Unfortunately on my Acmer P3 I’m struggling to get perfectly perpendicular lines and it’s really throwing my work out. I have ensured the belt tension is even but I don’t know how to make sure the gantry is square on this closed unit.
It was a lot easier with a traditional open gantry!
You can check it with a framing square or a rafter square; pick the biggest one that will fit inside.
The gantry looks like it has two Y axis belts, so perhaps one is misplaced by a tooth on its drive gear, as in this larger CO₂ machine:
It’s not obvious to me if that applies to the CoreXY drive on your machine, but something similar may be amiss.
Thank you so much for your reply! I definitely think the belt could be out by a tooth, however with this machine I can’t figure out how slip it again. Even with the belt tension screws fully loosened in the back there is still too much tension to actually change the belts position on the stepper pulley.
Here’s a diagram of the belt set up for this machine.
So it really is a CoreXY machine. The manual mentions only belt tension, which likely won’t get you where you want to go.
Asking Acmer how to square the machine might be productive, but you’ll need a better demonstration of just how out-of-square it is, as those game pieces look pretty good.
Using cardboard (cheap and readily available), engrave (do not cut) a large square with sides parallel to X & Y, then measure with a good (ideally steel) scale / ruler:
If the X & Y sides are not equal, then there’s an axis calibration problem that should not happen in a new machine. You can ask Acmer for advice or use LightBurn’s axis calibration; save a backup of the machine settings before changing anything.
That assumes Acmer’s firmware allows you to change the machine settings, which may not be true.
If the diagonals are not equal, then the machine is not square, which might be mechanical.
Verify the gantry is square to the frame using those squares I mentioned earlier or a piece of paper that ought to be pretty square.
If it is not mechanically square, then Acmer should explain how to make it square.
If they can’t and have no suggestions, perhaps you can make some totally unauthorized adjustments, but let’s save that until you’ve done the measurements & reported back. 
I engraved a 300x300mm square and measured it as accurately as I could. There is definitely a slight margin of error.
In a previous life, you were the chief metrology inspector in a tool & die shop, right?
With length errors of 0.03% and angular errors of 0.01°, that machine is as good as you’ll ever get and better than most.
Based on what you measured on that square, I recommend not messing with the mechanical setup, because there is no way you can improve it and many ways you can make it much much worse.
The spot looks nicely round, so the next step will be to measure the actual kerf in various materials:
With that in hand, you’ll have no trouble matching the results @bernd.dk showed off:
Fire The Laser!
While it does seem to be fairly close when measuring, the problem I’m having is that cutting out squares or rectangles and gluing them back to back (not face to face, would which actually mask the issue!) really accentuates the non perpendicular lines.
Here’s a picture of a straight horizontal line I cut in a piece that has a perfectly straight edge aligned with the frame/bed. I then turned it over and engraved the same line which shows the that the laser isn’t cutting perfectly perpendicular lines.
Maybe this is the best I can hope for?
IMO, that’s about what it comes down to.
There are a number of possible errors …
“a perfectly straight edge aligned with the frame/bed”: the honeycomb / frame is generally not precisely aligned with respect to the machine axes. Flipping a cut piece over and aligning it with the honeycomb doubles the angular error.
“gluing them back to back”: the kerf isn’t exactly straight, so the cut edge isn’t exactly perpendicular to the flat surface. This is particularly the case in wood with all its inherent irregularities. And, as above, the edges have double the frame’s angular error.
Doing precise work means finding and eliminating all the sources of inaccuracy. The measurements on that square show the machine is capable of really good accuracy, so the next step involves accurate fixtures.
Try this: clamp a piece of cardboard / MDF / wood to the honeycomb so it cannot move, then cut it parallel to the X axis. Using the cut edge as the reference, use another piece for the engrave-and-flip test. Because the cut edge is, by definition, parallel to the X axis, the new pair of lines should be pretty nearly indistinguishable.
With a laser beam of approx. 0.1 mm it is almost impossible to hit the same path again by rotating your test piece.
Maybe if you have built a template and a test piece that can be rotated 180 degrees…
In addition to that it could indicate that your frame/bed is not 100% parallel to your gantry.
I have mounted aluminum edges on my machine bed and spent a lot of time getting it parallel to the gantry (both X and Y). If the gantry is not skewed it should work.
This rectangle was cut out with the laser, so it’s angles are truly reflective of the axis relation to each other.
The gap measures 0.69mm with my callipers. That’s seems a lot and it’s immediately noticeable to the naked eye.
The square evidently consists of an aluminum extrusion held to the scale by a pair of screws. Have you checked the perpendicularity of the square?
Edit: The error looks like 0.13°, which probably doesn’t exceed any specs it might have claimed when new.
As the saying goes: Trust, but verify.
If it’s not perpendicular you should be able to refund the machine, if you can’t fix the problem yourself, that’s the most logical consequence.
I won’t be able to “live” with that.
No I can’t. I can’t send it to customers either, even if they wouldn’t notice.
Two glued pieces looking like this is just not precise enough for me.
Try the suggestions you’ve gotten for improving the results. Report back on what you’ve done and how it worked out.
Sorry Ed, I must have missed some suggestions? The kerf adjustment isn’t too much of a concern to my current work, just that the vertical and horizontal lines are perpendicular to each other.
Instead of the engrave and flip test I did (which was unnecessarily convoluted on my part) I just cut a decent sized rectangle out of a piece of scrap in the middle of the board and checked the right angles. That way I’m not assuming the honeycomb is perfectly straight in relation to the gantry.
With a square of unknown accuracy.
To summarize:
Again: Trust, but verify everything.
I agree that would not be worthy to show off your craft. Do you have a belt sander? A quick kiss on a 120 or 240 grit belt would clean that up, and remove the char at the same time.
Just thinking out loud.
