Re: following error question

["Richard Everett" <rick-at-skyko.com>]

Thanks Jon.  I knew velocity 10 would mean 600ipm, I was just trying to test
out the 1.4ghz machine to see if it would bog down.

I have not messed with CYCLE_TIME (it is still 0.001 for x,y an z), but I
know I have done more than 1000 steps per second.  I measured 5khz step
pulses on my oscilloscope.  How exactly does CYCLE_TIME effect max
steps/sec, and what do you think is going on in my case?

Richard


----- Original Message -----
From: "Jon Elson" <elson-at-pico-systems.com>
To: "Multiple recipients of list" <emc-at-nist.gov>
Sent: Wednesday, October 02, 2002 4:31 AM
Subject: Re: following error question


>
> Richard Everett wrote:
>
> > Hi again, In anticipation of getting the pulse multiplier for my Gecko
> > drive, I installed BDI-EMC on a 1.4 ghz athlon thunderbird machine.  I
> > then setup the emc.ini file with steppermod.o and input scale and
> > output scale set to 5000.  I set the max velocity to 10,
>
> 10?  Remember, that is in user units per second.  If your user units are
> inches, that is 600 IPM, which is
> a bit fast for the larger machines.  Perhaps the Fanuc control could
> deliver the power to do this.
>
> and ran EMC.  Much better performance as far as the slugishness of the
> interface.  When I tried to manually jog the motor 1 inch, I got a
> following error after it had moved part way.  Are there any other
> variables using steppermod that I should try? You should start with the
> jog velocity at a low rate, and move up in steps to find out what is
> happening.
> You should also learn how to make the logging / graphing function work
> (it is quite easy, just a few
> mouse clicks, but you have to install gnuplot to see the graphs).  Then
> you can find out why it is
> getting the following error.  But, most likely, it still can't put out
> step pulses fast enough.  The theoretical
> limit would be around 500,000 steps a second, but the port is not the
> bottleneck.  What is your
> CYCLE_TIME in the axis sections?  If it is still at .001 second, you can
> only get 500 (or maybe
> 1000) steps/second, I think.  If you command a move faster than that,
> then the commanded
> position will get ahead of the actual position, and a following error
> will occur.
>
> Jon
>
>
>