Update from older software version

[martyscncgarage]

It's not using mechanical switches at all. I only home by aligning to a mark then it looks for that servo pulse.

There are no sensors above the ball screw pulleys in use then?
Check repeatability and backlash.

[LeChuck]

There are no sensors. Only connections to the servo motors.

The repeatability is terrible right now, but it was good until yesterday. I did do these same operations several times and they weren't perfect, but nothing like this happened where it just lost its position. I just did another test. After homing, it took only one G30 move and one G28 back to get it to lose about 1mm on Z (made a mark to check). I would definitely have noticed that before yesterday's breakage, but the parts have been coming out fine on the Z axis, including really fine threading.

[slodat]

If you are losing steps, which it looks like you are, I HIGHLY recommend eliminating the resistor in the DMM cables and using the DB25 step/direction outputs. I have written this up in a thread or two on here. This has resolved my lost steps on my DYN4 machines and for other users on here. It’s easy, free and worked for many of us.

[LeChuck]

I think there's something wrong with the servo on Z and the pulse detection, or some configuration with that (I tried varying the frequency between 100k and 400k). And it's not really losing steps or backlash. This started happening yesterday and everything seems tight.

Another quick test:
- homed from the mark
- marked the position of cursor with piece of tape
- G30 then G28
- looked at cursor position, it's now about 1mm forward, after just one move
- moved cursor back to the mark and homed again
- it now homes to the new position where the G28 ended

The fact that it homes to that more forward position tells me there's something wrong with the servo/drive/reading/whatever. Even if it was losing steps and was not accurate, moving it back to the mark a couple mm and homing again (slow movement over just a few mms) should have brought it back to the original position.

I think maybe I should try swapping both motors and see what happens, but I need to make some pens as I've been out of commission for 5-6 weeks now so have to go back to the manual lathe.

[LeChuck]

If you are losing steps, which it looks like you are, I HIGHLY recommend eliminating the resistor in the DMM cables and using the DB25 step/direction outputs. I have written this up in a thread or two on here. This has resolved my lost steps on my DYN4 machines and for other users on here. It’s easy, free and worked for many of us.

Thanks, I will look into that. It's just surprising that Z was really smooth sailing for weeks until last night. It's really X that's been giving me a few troubles, simply because 0.05mm on that axis makes a big difference. But Z has been fine, and if it had diverted like it did yesterday, that would have had immediate effect. It became off by 2mm all of a sudden, then hasn't been consistent since.

[slodat]

That’s weird. Have you tried swapping drives or motors?

[LeChuck]

These things were made on this lathe over the past few days. Not perfect yet, but correct dimensions and shapes, with same settings, no bad breakage, just tweaking my X precision, so yesterday's issue just happened.

[LeChuck]

That’s weird. Have you tried swapping drives or motors?

I think that's what I want to do, I just need to make a few pens now to generate some side income. I calculated this lathe expense in terms of how many more pens I need to sell per month to pay it off, but so far it has cost me almost 6 weeks. I'm hoping I can get it right because I don't like to give up on anything but this is getting costly.

[LeChuck]

Only getting back to this now as I had to work with my other lathe today, but did this quick test:

Connected to the drive with the DMMDRV software. Homed then read the motor position when the lathe is at 0:0. 44695. Jogged around to make sure the drive does read different positions. Did a G30, then a G28 to go back to 0:0. Drive reads the same 44695 (indicating it's not losing steps?), however the Z axis is now 1 or 2mm to the left of the starting point. It's like the steps ratio is completely wrong but how would that have changed?? And it seems that only Z is affected.

[slodat]

That leads to a mechanical issue.. something slipping or similar in the motion components.