Re: [CAD_CAM_EDM_DRO] Rolled Ballscrews

I will repeat the inexpensive way I use to map ballscrew errors:

Buy a 6 and 12 inch micrometer test bar. These are a very
precise length standard. I think I paid about $30 for the 12"
at MSC.

These can be "cloned" with a simple stop and indicator setup
to make many accurate test bars. I get them close with a belt sander
(comparing the length to the standard) and then stone them to final size.
The ends do not have to be flat and parallel
as a micrometer face... in fact the store bought units have crowned ends.

I use 1/4x1/4" ground bar stock to make them.

These can now be stacked like gage blocks on an axis to
measure the linear positioning error. Hold them in place with tape or
magnets between a stop and the axis moving element. A little wiggling might
be needed to make sure the longest part of the gage is touching its
neighbors.You get points every 6 inches using a snap bore gage to measure a
small gap left between the test bars and a portion on the carriage. Or use a
good caliper for in between points.

The only thing you have to watch is temperature... a few
degrees of warming of the bars from holding them in your hands or any other
warmer or colder thing that contacts them
can cause thermal dimensional changes that can add up to thousandths in
several feet. Particular attention has to be paid
to this when the bars are being sanded to size...that will heat them up a
good bit and they will measure long until cool again.

I should also mention that the data points obtained are subject
to the Nyquist criteria as is any sampled system. I won't go in to the math
but will just say this:

The shortest periodicty one can measure is twice the distance
of the sample points.

I do a test over a short length with very close together
points to test for drunken threads. I usually find very little and
use 6" increments for the rest of the data.

Leslie Watts
L M Watts Furniture
Tiger, Georgia USA
http://www.rabun.net/~leswatts/wattsfurniturewp.html

----- Original Message -----
From: <ccs@...>
To: <cad_cam_edm_dro@yahoogroups.com>
Sent: Monday, November 05, 2001 7:38 PM
Subject: [CAD_CAM_EDM_DRO] Rolled Ballscrews


>
> I've been thinking about rolled ballscrews and their potential lead error.
> It seems this would be a problem for two reasons:
>
> 1) inconsistent preload
>
> 2) cyclic and cumulative positioning errors