Update from older software version

[tblough]

Go ahead and post photos of the complete encoder installation so we can see how it is connected to the spindle.

[martyscncgarage]

This is odd though: the spindle encoder has been set to -8000 since this lathe arrived (why negative?) but I'm counting 10000 ticks per revolution in my manual test. Shouldn't these numbers be the same? This might actually explain why it looks like it's making a 1.25 pitch instead of an 1.0 and also losing sync.

Mark a spot on the spindle at the 12oclock.
Go into the PID Menu
F1 Setup
F3 Config (137 password)
F4 PID
slowly rotate the spindle ONE full revolution and notate the value before and after the one turn. Subtract one from the other if it didn't start at 0. Was it 8000 or 10000? If 8000 then leave it alone. If its 10000, change the value in the Spindle section of the Wizard from -8000 to -10000

The - in front of the number is to let the control know the encoder is rotating the SAME direction as the spindle. Sometimes its needed, sometimes it is not.

Also, as you rotate the spindle, you will note an ASTERISK show up once per revolution and that should be at the same point. That is the index pulse.

[LeChuck]

Yes, that's what I did this morning. 10000 steps and it shows in the same spot.

This is as close as I can get. 2500 on there. So 2500x4=10000 steps per rev.

This seems to have fixed the pitch as it looks like it's doing a 1.0 now. But it's still cross threading some.

[tblough]

If it's still cross threading then there is probably noise on the encoder lines. Is the encoder cable shielded? Is the backshell on the connector metallic and connected to the shield? Is the shield soldered to the DE9 connector housing?

[slodat]

If you are losing steps on your DMM axis servos, give strong consideration to using the DB25 step/direction signals. I have done extensive testing on this on two completely different machines and both times I lost steps when using the Acorn screw terminal step/direction outputs. For whatever reason the schematics still show using these. In both of my (and many other users) cases simply changing to the DB25 step/direction outputs solved my lost steps issues.

If you want to know for certain you are losing steps go to an position on the axis, using the DMMRV software check the encoder counts for each DYN4 drive, move the axis around a lot. Use MDI to go back to the start position and check encoder counts. If they match, you are not losing steps. If they don't I highly recommend switching to the DB25 step/direction outputs and run the same test again.

[LeChuck]

slodat, thanks, I will look into that. I think I might still have to buy the DMM cable because mine seems fine for reading/saving data to the drive, but it doesn't look like it does proper RS232 communication to get the drives to move, but I'll go back to have a look at that when I get a chance.

So, on the shielding...I think this might be it! I don't think that the portion of the cable that goes to the encoder is shielded, but I'm no specialist. Looks like your standard soft black cable.

When I get a chance, I'll open up the connectors and look. I already had to repair one of them, as the green ground wire to the VFD had its solder joint broken off inside the DB9. I took the whole cable out of the bundle, pulled it away from the others, ran the job again, and lo and behold, a dirty but correct M10 1x0 thread! I need to do it in fewer than 20 passes though, and my last pass at 0.2 is actually way too big, but it's a learning curve.

I'll have to try several to see if it's consistent, because that's super important, I can't ruin an expensive piece of resin that I bought (only broke a thread once doing it by hand, and that was really my fault). I do have to go earn my salary now and put in a few extra hours, but a bit more peace of mind for now

[martyscncgarage]

What you should see when you open up the DB9 shells is that the drain wire is soldered onto the shells of all THREE of them. The splice point and at the DB9 on the Acorn. It is important. I believe the black cable has a shield.

[LeChuck]

Ok, I will look all those over.

[tblough]

I also recommend METAL backshells with shielded cable, not plastic.

[LeChuck]

OK, I'll look into that too.

Still not out of the woods on the threading. It's still cross threading a little, but if it happens early in the cycle then it's not very visible since it's a very shallow pass, but still there, then the rest of the passes finish the thread over it and it's mostly ok, but it's still not right. Maybe that means there could still be some shielding issue, just less than before?

Also, I noticed something that doesn't seem right either. The VFD and the DRO disagree on the spindle speed, not sure if that has an impact of the threading issue. The VFD is higher by a ratio of about 1.2, not far from that spindle rev/turn difference. So it looks like there's still a setting somewhere that isn't right.