Adjusting minor dimension errors by tightening/loosening belts?

LowRider CNC Advice
1

I know that @vicious1 has found that there may be some minor variation in tooth spacing on some belts (searched but can’t find the thread), and the standard method to correct is to adjust the steps/mm to compensate.

I am (re-)calibrating my LR3, and the initial measurements showed that on a 2700mm move (Y axis) I was actually moving 2698.25 on the non-rail side and 2697.75 on the rail side. On the X axis a 1760mm move was actually moving 1759.5mm.

Because there were different correction factors on each belt, I started to think that it wasn’t actually the belt tooth spacing at issue (at least on my build), but rather minor variations of tension on the different belts was causing the overall differences.

Because the pitch on the belt tensioning screws is 0.5mm/thread, I adjusted the X belt 1 full turn tighter, the non-rail side Y belt by 3.5 full turns tighter, and the rail side Y belt 4.5 full turns tighter. I checked the belts and they don’t seem to be overtensioned, and after re-measuring the travel, all three distances are exactly the same as the move command (within <0.25mm).

PXL_20250724_1955581011
PXL_20250724_19555810111920×1446 127 KB

Just thought that I would post this info, so that others trying to calibrate their machines can hopefully dial in their distances without resorting to modifying their steps/mm.

Comments and/or thoughts…?

3 Likes
2

My general fear with calibration is that people are either not able to measure to the needed level of accuracy, or there are systematic errors showing up as calibration error and correcting it will hide the issue instead of fixing it.

However, you’re talking about fine tuning and your machine is well established. So I support you and this method.

3 Likes
3

This is my first stop in belt length (pitch) calibration, after all you’re tensioning them… low/high tension is going to cause them to be under/over the expected pitch and mismatched tension between belt pairs is going to show up as different pitches in belts from the same batch.

…ignoring longer spans and gravity (not an issue for the LowRider).

…ignoring high accelerations of the motion system and loads on the cutting tool (you may end up at the target, but not to the points you expected while under motion).

1 Like