Belt upgrade worth it?

Hi all. I make these pistol grips with a small-size Shapeoko 3 I bought in March 2017. It’s got the 9mm belts, but not steel core. Through tons of iterations, I can now get the grips to come out near perfect with very tight tolerances. However, the challenge I run across most often (especially with really hard woods) is the bushing holes on the back can slightly mis-align with the holes milled from the front. Luckily it doesn’t affect the function, is hidden when on gun, and so slight it’s barely noticeable if you’re looking for it. I assume it’s due to the stretch in the belts being magnified when milling same-center holes from opposite sides with opposite forces acting on the spindle (opposite from the part’s perspective once flipped).

Would a steel core belt upgrade help this significantly??? Are the belts the main culprit??? Thanks…


4 Likes

What method are you using to align and zero the parts when flipping them over?

2 Likes

Steel dowels on a jig. I have a top stage on the jig to drill the alignment dowel holes, then a bottom stage on the jig with steel dowels sticking out that stick them on after flipping. Pretty sure it’s getting perfectly aligned to my CAD on the flip.

Have the belts ever been replaced or recalibrated? I’m a big fan of the steel belts, but if you’re getting one side that fits dimensions, there’s no reason the second shouldn’t. Is the machine powered off between the two sides? If so, could be some error coming in when homing (a loose switch?).

2 Likes

Also, are the holes actually round? If there’s belt stretch it’s likely to show up in off round holes when cut.

Yes indeed, I think they’re off round a tiny bit when I do notice a problem. Think steel belts would make a noticeable improvement?

mikep - No I’ve never replaced the belts, but I’ve calibrated for stretch pretty accurately. No the machine isn’t powered off in between.

I don’t think it’s a problem that one side fits dimension and one doesn’t, as much as center on center holes really make any slop in the belts noticeable.

You’re a big fan of steel belts? Is it because you found they took out a lot of slop? I just wanted to ask everyone before I coughed up the $70 for the kit and took the time to change them. Also wondered if the HDZ is really what I need to improve this?

I have the HDZ and the kevlar belts from reprap.nl

My experience with the kevlar lines up with the excellent work in the belt stretch thread in that they seemed to roughly halve the positional error and backlash I was seeing with the original (non steel) belts which were about 9 months old.

OTOH, using the bitzero I am still consistently 0.1 to 0.2mm off from the side of the stock when setting zero, I have to put ± 0.1mm of backlash compensation in the cuts and this varies from the side to the middle of the machine so I am really not in any position to tell anyone how to kill their backlash or positional inaccuraccy issues :cry:

Over to the more experienced folk

4 Likes

Thanks. Looks like I should try steel. I searched and see that mikep has tried both steel and kevlar and says he’s a fan of steel now. I hadn’t discovered https://reseller.reprap.world/ yet! Thanks! Looks like I don’t have to get the carbide repair kit after all. I can just order belts off reprap. I assume this is the same as what’s in the Carbide 3D repair kit?

GT2x9 mm Steel-reinforced timing belt (per meter)
https://reseller.reprap.world/products/mechanical/timing_belts/gt2x9_mm_steel_reinforced_timing_belt_per_meter

It will make a huge difference over fiberglass core. As in night and day noticeable. I have been running 9mm steel core belts for a good while and love them, but I just switched to 10mm wide and so far, so good.

That reminds me, in a few weeks I need to ask @Julien to reopen that thread on stepper holding/belt stretch measurements. I am about to mail some gt3 and 10mm steel core to @The_real_janderson for some measurements to add to the mix.

6 Likes

Geeze Scott, don’t know how I missed this thread with all my searching - “Belt Stretch and Stepper Holding Measured”

This was just what I was looking for in trying to figure out if steel was worth it.

Thanks!

1 Like

Ah,

Have you tried out the GT3 belts then?

I just ordered one to try out on my X axis.

I haven’t yet. The 10mm wide belts I ordered over 2 months ago finally showed up just as my x-axis belt was giving out and I decided to throw those on as I need my machine to be cutting reliably right now and I dont want to deal with recalibrating for belt stretch, as properly tensioned steel core belts have been pretty much “plug and play” for me with no calibrations needed from belt to belt. Granted, my work requires lower levels of precision compared to most people on here. Upon my caveman-level inspection of them when I opened the box, I did notice they seem a tiny bit thicker than my gt2 fiberglass core belts. Otherwise they are indistinguishable from the gt2s.

4 Likes

Thanks,

I’ll try to do some stretch and comparison measurements on my GT3 when it arrives and post the results, we’ll see how they work out.

3 Likes

I would also diagnose this on something similar and less complex. I could be wrong, but this thread struck me as jumping right to belts without ever looking back. It sounds like you’re doing a through hole bore from the outside, followed by a partial depth counterbore from the inside?

Flipping a part and maintaining concentricity strikes me as way more likely of a culprit than belt stretch. If I understand the question correctly, can you do this on a test piece, doing both boring operations from the same side? If they were perfectly aligned, that is a quick test to see how much belt stretch is really affecting you. If they are concentric and aligned, stretch is not an issue.

As another test, if the thought is that it’s cutting stresses, you could machine the through hole under-sized, then come back with a really light clean up pass? If the hole was made imperfect the first time due to hogging out all that material and being deflected due to cutting forces, hopefully the touch up pass could fix any inaccuracies. The counterbore already looks pretty light to me, but you could do this in two passes as well.

Also, for the holes to be offset to one side, wouldn’t belt stretch have to play out as machine deflection in a particular direction? As in, the wood deflects the tool only in X but not y, or the Shapeoko requires much more force to move along one axis than the other? Otherwise, I would expect deflection to act uniformly and your ID would be too big/small (but aligned) vs. offset like that.

Anyway, main point is that this is worth at least a couple ~30min tests to verify that belts will solve your problem!

3 Likes

This topic was automatically closed after 30 days. New replies are no longer allowed.