John Kasunich wrote:

> I'm curious, what is the resolution of the scale?  and the encoder?
> Chris and I were discussing this last night a bit - I'm of the 
> impression that its a lot harder to get high counts per inch in a scale, 
> since you don't have the mechanical advantage of the screw working in 
> your favor.
> 
> 
Exactly the problem!  The only way to get really high resolution 
in a linear scale is to make the scale analog and use an 
interpolator.  A high-res scale is 2 um, that is about .00008",
which isn't so bad, but it is only 12700 counts/inch.  I'm 
running 20,000 counts/inch on X and Y and 40,000 counts/inch for 
Z on my Bridgeport.  This has tach feedback to the velocity 
servo amps, and runs quite smooth.  Without the tachs, however, 
I think it would be a rattling monster!  I am running 128,000 
counts/inch on my minimill, and it gives nice smooth motion 
without tachs.

My guess is the huge scales on the G&L are not going to be that 
high a resolution, maybe more like 10 or 20 um.  You CAN"T use 
that as the only position feedback, it will just be too coarse.
It either needs to be interpolated (if you have the hardware) or 
it has to be combined with the shaft encoder.  If the ballscrews
are tight, maybe the simplest thing is to use the shaft encoder 
for main feedback, and use the linear scale with a special 
program for creation of the compensation table.  Oh, but you 
wanted live compensation for thermal correction while the 
machine runs.   Hmmm, that is going to be a lot tougher.
And, localized heating of the leadscrew can lead to all sorts of
non-uniform errors, so you can't just assign an overall 
correction factor.  I think this gets a LOT more complicated if 
the scale has low resolution.

Jon Elson (Who is glad his machines have relatively short travels!)

-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2008.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
_______________________________________________
Emc-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/emc-users

Reply via email to