Centroid Community CNC Support Forum

I have a small Logan lathe using Dmm dyn4 AC 640 dst 400w servo running on Acorn rev 3 with:
2:1 timing belt drive
Lead screw = 10 tpi with .003 backlash
Motor revs/inch=20
Drive pulse = 2000 , control = 8000, 200,000 hz
Drive has been tuned using dmm tuning software and running smoothly.
CNC12 ver 4.6 free software
I'm trying to adjust +/- axis movement with dial indicator to set exact revs/inch.
Problem: When I move X+ direction, commanded results are correct within .002 per each inch travel every move. When I move X- direction, commanded result is 0.014"/inch short (.986 vs 1.000). If I then move back 1 or 2 or more inches X+ direction, total movement will be within .002 of that X+ commanded, but it will be off .014" for each inch traveled in X-. So, its loosing original starting position.
I've checked belt tension, both timing belt sheaves set screws and tried different backlash settings. It looks like motor/drive is possibly loosing steps, but in one direction only.

sounds like backlash. Have you measured backlash?

Yes, backlash measures .003 - .004, depending upon where on lead screw its measured.
Servo Motor just doesn't turn the full required 20 revs on a 1" X- move, while it does turn 20 revs, almost exactly on a 1" X+ move.

I assume you are using the original leadscrew, driven by either the half nut or the key along with no thrust bearing to take up the back and forth slop? Going to be a lot of slop in that mechanism, Not to mention leadscrew wear.
When fine tuning the dro, always move in the direction you are testing before DOING the measurement test.

Try Marc Leonard's backlash test program. You'll just need it for Z
http://www.cncsnw.com/Lashhowto.htm

Marty

Are you using DMM cables to connect to the Acorn? If you are there is an issue with them.

Ignoring how far the cross-slide may or may not move, for now: if you get different amounts of servo motor rotation for the same commanded axis travel, with zero backlash comp, then you are dropping counts between the Acorn and the drive.

It’s easy to determine if you are losing steps...

Note encoder counts in DMMRV
Move the axis around via MDI
Move axis back to start position and note encoder counts. If they aren’t very near the same. If you are losing steps, use the DB25 step/direction outputs. I’ve documented this well if you want to read more on the topic.

Problem solved! Thanks all for your input. X axis movement/positioning overall is now +/- .001". This little bit error can be mostly fixed I think, with anti-backlash nut or new screw. As it turned out, the DB25 cables from DMM were causing this drifting issue. DMM supplied new wiring schematic (attached earlier in thread) of new DB25 cable. On to the spindle now, if I can just find a bit of time away from work!!??

I don't see the DMM post you are referring to. Please repost it?