Re: [CAD_CAM_EDM_DRO] EMC, freqmod.o, A-axis, .ini settings, oh my!

"j.guenther" wrote:

> Hi Tom,
>
> Welcome to the club!! I have been fighting this problem since the end of
> June with no joy. It appears to be a granularity issue, and the lower the
> steps per unit the worse it gets. I am going to reload my EMC setup with
> BDI 2.16 so I can try it with steppermod.o as I am currently running the
> BDI-TNG which only works with freqmod.o. I really wish one of the EMC
> developers would track this down and fix it. I have thought about running
> mine with input and output scale set to 28800, and units set to 1. That way
> a 90 degree move would be g0 a0.25 which is not the way I want to run it but
> does work. Sounds like you have a Sherline mill and rotary table. I have
> the Sherline 5400 mill and got the cnc rotary table for my 55th birthday.
>
> John Guenther
> Sterling, Virginia
>
> > -----Original Message-----
> > From: Tom Benedict [mailto:benedict@...]
> > Sent: Monday, August 19, 2002 11:24 AM
> > To: CAD_CAM_EDM_DRO@yahoogroups.com
> > Subject: [CAD_CAM_EDM_DRO] EMC, freqmod.o, A-axis, .ini settings, oh my!
> >
> >
> > Here's the setup:
> >
> > I've got a stepper-based mill that's got four axes. The A axis is a
> > rotary table that's bolted to the mill bed so the axis points parallel to
> > the X-axis.
> >
> > I'm running the latest rev of the BDI, and the motors are set up to use
> > freqmod.o.
> >
> > I'm trying to tune the .ini parameters for the fourth axis so I can
> > address the rotary table in units of degrees. The linear axes are being
> > addressed in units of inches.
> >
> > I've got 200 step/in motors that are being half-stepped. The linear axes
> > have 20TPI leadscrews, so that winds up being 8000 steps/inch. The rotary
> > axis has a 1:72 worm gear on it, and uses the same motor as the linear
> > axes. So at 5 degrees per revolution of the motor, that's 80 steps per
> > degree.
> >
> > Now for the troubleshooting:
> >
> > I've got the linear axes pretty happy. I don't have my .ini file in front
> > of me, but I use an INPUT_SCALE and OUTPUT_SCALE of 8000.0 0.0,
> > LINEAR_UNITS of 0.03937007874016 in [TRAJ] and UNITS of 0.03937007874016
> > inside each [AXIS_*] section (I'm using inches), and a P of 500.0.
> >
> > I've got ANGULAR_UNITS set to 1.0 in [TRAJ] and UNITS set to 1.0 in
> > [AXIS_3]. With all the other parameters set the same, if INPUT_SCALE and
> > OUTPUT_SCALE is set to 8000.0 0.0, the axis works, but it's not being
> > addressed in terms of degrees. If I change INPUT_SCALE and OUTPUT_SCALE
> > to 80.0 0.0, the axis ceases to function.
> >
> > I've tried dinking with this a fair amount. I can marginally get it
> > working with the SCALEs set to 800.0 0.0, and address it in terms of tens
> > of degrees (so G0 A9.0 would rotate it 90 degrees), but obviously this
> > isn't what I'd like it to do.
> >
> > Has anyone got a stepper system using freqmod.o to work with a rotary
> > axis? Can anyone share a success story that would help me get this thing
> > working?
> >
> > Any and all help is much appreciated.

Well, I don't understand the causes of this, but I will look into whether my
Universal Stepper board will relieve this problem. Since it moves the generation
of steps outside the PC into hardware, it gets rid of most of the software restrictions
of freqmod/steppermod. This board emulates a servo system with an encoder counter
and a step pulse generator. (If open loop, then the counter counts the step pulses.)
Step pulses up to 300,000 steps/second can be generated, with very smooth ramping
and fine (100 nS) resolution of step pulse timing. I will attempt to set up such a
4 axis system and test it when I get back home. (I'm out of town this week.)

See http://pico-systems.com/univstep.html for more info on these units.

Jon