I’ve called out 6 squares to be cut 0.680 in. in both X&Y axes. The resulting holes are 0.634 to 0.644 in both X & Y Axes. What would be the cause of this?
Thank you
update: Since I’ve asked this question I’ve gone ahead and checked the calibration of both X & Y and they are spot on.
Deflection could cause this, usually caused by loose vwheels, or belt stretch.
Your bit might not be the diameter it is advertised as.
Backlash could also be a problem. To check during calibration go both directions + and then -
Both above can be in combination with deflection as Seth stated.
Well, I’m assuming its an xcarve running belts since he used the category of X-Carve… but i guess if he modded it to use lead screws, then yeah backlash could be a cause as well…
Hi Seth, thanks for the input. Along with checking the calibration on the X & Y axes I’ve also gone through the adjustments on the v-wheels and the belts. This has made it closer but it is still off a little. I suspect it is the bit diameter.
Sorry, yes it is a new X-Carve that I’ve just set up in the last month. No mods, yet! lol
Thanks Mark, I believe you are right, that is my suspicion as well. Thanks for your reply.
My machine out of the box did this as well. About the same amount of deviation even with the lightest cuts. I tested the machine accuracy it by doing test traces with a fineliner on paper (so essentially a plotter) in place of the spindle, so essentially no drag. Dimensions were perfect with a pen. Here’s a thread on that: Everything i cut is undersized - #13 by TravisBrown1
So I threw out the garbage Z-axis that came with the machine and built a new one. Stiffened up the X and Y. Changed to 9mm belts. Problem went away. So basically, make the X-Carve what it should have been out of the box.
Changing the Z-axis slide will be the biggest step to getting you more accurate. You can side-step other deficiencies with lighter passes.
Also, make sure you have sufficient and even tension in your belts. I go for 5lbs @ 1.5 inches. And calibrate your steps after.
Has anyone ever held in their hand a bit that’s .03 out? That’s ~.75 of a millimetre.
I can understand accidentally using the wrong bit, like accidentally grabbing a 3mm cutter instead of .125" if one was prone to mixing their metric and SAE tooling, but have yet to find even the cheapest bit that’s out of specified diameter by any meaningful amount that would be discernible with bench top milling.
It’s that sort of thing you check first, just because; but I’m interested if anyone’s had a cutter that’s out by more than the accuracy of common digital callipers.
Reminds me of that scene in House, MD where one of his team during differential diagnosis asks, “Is it lupus?” and House replies, “It’s never lupus!”
Belt slack will effectively give you backlash.
I’d have to disagree, this is now a discussion of terminology …
Backlash refers to meshing of gears and the movement variance caused by slop in that meshing… granted you could have backlash on a belt system I guess… but only if your pulley pitch did not match your belt pitch…
Instead the term used on a belt drive system is actually “belt stretch” because the issue is not caused by gear meshing. But rather the belt is stretched out and may require replacement or retensioning.
This is why I wrote “effectively give you” rather than “is.” Belt stretch can in effect be the same as backlash.
But note you are referring to belts gradually slackening, which is a different issue. Kind of like guitar strings, belts will initially stretch under the applied load until they settle into their happy place. Then additional stretching tends to be minimal. This is a pretensioning period. This is how guitar strings stay in tune once broken in, and you aren’t continually picking up the timing belt from under your car.
But by “effectively backlash,” I’m meaning that a belt under normal tension has a degree of stretch proportionate to its tension and elasticity, and will react with such in a backlashy way to changing loads. You can reduce the degree by increasing the preload (e.g. increasing tension) and/or reducing elasticity (e.g. by replacing with a wider belt, or one of different material, that requires greater force to stretch to same degree.)
A belt slightly stretching and recoiling with any change of direction or velocity of that which it is driving is essentially system backlash by a different name. So yes, terminology.
True backlash in larger toothed belt-drive systems is a real thing, but tends to be smoothed out by the elasticity inherent in the system. I do agree that this “true” gear backlash of Gates belts is negligible in a system this small.
Fair enough
Thanks for the reply Travis. I’d like to think I wouldn’t have to replace the Z-axis since the machine is brand new within the last month and came will the upgrades to all the axes.
Additionally I’ve found the bit to be different than the size called out by Inventables in the Easel program. The bit I was using, 1/8" double flute straight end mill is actually 0.1196 as apposed to .125.
Hmm. I don’t know that a deviation of .0054 in bit diameter is going to cause .036-.046 difference in your cut. Even if you double the variance for opposing sides of your undersized squares, that’s only a total of 0.0108. That leaves .025 of slop coming from somewhere else. I’ve never seen someone complain of oversize cuts, and when I was having the issue I got the same amount of undercut regardless of the cut dimensions, so didn’t appear to be any sort of compounding error. I think the balance of your issue which leads me to believe the rest is a rigidity issue.
But change the cutter size in Easel to actual diameter and see if it corrects sufficiently?
(p.s. The new XC Z slides look much better than the original design. So if you have that, forget what I said about Z-slide.)
This is what I did that corrected the issue for my machine.
Since this was a relatively new build of a post November '21 machine:
