Jobs abort with GRBL ALARM:3.
I can run the same jobs I'm using to show the issues below by exporting the gcode and directly feeding it to the machine. And they finish. Which leads me to the conclusion, that the gcode and the machine itself is fine, but serial communication by Rayforge is borked.
Case 1: rayforge_debug_2026-05-25_08-45-08.zip
Steps to reproduce with 'Poll device status during jobs' enabled:
- Draw square, edge length 100 mm
- Position square at 0, 0
- 'Cycle laser head around occupied area'
- Device stops a bit right of the top left corner with ALARM:3
Strangely it appears to be dependent on the size of the square:
98 mm: similar to 100 mm; stops some mm after top left corner.
97 mm: does one loop and then stops some mm after top left corner.
96 mm or smaller: runs without alarm.
Larger than 100 mm: stops on the left edge after about 105 mm traveled. About the same total distance as with the smaller squares.
With 'Poll device status during jobs' disabled, the 'Cycle laser head around occupied area' loops run without error.
I could live with that, but even with 'Poll device status during jobs' disabled, there is another issue.
Case 2: rayforge_debug_2026-05-25_09-27-03.zip
The job in the attached debug log stops with ALARM:3 even with 'Poll device status during jobs' disabled, shortly after starting in the upper left corner on the way down (see screenshot below). If I delete the ellipses, leaving only the 200 mm square, the job finishes.
Expected behavior
Jobs should not abort with ALARM:3.
Versions:
1.7.10
Jobs abort with GRBL ALARM:3.
I can run the same jobs I'm using to show the issues below by exporting the gcode and directly feeding it to the machine. And they finish. Which leads me to the conclusion, that the gcode and the machine itself is fine, but serial communication by Rayforge is borked.
Case 1: rayforge_debug_2026-05-25_08-45-08.zip
Steps to reproduce with 'Poll device status during jobs' enabled:
Strangely it appears to be dependent on the size of the square:
98 mm: similar to 100 mm; stops some mm after top left corner.
97 mm: does one loop and then stops some mm after top left corner.
96 mm or smaller: runs without alarm.
Larger than 100 mm: stops on the left edge after about 105 mm traveled. About the same total distance as with the smaller squares.
With 'Poll device status during jobs' disabled, the 'Cycle laser head around occupied area' loops run without error.
I could live with that, but even with 'Poll device status during jobs' disabled, there is another issue.
Case 2: rayforge_debug_2026-05-25_09-27-03.zip
The job in the attached debug log stops with ALARM:3 even with 'Poll device status during jobs' disabled, shortly after starting in the upper left corner on the way down (see screenshot below). If I delete the ellipses, leaving only the 200 mm square, the job finishes.
Expected behavior
Jobs should not abort with ALARM:3.
Versions:
1.7.10