Considering a Full Rewire on a Working Schaublin 125 CNC

Has that bug been reported to the Devs?

RotarySMP wrote:

Has that bug been reported to the Devs?
 

Yes.
github.com/LinuxCNC/linuxcnc/issues/3455#issue-3089348330

There's another spindle synchronization bug, possibly related, that I encountered at the same time last year.  @Aciera found a workaround for it, but I don't know if the two are related.
github.com/LinuxCNC/linuxcnc/issues/3556#issue-3398986003

I don't have the skills to really figure out what's causing the two problems, much less contribute to a fix.

EDIT - I somehow messed up pasting the forum links when I first reported the bug, but one can highlight the link and right-click to open the forum thread.

I had time to test some more today.
Let me start with a big thanks for the great, helpful feedback.

I started trying some runs with G33. Multiple runs with the same settings resulted in only one tool mark.
Increasing the RPM did in fact result in a slightly lagging toolpath (cutting on the right side of the insert).
This was not very pronounced, even between 100 and 900 RPM the difference was visible, but not more than a few tenths of a mm.
At 2000 RPM, the new tool marks were close to the middle between the earlier ones. But I don’t really intend to thread at 2000 RPM anyway.

Going back to G76, but with Q, H, E, and L all set to 0, gave me my first nice thread.
Reintroducing Q29 still gives a nice thread and cuts on the left of the insert, as one would expect.
H2 also still worked, and after that E0.75 but L0 gave the same good result.
So only L2 was left, and I am now using the same G76 line as before, so this has to be the problem.
But no it also cuts fine.

At this point, I am uncertain if I should be happy that I can cut threads now or worried that I have no idea what was wrong until now.
I cannot think of anything I changed in between.The only thing is, I had a look at the belt connecting the spindle to the encoder.
I had this apart and inspected, checked the belt (nothing notable to see), and put it back together. I don’t think that should have changed anything, maybe slightly different belt tension?

I was at it again today.
Threading does not work again, absolutely the same behavior as before.This time I could try G33 with the problem present. Running the exact same code multiple times, every subsequent run cut slightly to the right of the last by very even amounts.Now I am really confused. It worked a few days ago. That day it worked over and over, and now it does not work at all.
RPM does not seem to make a noticeable difference, or at least it does not work no matter how low I go.I don’t even have an idea what that tells me. The hardware is obviously capable of working. Is it some intermittent contact somewhere?The encoder being the problem is really the only thing that makes sense to me, but all tests I did point to it being OK.If anyone has an idea what I should test to find this, I would highly appreciate it.

Could the coupling between the Z axis motor and ball screw be slipping? 

I don’t think it can be the Z-axis clutch, as I have a glass scale on Z.
So a slipping clutch should either be corrected or, more likely, end in a stop due to following error.

Sorry, I forgot that. Can you get a scope onto the spindle encoder index pulse?
 

You don’t have to apologize at all. I had a look at the index signal before, but not with the rest of the machine running. So I will test this with the axis drives and VFD running to make sure.