Squares sides not cutting at 90 degrees

I sent an SVG rounded square to GF, cut it from acrylic but when I flipped it it didn’t fit in its own hole

The square had come out sheared, as a parallelogram, Over 10cm it had diverged from perpendicular by 2-3mm.

I can’t make reliable interlocking parts like this… what do I do?

1 Like

I cut the square three times in the center and to the left and right and the same shear occurred.

Feels like an issue with motors?

There is at least one other owner wth a similar issue.

Are you sure you are not just seeing the effects of the beam divergence? Is the angle greater on one side then the other?

@jkopel That’s a great link, but not applicable here. the other user you quote had a different problem : that the cut edges are not normal to the material surface.

My problem is that the laser doesn’t faithfully preserve angles between lines in the plane of the material. everything comes out a little skewed. I think I have seen it mentioned one other place.

Sounds like your gantry isn’t aligned properly. You’re probably going to need to hear from support staff on this.

I’m sorry to hear you ran into trouble! I’m afraid I’ll need a bit more information to look into this.

Could you please provide the following?

  1. The design file for the print you are referring to
  2. A photo of the print that shows where the shape isn’t printing the way you expect
  3. The approximate date and time of the print

You can post the information on this thread, or send an email to support@glowforge.com with the attachments if you’d prefer. Thank you in advance!

Here’s the input. The problematic item is the outer rect (currently black filled) when executed as a proofgrade cut. It’s possible the other lines are also slanted, but that would only be a cosmetic issue. in GFUI I scaled the design to about 3.75" wide.

5%20layer%20linebased

And here’s the output on dark (non PG) acrylic, placed on a set square. you can clearly see the divergence: the metal ruler marks show about 1/16" at the bottom.

This is a little different than my issue where the cut wasn’t square to the surface. When I first started using my GF I found that squares weren’t actually square. I used a machinist’s square and carefully re-squared the gantry and haven’t had problems since. If you do this and find problems again, you will need to contact support.

2 Likes

Thanks Ben! How do you resquare the gantry? Just gently shove it to the back?

That reminds me… Before I did this bad run it spent 10 minutes “calibrating” by snapping the head against the back over and over Not sure if that’s supposed to cure or cause the problem I’m having. After a while I just turned it off, moved the gantry to center and turned it on again.

I used a machinist’s square and squared the aluminium X track against the aluminum Y track. I had to twist the gantry with one hand while holding the square in the other. I could feel the stepper on one side jump into the next alignment when it was square.

4 Likes

Thanks Ben. I really enjoyed your other post with all the diagrams and tests.

I have given it a yank. I wonder why calibration doesn’t fix that; it seems like a calibration issue.

Glad it was enjoyable. A side effect of working in tech support and also of being an engineer who has done testing in the past. If you can provide more information and definitive results it makes problem solving that much easier.

Does the GF actively attempt to square the gantry during callibration or does it simply depend on luck to be square?

2 Likes

@alexbbrown how is it working now?

I avent repeated the square test yet. I will this weekend.

We are good; I gave the gantry a solid twist and checked with a set-square.

Why doesn’t glowforge know it’s misaligned?

1 Like

I am guessing the camera isn’t able to detect small misalignment yet.

I wouldn’t characterize it as small; over the bed width that’s a 5-10mm skew

1 Like

I agree, but from what I can see the camera only uses one point of reference, which would make it small. Either that or the gantry is run against endstops to square it, though I don’t know for certain.

I’m glad you were able to fix it.

Thanks for the feedback, I’ll pass it on to the rest of the team.