I am making an assembly out of 6mm thick cast acrylic (clear). I have one piece that has tabs that insert into another with holes (perpendicular parts). I carefully calibrated for the kerf and used the measured kerf in Fusion 360 to generate my design files using the colorific plugin. In order to increase clearance and allow parts to slip together, I intentionally reduced the kerf width inside Fusion 360 by 100um.
With some parts this worked great, but with my 140mm diameter disk (with holes to insert tabs) the hole size varies from place to place. If this were a clock there would be a tab hole at noon, 3, 6, and 9. Each slot size is different. I get a loose sloppy fit at noon, but can’t fit at all in 6 without forcing it. At 3 and 9 the fit is too tight, but a bit looser than at 6.
The obvious concern would be that the part was not flat, but I think it was. I’ve double checked the remaining acrylic and it is perfectly flat. There was nothing in the crumb tray tilting the part and there was nothing under the crumb tray.
I cut the part again using 6mm thick black cast acrylic. Same exact result. I double checked that the crumb tray was flat and there were no scraps under the acrylic.
I measured the thickness of the acrylic from one side to the other. It varies by about 150um (out of 6mm) on the black acrylic and 80um on the clear acrylic.
If you’re saying the cut closer to the front is consistently smaller than ones at the rear, there’s no mechanical reason for that. The beam path is fixed as the optics are all mounted on the gantry, so dirty or not, the result would be identical.
The issue mentioned above happens as the head moves away from the fixed beam guide on the left side (under the top panel), although it’s usually an alignment issue that causes the beam to not be centered on the window on the head as the head moves further away. Dirty optics might have a small impact but that wouldn’t apply in your situation (if I understand it correctly.)
I just cleaned all the optics. The problem might be just a bit better, but still a significant issue. I cut the last part from a different area of the bed to see if that helped.
I checked level of the top of the crumb tray and the top of the cutting head. Both are as level as I can measure with my phone. My crumb tray is fitting perfectly into its grooves.
As @eflyguy suggested, it might be misalignment. Is there a way to test that, or correct it?
OK. Let me clarify. If I’m looking down at my component laying on the crumb tray (with the laser to the north and the button to the east) . The following information holds:
East - by far the loosest fit. The slots are largest on this side. They appear to be larger with regard to both the N-S and E-W dimensions.
West - appears to be the tightest fit.
I do have a set of calipers. Measured from the top of the crumb tray to the top of the laser trolly (has glowforge logo)
NW - 114.2mm
NE - 114.2mm
SE - 113.8mm
SW - 113.9mm
There is some apparent variability here, but I don’t think that it is likely more than my error in placing the caliper reliably.
That amount of variation in head-bed height won’t make any difference. Mine is more than that.
Do you have scrap you can cut a handful of slots (five or more) into on the left side, then again on the right, and see if they are consistent?
What I’ve done (for purpose of perfect fit, not issues) is made small test pieces with different sized tabs, notches and slots, small squares with tabs that vary by a small fraction, slots and notches the same. I print them on my material and then find the best match for my needs.
You could make a similar “tab” gauge, then print a bunch of slots as I said above and see which fits best on each side. Use it like a feeler gauge.
Not saying you should adjust your design from there, but it would at least allow you to verify that the issue is consistent.
We see that the vertical kerf varies by as much as 60 microns and the lateral kerf varies by as much as 80 microns. This seems like quite a lot to me. Certainly a limitation for any sort of precision machining, which is why I bought the glowforge.
I have to suspect that this is a power fluctuation because of beam misalignment. I should note that in the process of trouble shooting this issue I found impact damage to one corner of my crumb tray. It is not enough to mess up the planarity of my tray, but it suggests that my system took a beating somewhere along the line. It must have been during transit, because I’ve babied it.
I’m open to suggestions. I suspect that I should move this thread to support.
I just cut five .5" squares from 3mm acrylic, one in each corner of the bed and one in the center. I scored each so I could track where each was cut.
Dimensions range from 12.44-12.62mm, which is pretty much in-line with your results. What’s interesting is that even across an individual piece, it varies as much as .06mm (i.e. the width of each piece when measured near the top, center, and bottom.)
I’m intrigued that you’re seeing what I’m seeing. I can’t believe that there is that much density variation in cast acrylic. I’ve done a decent amount of work with other polymers and density variations like that just don’t happen. Not in a polymer like acrylic.
I think that it is more likely that the acrylic is melting and depending upon the airflow it is cooling with different geometry. The airflow definitely varies across the bed. It will also vary depending upon whether the cut is vertical or horizontal.
What system and powers did you use?
I have a basic system using Full power at 110 speed.
I’d like to see if this variation is similar with other materials.
After reading your detailed post in P&S, I had a thought… (I don’t have “TV”…)
In short, I measured the holes left from the squares I cut, and they have as much “variability” as the squares I cut from them. I don’t have a way to match up each piece as I used scraps for the test, but this leads me to a conclusion which supports my past experience - it’s not the kerf that’s varying, it’s the cut path/precision of beam placement.
There was another thread on this a while back, where someone had noted a difference in size depending on axis. I had done some similar tests during that discussion and found the same results - the machine simply isn’t as precise as they were hoping for.
Glowforge is sending a replacement crumb tray. That will not likely solve what I’m seeing. You can see from the latest update that the variability seems to be highly related to temperature of the cut. The odd thing is that the NS and EW kerfs are quite different. It could be a speed difference. I’ll explore more in a few weeks when I return from vacation. Thanks for your input.