Re: AMAZING STUFF!

On 12 Mar, CAD_CAM_EDM_DRO@onelist.com wrote:

> Message: 16
> Date: Sat, 11 Mar 2000 19:18:39 -0500
> From: "Matt Shaver" <mshaver@...>
> Subject: Re: AMAZING STUFF!
>
> > From: "Tim Goldstein" <timg@...>
> >
> > Terry,
> > Also, Alan up in Canada was
> > working on an external PIC based solution a while ago.

> What I wonder is what will the control algorithm do when it senses that the
> axis is lagging? In a servo system the amplifiers are commanded to increase
> the motor speed. In a stepper system this would imply increasing the step
> rate. If the motor has stalled, increasing the step rate won't do any good.
> Steve Stallings stopped by today and we discussed this very issue. He felt
> that lost steps were most often due to resonance problems and if you've just
> lost a few steps because you went over a "resonance bump", then it may indeed
> be possible to "catch up".

Matt;

My "solution" - now on the backburner, as a result of
upgrading to ballscrews - was designed to get around the problems of
backlash in the leadscrews.

Simply, my system has Step and Dir inputs AND outputs on my homebrew
DRO. ( A PIC, an LCD and some rotary encoders )

The PC sends S&D to the DRO and the DRO "echos" the S&D info to the
motor controller.

Should the DRO find two consecutive steps issued with no corresponding
movement detected, it "pauses" the CNC software. ( CNCPro has this
feature ) The DRO then issues additional steps ( unbeknownst to CNCPro )
until suitable movement does occur.

Once the axis is "lined up", the "pause" is released and operation
continues normally.

This solution seems to me to be innocuous, as the "stopping and
starting" should only occur at the VERY beginning of a move, and
really, only at a direction change, so any "messup" of the finish
should take place at the very beginning of a cut.

It was never my intention to fix a resonance problem by this means. To
do so seems to me like fixing a poor idle problem on a car by upping
the idle speed to 4000 RPM - not really a fix.

Resonance should be cured, not just glossed over, methinks. DanM's 1/4
step controllers should SOLVE a lot of people's problems in this area.

Alan

--

Alan Rothenbush | The Spartans do not ask the number of the
Academic Computing Services | enemy, only where they are.
Simon Fraser University |
Burnaby, B.C., Canada | Agix of Sparta