Re: [CAD_CAM_EDM_DRO] DRO Encoders

james owens wrote:

> As pointed out to me the other day by Jon Elson USDigital do not make
> a 1000 line encoder disk let alone a 2000, it is 500 lines per rev in
> quatrature. The next point is the value for Pi. What is yours? When I
> was at school in the dim and distant past Pi was 22 / 7, which in
> decimal is 3.142857143

This is fine for a slide rule, but a more accurate value is 3.141592654

> my calculator does not have any digits. What I'm saying is that the
> normal measure we use for Pi is 3.142 so clearly there is an error
> here. Then there is the problem of how the disk is mounted on the
> machine, can you honestly say that you can machine to 0.001" as
> measured on every micrometer. Your own, yes maybe but if I came alone
> and measured any piece you made I would probably get a different
> number. I am not demeaning your skills, what I am saying is that
> homeshop machinists we do not usually have the very high accuracy
> measuring equipment that industry can afford. We machine to a fit and
> use measurements as a guide to achieve that fit.
>
> So what we get when all the errors are added together, with the
> inexact Pi, the small disk with its error in diameter traveling a long
> way, in relative terms and all this additive error! A DRO which is
> harder to impliment than a linear scale.

Well, assuming the disk is round, you can always measure linear travel
with a dial test indicator and
gauge blocks, and then compute the precise constant to divide encoder
counts by to get inches,
or mm, as you prefer. I performed this test with my machine, and found
the screws were accurate
enough that no correction was necessary. A full inch travel came out
well within one encoder count
of the design value. That's with precision ground leadscrews, of
course. If you turn friction wheels
to convert linear travel to rotary, you could possibly get it dead on,
but as long as the wheels are
round, you could calibrate for whatever circumference they come out to
be. Cyclical errors would
result from out-of-round wheels.

The kicker with the friction wheel idea is that dirt will cause erratic,
non-repeatable errors.
The string and drum approach can, at least in theory, be shielded from
the chips. I suppose
you could enclose the whole track the friction wheel rolls on, to keep
chips and debris from
causing trouble.

Jon