Reversing direction causes motors to jerk

I'm in the process of putting together a kit from Avid CNC (CNC Router Parts). Everything is running, but I'm stuck on a problem. The motors run smoothly until they reverse direction. When they reverse, they jerk and make a CHUNK sound.

I don't think the problem is with rack and pinion setup because

1. It happens to all 4 motors, even the Z-axis ball screw.
2. The motors jerk like this even when I run them disconnected from the machine.
3. All of the motors run alright until the direction changes.

So here's everything I know:

1. I've tried lowering the acceleration, but it didn't seem to have any effect.
2. I've tried both using backlash compensation and not.
3. The steps/inch configuration is correct as measured with an indicator.
4. Whatever the problem is doesn't cause much backlash. I generally get an error of ~0.003in the first time I move a motor, but after than I can move it back and forth a lot, often with no drift.
5. So far, changing the tension on the rack and pinion parts hasn't had any effect.
6. It happens to all 4 motors, even the Z-axis ball screw.
7. The motors jerk like this even when I run them disconnected from the machine.
8. All of the motors run alright until the direction changes.
9. I've tried different microstep configurations from 2 up to 16. None of them had an effect.
10. I've attached my linux-cnc config.

This is my setup:

Linux CNC
NEMA 34 stepper motors
Gecko GR214V drivers
Mesa 6I25/7I76 I/O cards

This is how I have the drives wired to the break out board:

Gecko P1 pin 2 (stp) -> 7I76s TB2 pin 3 (step0+)
Gecko P1 pin 3 (dir) -> 7I76s TB2 pin 5 (dir0+)
Gecko P1 pin 4 (+5v) -> 7I76s TB2 pin 6 (+5vp)

I've tried several timings, none of them had any effect. Here are two settings I tried:

DIRSETUP = 100
DIRHOLD = 600
STEPLEN = 2000
STEPSPACE = 1000

DIRSETUP = 1000
DIRHOLD = 1000
STEPLEN = 4000
STEPSPACE = 3000

Here are two videos of what's happening:

https://vimeo.com/381842213
https://vimeo.com/381840195

Anyone have an idea as to what the problem might be or what I should test next?

That looks like not enough current to the steppers.

Quote:

---

Originally Posted by nlancaster

That looks like not enough current to the steppers.

---

Hmm ok. I'll try to verify the current getting to the motors. Why would a lack of current cause this?