In article et, Doc says...

... It
occurred to me that I never said that I'm drilling into wood.
Let me amend that, plexi that's been glued over wood.

First off you cannot drill anything with the degree of precision
you mention in your topic - namely "absolutely." Don't
even bother trying to do that. Instead take the time to think
about the tolerance for the positional accuracy of the hole,
based on how it is supposed to function. Then put a *number*
on that tolerance. In other words, "the center of the hole
has to be located within 1/64 of an inch."

Or, "the center needs to be true +/- 0.001 inch."

That sort of thing.

You should be aware that one can put a very accurate hole in
place this way, and get it within five thousanths or so:

Scribe the position using layout dye and the appropriate
measuring tools.

Use a sharp scriber to indent the workpiece at the intersection
of the two lines. A magnifier helps here.

Pick up the indent with a center punch, and punch it.

*Inspect* the newly formed center punch with the magnifier.
It will not be at the exact center of the crossed lines.

Drift the center so it *is* at the center to the best of
your optics.

Place the workpiece in a drill press and use a small drill
to start the hole. 1/16 or so is about right for most cases.
The drill will _PICK_UP_ the center even if the spindle of the
machine is not perfectly aligned with the centerpunch. It
does not need to be.

Drill through with the small drill. (I do NOT use a center
drill or rigid spotting drill for the technique. I *want*
the smaller drill to be somewhat flexible so it picks up
the center on its own)

Drill up with the correct final size, and the larger drill
will follow the smaller hole.

Basically you cannot count on drilled holes to be round
to better than a few thou. So by the same token don't try
to spot them better than than that: +/- 0.003 or so.

Any design requiring sub-thousandth precision on drilled
holes is doomed from the outset. If this is the case,
change the design _now_.

Jim


--
==================================================
please reply to:
JRR(zero) at pkmfgvm4 (dot) vnet (dot) ibm (dot) com
==================================================