Motor won't step at low jog rate

Hello, i have a strange issue which is that one of the two motors controlling my Y-axis won't move - but only if the jog rate is very low, if it's up it works totally fine. The machine has been tested with another piece of software so i know the hardware isn't at fault. I'm wondering if this could be a problem with microstepping as it's only doing this at very low feed rates?

I'm using a Gecko G540 and four Nema 23 380oz motors, the Y axis is driven by two motors (Y and A) and i just doubled the Y step/direction on the A pins for the dual motor control.

So when i first fired it up today it worked perfectly with LinuxCNC, after ~10 minutes it started doing this weird thing. What happens is that one of the motors takes a second or two to start rolling, again only at low jog speeds.

I tested swapping Y and A - the problem is with the A driver quite clearly and not the motor. Again, it works perfectly with Mach3 so the hardware is not the issue (i just reverfied it).

I couldnt' really find much info on how to configure my machine (a KRMx02) so i did some research and came up with the following (which may or may not be entirely right):

I've set it up as a XYZ axis machine and used the "Gecko G540) driver type as the driver timing seconds - any chance this template is whacky? I've done the jitter tests and abused my computer.

Parallel port setup: (for the G540)
2 - X Step
3 - X Direction (inv)
4 - Y Step
5 - Y Direction (inv)
6 - Z Step
7 - Z Direction
8 - Y Step
9 - Y Direction
16 - Charge Pump

Axis setup:
Steps per rev: 200
Microstepping: 10
Pully teeth: 3:1
Leadscrew pitch: 2.8647
Max velocity: 15 in/s
Max acceleration: 30 in/s2

Same settings for all axises except Z that has a leadscrew pitch of 6.

Many thanks for any insight!

Hello,

Are you saying that you are using the same parallel port pin to drive both Y axis?
If so I wonder if it could be a low voltage problem on the parallel port.
What mode is your parallel port in? It can some times make a small difference in the output.
You could also move the second Y driver to a different parallel port pin and just combine them in you hal configuration file.
Also double check your cable connection.

Rick G

So as for the parllel port, it's set to EPP in bios - i had the option of using 1.7 or 1.9 and i went with 1.7. I'm not using the same parallel port pins, but rather i configured the pins that are supposed to control the A axis to respond to the Y direction and step. This was the only was i could seem to slave the two axises together, but if there's another better way i would absolutely try it. I will check the HAL config, i haven't even looked at that.

I will def. test the other step settings for the Gecko and report back.

What really weirds me out is that the problem doesn't happen until 10 minutes and that it's not always the same response; some times the motor doesn't turn at all, some times it just has a time-delay.

Take a look here...
www.linuxcnc.org/index.php/english/forum...e-axes-in-emc2montec

Rick G

well i tested the new timing numbers and it started screwing up after about ~5 minutes anyway. Restarting the controller and LinuxCNC seems to do the trick for another ~5-10 minutes and then it happens again. I've tried it with several different kinds of gcode too of course.

I'm trying to look into how to configure this as a true XYYZ axis machine with HAL, maybe that will help.

okay so i modified the gantry model and it works except the A-motor isn't running when i jog the Y axis, i've got this in the INI:

[TRAJ]
AXES = 4
COORDINATES = X Y Z Y

and in the HAL:
loadrt gantrykins coordinates=xyzy
setp gantrykins.joint-0 0
setp gantrykins.joint-1 1
setp gantrykins.joint-2 2
setp gantrykins.joint-3 1

if you are running steppers with two axis that are the same and you are not going to install separate home or limit switches you can just output the step signals to two sets of parallel port pins. Just use the standard XYZ setup and something like this...

x axis
net xdir => parport.0.pin-04-out
setp parport.0.pin-05-out-invert 1
net xstep => parport.0.pin-05-out
setp parport.0.pin-05-out-reset 1

slave axis
setp parport.0-08-out-invert 1
net xdir => parport.0.pin-08-out
setp parport.0.pin-09-out-invert 1
net xstep => parport.0.pin-09-out
setp parport.0.pin-09-out-reset 1

That should be close but from memory. The one is reversed here because if you had two motors at opposite sides and they both pointed inward a move would be with one clockwise and the other counter clockwise, but it is easy to switch.

If you go this route you do not need gantrykins

Rick G