Re: Threaded Rod WORKS! & cyclic error

--- In CAD_CAM_EDM_DRO@y..., ccs@m... wrote:

>
> And many hobby machines may not need much accuracy, however there is
> one aspect that I will call 'incremental precision' which they do
> need. The primary reason I am building cnc machines is to make
> calculated curves for musical instrument tooling. I did some of

this

> manually with a crude digital readout, not realizing that the

readout

> had a very small (maybe .002") amount of cyclic error due to the
> unwise way I connected the encoder to the screw. This showed up in

my

> finished part as a very noticeable ripple superimposed on the

contour

> - it even transfered from the bell mandrel I had machined through

into

> the first few trumpet bells, until I carefully filed it out.
>
> So, absolute accuracy may not be important for many applications,

but

> cyclic error can really ruin your day if you are machining curves or
> even diagonal paths. Of course, cyclic error over a short distance
> may be as likely to result from poor screw mounting as from errors

in

> the screw itself.
>
> Chris Stratton

Cyclic error can show up in the most surprising places and takes a
real PITA effort to eliminate or minimize. The printing industry
uses halftone dots to simulate the appearance of shades of gray or
color. Take a look at a B&W newspaper photo. Before the days of
digital imagesetters, halftones were created from glass masters,
created by ruling engines. This technology is well over a hundred
years old, consisting of 4 to 5 pitch leads screws and LARGE (2 - 3
foot)divider disks. The leadscrew moved a scribing head on a gantry
over a sliding bed, scraping a resist off a glass plate that was
later etched and filled. Typical rulings ranged from 65 line pairs
per inch up to about 200. Two of these etched and filled glass
plates were cemented together with the pattern at right angles to
produce an array of "windows", which eventually produced the
different size dots you see on the newspaper photo.

Film copies of these patterns, duplicated from one film to another
dozens of times, still show the characteristic ruling defects. The
worst is "thread drunkeness" where the geometric center of the thread
is a corkscrew pattern, rather than a straight line down the center
of the material.

In this application, because these pattrns are overlaid on each
other, the BIGGEST concern is the elimination of cyclical errors, and
absolute precision is less important.

When we first talked to machine builders about creating an array of
lines, they proudly stated their positional accuracy. When pressed
further, if they would guarantee the placement of 200 lines per inch
over 60 inches, all gaurantees vanished.

The solution arose in the form of a HP laser interferometer to
measure absolute location using custom electronics, a real good,
large diameter leadscrew, a rotating nut, a randomizer, and debunking
every vendor's claims.

All of this equipment is retired with the changing of the printing
industry, but the spare HP system is still available. $13,000 new,
tested but never used. :(

But threaded rod has its place.

Larry Goldberg,
Carlstadt, NJ