Hello Guest it is March 28, 2024, 01:23:37 PM

Author Topic: Problem: Losing position if more than one axis wired up to breakout board  (Read 11634 times)

0 Members and 1 Guest are viewing this topic.

Offline Bfive

*
  •  13 13
    • View Profile
Hi all,


My system:
I'm using Smoothstepper USB to a breakout board with step/direction servo drives. The machine is using ballscrews and linear bearings, no backlash compensation, very little friction/sticking in axis movement.

The problem:
I've been struggling with a mysterious loss of position with my machine. It is extremely bad such that I'll run some G-code for 20 seconds then send it back home, only to find it off of zero by an inch. It seems to only happen mostly to one axis, although sometimes I can readily see both X and Y positions off. Also, if I feed rate override to a lower speed (like 10%), it seems to exacerbate the problem, or at least make it more obvious. It's strange because I can run a lot of code at high speed (feedrate override 200%) and be off by less than if I ran the same code at a low speed (feedrate override 10%).

My troubleshooting:
I've swapped drives, swapped signal cables, fiddled with Mach3 settings, all without a good result. I think I may be on to something, though. Last night I tried disconnecting all the signal cables (enable, step, direction, GND) from every axis except one. With only one axis wired up, it seems that the problem goes away. However, once I connect another axis to the breakout board, the problem resurfaces. It doesn't matter if the drive is disabled in Mach, just the fact that the other axis is wired to the breakout board seems to induce the position loss.

Any ideas of the cause?

Many thanks!
« Last Edit: April 11, 2012, 03:42:25 PM by Bfive »

Offline Hood

*
  •  25,835 25,835
  • Carnoustie, Scotland
    • View Profile
Do you have shielded cables, especially for Step/Dir signals?
What kind of drives?
What kind of breakout board?
Hood

Offline Bfive

*
  •  13 13
    • View Profile
Hi Hood,

Thanks, for replying.
I have unshielded ethernet cables for step/direction, and realize that this could be a source of interference. I'm using the twisted pair of wires, one as signal, one as ground for each of my step/direction signals. I do have some shielded cable that I will try to test with. Should I ground the cable shield to the breakout board or to the drive side or does that not matter?

Also I'm using a really really old CNC4PC C10 breakout board. I think I got it sometime in 2005 or something, so it's a very early version. My machine has been sitting around most of the time since then, half-finished and only partially tested. Have there been problems with early versions of this board? I had some issues with lost position even when I used the parallel port at a low frequency. I also ordered a new breakout board to test or use wiith other projects.

Another question: do I need to power the breakout board with +5V? It seems to run with just the power from the parallel port, and I'm not running any relays or more power-hungry things from it (just step/direction/enable). Could I be creating a ground loop by powering the breakout board with a 5V power supply, which has a different ground from the 5V of the PC side? Sorry, my understanding is kind of vague. I've read a lot but not quite digested it all.

Thanks!
« Last Edit: April 12, 2012, 07:20:44 PM by Bfive »
i would power the breakout board from the pc.

Offline Hood

*
  •  25,835 25,835
  • Carnoustie, Scotland
    • View Profile
Sorry dont have a clue about the CNC4PC BOBs, never used them, I would imagine it best to look up their documentation and see whats required. Regarding shields, you take one end of the shield and connect all grounds at a single point such as a stud in your control cabinet.
Hood
just reread your post. yes, the c10 must be powered from the pc. either by tapping the usb port or just connect into an unused connector off the pc power supply. red 5v+ , black-. also the enable jumper on the board. i have 3 of these boards and have not had any problems.
Just to see if it corrects the issue.
Have you tried turning on the rapid override on the settings page, and setting it to a lower value (try 100-150)? Not sure how fast your rapids are, but start limiting the rapids and see if you see results.

If it does that would suggest to me that there might be an issue with motor tuning.
When I tune one motor at a time they seems fine, but when they all start moving together they have issues. So I don't push the limits at tune, I shoot for 70-80% of max rate at tune.

Wish I had a better handle of the proper way to tune and use the max rate for motors.

Offline RICH

*
  • *
  •  7,427 7,427
    • View Profile
Quote
Wish I had a better handle of the proper way to tune and use the max rate for motors.

Can you post some specifics Ya-Nvr-No?
What kind of motors ...steppers or servos?

It would help me in a write up that I have been picking away at.

Bfive, hope you don't mind my post and don't want to highjack your thread.
Thanks,
RICH
« Last Edit: April 14, 2012, 06:49:23 AM by RICH »

Offline Bfive

*
  •  13 13
    • View Profile
Thanks for all the input guys,
I've played with motor tuning (now set very conservatively) and other software settings, but I think the issue was/is hardware. I got a new breakout board, powered from a separate 5V usb port (and 5V jumper enabled on smoothstepper). So I have one USB port going to the smoothstepper as signal, another USB port providing 5V power (all from the control PC). So far it seems to have resolved the problem. I've only done very limited testing, but it seems to no longer have a massive loss of position after a few moves. I'll update if anything else comes up.
« Last Edit: April 18, 2012, 02:43:26 PM by Bfive »
Hello Folks,

I am having a similar problem with a new 2 axis machine I am building. It is using Baldor servo motors with Microflex drives, a Smoothstepper and a C25 BoB plugged directly into the SS.  There is also a joypad plugged in which is used to jog the machine.

The motors have been tuned and the steps adjusted. If I tell the axis (via MDI) to move x10z5 everything is fine. If I then tell it, via MDI, to move back to x0z0, the z axis wants to move in the wrong direction and trips the limit. However, if I send it first to z0, it will go there and then x0 it will do that too. Seems strange.

If I use the joypad and the "jog to gcode" to record some motion, it gets completely lost and misplaces several inches on each axis!!

Any thoughts on this?

Thank you,
Davey Boy.