Windows 11, Dell Optiplex i5-4590, Sainsmart 3018PROver, Sainsmart 5.5W module
Importing a DXF file of text and lines. On the Lightburn screen, everything looks correct and aligned properly. When I run the laser engraver, some text letters are misaligned. WHITE in the middle of picture 2, WHITE and GREEN in the bottom of picture 3. The misaligned letters changes every time I try to do an engraving.
I thought it might be the engraver steppers missing steps, but the misaligned letters are in the beginning of the project and if it was missed steps the entire engraving would be off.
Same results at 1800 or 900 speed, and same results if the engraving is done in multiple passes. Picture 1 was three passes at 1800, picture 2 was 900 single pass.
Is that also true of the Preview display, where it shows the laser head path?
If the Preview is correct and the results aren’t, then the problem is (almost) always mechanical.
A loose setscrew in the leadscrew coupling on the motor shaft is typically the cause of intermittent backlash, because the screw / shaft can jam in one position until enough torque slaps it to the other position. The screw may feel tight, because it’s jammed, so loosen the screws and retighten them while wiggling the coupler to ensure they’re seated on the flat.
This test pattern helps identify mechanical problems:
Scale it to fit the platform and run it as fast as it will go in Line mode with optimizations turned off and power set to mark a sheet of cardboard. Any differences from the design will be informative; a photo will let us look over your shoulder.
Protip: Use cardboard / paper / whatever, rather than spendy stuff, until you’re certain you’ve found and fixed the problem.
Trial cut on chunk of plywood, running at 3000mm/min. scaled to 70% of the original. Still has similar text line offsets, which don’t show on the preview.
Couplings are tight, I use the split clamping couplings, not the cheap setscrew couplings.
My thinking is if it was a mechanical issue, it would be more random. Scaling puts the potential errors in different physical locations, but every test I’ve done has the text errors lined up in the same vertical drawing segment, even if the engraving is a different size. If it was a mechanical error caused by a loose coupling or damage to the drive shafts it would be in a similar physical location each time. But the errors are in similar logical locations relative to the text, not relative to the physical size of the engraving.
And the errors are at the beginning of the engraving sequence, if it was missed steps the offset would continue throughout the rest of the engraving sequence, as the engraver is not a closed loop system.
Load this file, scale it proportionately to fit your bed and run it at high speed with just enough power to get a good mark on cardboard. Any discrepancies from the file will indicate machanical problems. Upload a photo of the results so we can look with you. BacklashTest.lbrn2 (68.8 KB)
Test pattern sample. As far as I can see it’s perfect.
Looking at the DXF file, I suspect a case of computer voodoo, something is wonky in the transition from DXF to g-code, but I don’t want to spend three week digging through the 500kb DXF and 11,000 lines of g-code of my original design to figure out why it’s jumping.
I took a screenshot of my CAD software used to prepare the DXF file, then loaded the resulting .PNG file into Lightburn. Engraved the .PNG file on to a scrap chunk of MDF. Here’s a photo of the results. All text is aligned properly, so my guess is the offsets in text seen in the original engraving attempts are something wonky going on in the CAD > DXF > G-Code sequence.
It’s easy enough to make my workspaces as .PNG or .TIF files, and Lightburn does a good job of scaling as required so I won’t spend any more time trying to figure out why the DXF files don’t translate perfectly.
You can eliminate speculation fairly easily and avoid potential future red herrings about future if you can provide the Preview and g-code examples if you’re inclined.
