SOLVED: Tuning stepper motor with Gecko 201 driver

As I mentioned in a previous thread, I'm a LinuxCNC newbie attempting to convert a lathe over from TurboCNC. My X axis has a Gecko 203V drive while the Z runs a 201, and both steppers run satisfactorily in TurboCNC. But the 201 runs terribly rough in LinuxCNC, even after trying many different timing settings.

The 203 loves LinuxCNC, running smoother and faster than it did on TurboCNC, even with less than optimal timing settings. The 203 can't do anything wrong in LinuxCNC, while the 201 can't do anything right.

So I tried experimenting with a wide range of stepper timing settings in StepConf. It didn't seem to make much difference. For now I settled on Step Time 1000, Step Space 5000, Direction Hold 20000, and Direction Setup 1000.

I tried tweaking the STEPGEN_MAXACCEL and that didn't seem to make much difference, so for now it's set at the default 1.25.

Since the 201 runs fine in TurboCNC but runs like crap in LinuxCNC, I'm going to assume that the hardware is fine and that the fault lies in the LinuxCNC setup.

I'm looking for suggestions of things to try that I haven't already tried.

BigJohnT wrote:

Does it run rough at all speeds or just in the mid range?

At all speeds, to the extent that it has much speed. It sounds like it is grinding gravel.

I don't have a spare 203V laying around. The lathe originally had 201's and then when a 201 failed I replaced it with the 203. Like you I am impressed with the 203.

I did try using much longer timing numbers and it didn't seem to make much difference.

I'll do some more reading and searching on the DoubleStep thing, and I'm also curious about the "opendrain" command?

But surely some other EMC users have some experience with 201's? Of course their setup might depend on whether they are using an isolated breakout board. I'm running a non-isolated breakout board.

BigJohnT wrote:

So your driving the Geckos straight from the parallel port?

You could say that. My BOB is a "straight-through" BOB with no signal conditioning or isolation.

BigJohnT wrote:

What might be messing with the 201 is doublestep.

As near as I can figure, the doublestep feature was enabled by StepConf.

From the HAL manual: "To setup DoubleStep on the parallel port you must add the function parport.n.reset after parport.n.write and configure stepspace to 0 and the reset time wanted."

My HAL file does have the partport.n.reset - time 1000, and it does have stepspace 0 for both X & Z.

I'm not sure if the DoubleStep matters but I will try it without DoulbleStep and see if it makes a difference.

It is inverted. If I un-invert it, the z does not run at all.