against the fence as needed to subtract from the base length rather than moving the fence, because the fence is a non-precision measurement. On a mill with a table substantially shorter than x, this could be quite tricky to set up.

I have cut aluminum extrusion (though not for a printer) with a decent miter saw (I used a carbide tooth blade) but the cut angle is set only to within tolerances considered acceptable for woodworking. You might get lucky, but it's also totally possible that your cut will be nice and flat with a beautiful surface finish—and 0.5⁰ off of square, which looks great until you assemble the frame. I agree with DD that if you are going to use the square ends as references (and it's easier and more stable to do so), ordering cut-to-length extrusions is worth it, at least for pieces that butt up square against another piece.

If you've decided to go the non-precision route with a hacksaw, one way to improve squareness of your cut is to clamp a sacrificial fence to both sides of your material and saw against the fence. I have used this technique (though not for building printer frames) and I found it quite frustrating because not only was it slower, it still wasn't as square as I wanted... ☺

I read about people, who even milled the faces of their T-slot extrusion to have a straight base for mounting linear rails.
If there's any unevenness on the faces, why bother milling the ends strait but leave the faces untouched? ( like many people do )

There's a lot of different t-slot out there. Some has flat faces, some concave. The concave faces are intended to act as a spring so that when you bolt stuff to the the t-slot, the aluminum will distort and act like a lock washer by keeping tension on your bolts. If you have the concave-faced stuff and want to mount linear guides, you either mill a flat surface into the face of the aluminum or you bolt on a piece of tooling plate. I have done both.

There are techniques for milling long pieces to accurate/matching lengths even on a mill with a short table. You can clamp the work pieces together, etc.

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

Quote
mcdanlj
You might get lucky, but it's also totally possible that your cut will be nice and flat with a beautiful surface finish—and 0.5⁰ off of square, which looks great until you assemble the frame. I agree with DD that if you are going to use the square ends as references (and it's easier and more stable to do so), ordering cut-to-length extrusions is worth it, at least for pieces that butt up square against another piece.

I was thinking that was the likely outcome. Still, i snagged a vintage NOS Makita LS1020 off eBay, which only has single miter function (not compound), so there is only one angle to dial in. Once set, I bet it is fairly repeatable.

I've ordered Misumi in the past, and while the cuts are square, they do vary from piece to piece (+/-.5 mm). If you need two short rails to be very parallel it's probably on the limits of OK. Their higher precision option of +/-.2 wasn't when i ordered, mine came out at more than .5 different. I'd stick with the regular option and hope for the best.

Quote
mcdanlj
If you've decided to go the non-precision route with a hacksaw, one way to improve squareness of your cut is to clamp a sacrificial fence to both sides of your material and saw against the fence. I have used this technique (though not for building printer frames) and I found it quite frustrating because not only was it slower, it still wasn't as square as I wanted... ☺

I just did an experiment. I took a 10cm-wide piece of of 18mm (3/4") aluminum stock. I used digital caliper depth gauge to clamp fence pieces against both sides at the same depth, and then used a dry-lubricated hacksaw to cut as squarely as I could. I put the cut piece in the mill and set my zero depth to have the endmill just barely touching the highest point. A .02" pass barely cut a few high spots. Another .02" pass cut it almost fully smooth, with just a few nicks. A final .01" pass removed all the nicks. therefore, the total variation in my cut was 0.05" or about 1.25 mm with almost all of the variation in 0.02" or about 0.5 mm. There were not obviously local substantial minima in variance, either. You might do better than me, but it still won't give a square printer without some aid like external corner plates.

In retrospect, I should have done .01" passes for better data, but my subjective impression was that most of the variance was a bit less than ~ .01" / ~ 0.25 mm, just from looking at the line where the mill engaged with the aluminum it was cutting. I'm not motivated to do another experiment, though. Hacksawing is hard work! Time to mill that into a new extruder carriage mount!

Quote
gmedlicott
I was thinking that was the likely outcome. Still, i snagged a vintage NOS Makita LS1020 off eBay, which only has single miter function (not compound), so there is only one angle to dial in. Once set, I bet it is fairly repeatable.

If you can "set and forget it" to be square, that might be a good option. You may be able to use a rubber mallet or dead blow hammer to gently nudge it to be accurate. If I recall correctly, I spent about an hour and a half tramming the head on a mini-mill, which would be a fairly similar operation. I'd be curious to hear how well this project works!

Quote
gmedlicott
I've ordered Misumi in the past, and while the cuts are square, they do vary from piece to piece (+/-.5 mm). If you need two short rails to be very parallel it's probably on the limits of OK. Their higher precision option of +/-.2 wasn't when i ordered, mine came out at more than .5 different. I'd stick with the regular option and hope for the best.

That's not as precise as I assumed. I had never gotten as far as looking into precision options there.

I kept not building my own CoreXY from scratch, and in January I bought a Tronxy X5S to bootstrap the process, expecting it to be a learning about lots of things not to do. It was. Given that my rebuild is reusing the frame, steppers, and most of the Z axis (but replacing the leadscrews) I was feeling that perhaps that was an expensive source of pre-cut-and-drilled extrusion. I'm guessing they must have cut and drilled the extrusion with jigs, then, because for all its other faults, the pieces were cut sufficiently precisely to length that I couldn't detect any deviation from square with my machinist's square, on any of the angles in the frame. I added internal corner braces everywhere I could to add tension to make it more rigid, but they weren't to fix deviation from square, just to preserve it if the machine is bumped/transported. The extrusion is flat, too; when I placed a piece across a reference plate I couldn't see light under it, and screwing linear rail down to it flattened out the linear rail and made it run smoother. So they did something right there!

Edited 1 time(s). Last edit at 09/03/2018 01:04PM by mcdanlj.