Can't get DoubleStep to work ? SOLVED

Hello,

I'm trying to use DoubleStep as described here to get more speed for 2 of my new mill axis.

So in HAL file, I've added the reset function to base thread :
addf parport.0.reset base-thread

set the reset interval :
setp parport.0.reset-time 5000

set the required step space to 0 on required stepgen :
setp stepgen.0.stepspace 0

So far, all is well, but when I add the lines that should activate the reset for the selected pins, for instance :
setp parport.0.pin−02−out−reset 1

LinuxCNC won't start anymore, and I get an error message stating that there is no such variable or parameter (can't remember exact wording, sorry). I've checked and rechecked the lines, can't figure out what's wrong.

Any idea ?

Thank you for your help.

Looks like you have accidentally corrupted your hal file (wrong editor?)

setp parport.0.pin-02-out-reset 1
setp parport.0.pin−03−out−reset 1
setp parport.0.pin−04−out−reset 1
setp parport.0.pin−05−out−reset 1
setp parport.0.pin−06−out−reset 1
setp parport.0.pin−07−out−reset 1

Hello,

Sorry, I'm doing this at home from my Mac and basic TextEdit, the file on the mill PC looked fine. The 'problem' lines looked like :
setp parport.0.pin-02-out-reset 1


[EDIT] I actually edited the file on my Mac, transfered it to the PC, and edited there back and forth trying to find the source of my problems. It's quite possible that I did not actually re-typed the lines there, just copied/pasted portions of the text. Can you confirm that the lines should be OK ? [/EDIT]

still looks broken...
One thing to try is comment out the bad lines, start linuxcnc and then
type

halcmd show pin parport* > realpins

in a terminal window

and then check that the pins are there and use realpins as a copy-paste donor
file

OK, sorry for the stupid mistake.

I'll follow your advice tomorrow and report here.

Current max speed is 12 m/mn, I'd be happy to move up to 15 m/mn with the same resolution and (not so good) old PC.

Thank you for your help !

Is there a reason you are not using stepconf?
It will set up double step automatically.

Chris M

Hello,

This is now fixed, thank you all for your help.

@PCW : you were right, and I did have broken text in my .hal file. from editing the file on my Mac.
This was easily fixed by retyping the lines on the PC directly, I'll make sure I don't get bitten by this again

@cmorley : I did use stepconf to get started, and it was a great help. After that, I couldn't resist the urge to tweak and try to understand what is what. This is a slow and error-prone process, but I find it oddly satisfying, at least when I make progress...

As a finale note, I add to lower the signal filtering (P1-00) that I had set up on my Delta ASDA-A2 drives : for a while, the motors wouldn't move with DoubleStep activated, although the signals _were_ generated. Halscope was a great help to troubleshoot this issue.

Bottom line is that my gantry machine is now moving at up to 250 mm/s, with a resolution of 1/100° mm per pulse. I'm using an old PC with average latency, parport and a cheap 5 axis BOB... What could be better ?

Next step is to find a way to get a stable spindle speed from the PWM->0-10v signals, at the moment it's a bit wobbly.

Thank you all for your help

Hello,

I'm revisiting this "solved" issue, as I've just discovered that double-stepping was responsible for the loss of precision of my machine. What happened is that, with DoubleStep active, the computer did not send the exact count of pulses required for a given move, as seen from the servo monitoring software. This created unpredictable and rather large errors on dimensions, repositioning etc...

With DoubleStep deactivated, and everything else unchanged, the drives get the exact count of pulses expected for a rather long movement (3500 mm), and come back exactly to 0 as expected.

So, I'm left with a question : have I hit a hardware limitation of my rather old PC : Dell Optiplex 760 (Dual Core E7300 @ 2,6Ghz (?) - 2 Go RAM - 80 Go HD - Graphical Card - 64 bits -> AMD64), or is there a chance to get this working with better parameters on the double-stepping ?

Thanks for your help

I dont think that the parallel port driver can generate the wrong number of
pulses with a slow CPU speed because of the way it uses feedback but
The pulse length will be _much_ longer in the "single-step" mode.
You might try increasing the reset time to see if you still lose steps.
Are you using a breakout board? sometimes isolated breakouts are quite slow.

Thank you for your reply.

I am using at the moment a basic Chinese "5 axis" BOB, but don't have detailed specs of it. The plan is to upgrade to something much better when the rest of the machine is validated.

So you are saying that if the reset time is too low, the pulse length might be too short for the BOB ? This means that I should also check the minimum requirements on the drive's documentation. I'm using Delta ASDA-A2 servos.

I will test longer reset times and report.

Thanks again for the suggestion