Y Axis Travel Error

Has anyone seen an error on the Y Axis? I am getting set up to run an indexed job where the spacing between sections is 22". I laid out my scribe lines on my stock (16Ga CR Steel) at 0.5, 22.50, and 44.50.

My scale is a Lee Valley tools 24" stainless steel ruler.

On setup, I have set Y=0.00 at the 0.5" mark (my torch holder has a laser line projecting onto the workpiece so relatively easy to check alignment). I then moved the gantry Y22.000 and I was surprised that the position was not on my next scribe line. The gantry had moved 0.097" too far.

Thinking I messed up laying out the line, I measured it with the ruler and the scribe line was on the money. Next thought was my scale wasn’t really that accurate, so I checked the X movement. It moved exactly 22" according to the scale, so one of the two axis was wrong.

Since these are Acme screws driving the two axis, I found it hard to believe there was an error, so I checked for slop in the gantry, including the Z gantry, which I added. It appears to be tight and there isn’t any movement to the line projected on the sheet as I apply moderate pressure to the torch holder.

So I tweaked the Y Axis motion settings and, where the original value was 12800, I tuned it to 12758 to yield exactly 22.00 inches of movement.

So, my question is, again, has anyone noticed an error on either axis? I suppose the threaded rod COULD be off by 0.44%, but in this case it matters. Fortunately, it appears that I could tune the axis to correct the error. But I am surprised…

It occurred to me that, rather than the lead screw being inaccurate, the Y Rail Square tube could have a slight curve in it. That would project to a larger error at X=23 (where I was measuring) compared to X=0. I’ll check that out after lunch… :hamburger::beer:

Ok, I had my lunch (no beer) and I’ve measured the Y error with X=0. The error IS smaller (about 0.050") but not as small as I would have expected with a slightly curved Rail. So, at the moment, setting Y Axis Steps Per to 12758 is providing a usable result. This will cause alignment errors at the section joints, but hopefully they’re within a half of a kerf width.

Still looking for an explanation of this behavior!