Re: [CAD_CAM_EDM_DRO] ...but how much precision????

Bill Kichman wrote:

>Has anybody truly measured the degree or precision you are getting from your installation? Is 1 thousandth attainable from home cnc?
>
>

Sure. I did it with a straight edge and a stack of gauge blocks. I
could only measure over one
foot with just gauge blocks, then I had to use other extension blocks of
lower precision.

My high-accuracy ground screws were dead on over the full length of
travel. But, I
found backlash of .001 to .0015" on the various axes. There are a bunch
of components
that have slight flexibility, and when they all add up, you get that
much lost motion
even though the screws and nuts have backlash in the range of .0001"
maximum.

My table, saddle and knee are pretty worn, so some of that may be table
rock, etc.
(The underside of a Bridgeport-type machine gets worn concave, the top
of the saddle
gets worn convex, and this allows the table to rock slightly when it
reverses direction.
If you mill a circular path on a piece of material, then go to the
center and sweep an
indicator around the circle, you can see the effects of this. You get a
saddle-shaped
sort of path that tells you the magnitude of the slideway wear. It is
around .001"
for a 7" diameter circle on my machine.

To directly answer your question, it is possible, but costs $. A
Sherline with stepper
motors is WAY off this level of accuracy, although it may be able to
make a very
small part to this accuracy. I doubt any of the mini-mills can hold
this, either, over
more than a few inches. I can tell you that a brand-new Bridgeport CNC
machine
can't be trusted to hold .001" over the full travel of the machine, in
all axes, together.
It can probably be trusted to hold .001" throughout a one foot cube
after careful
adjustment. The big CNC makers go to incredible lengths in both basic
mechanical
structure, fitting and alignment at the factory, and realigning at the
customer site
to get this level of accuracy over the full range of the machines. They
run temperature-
controlled coolant through their hollow ballscrews, and have temp-controlled
chillers on the spindle head, for instance! You should be aware
that picking up a one-foot square component (presumably a hollow box if you
actually are going to lift it) handling it a moment so it absorbs some
body heat,
and then measuring it will cause it to have expanded by .001" or more.
Therefore,
unless you make a light, slow, finish pass on the part while it is
flooded with
temperature-controlled coolant will cause it to come out to the wrong size.

Especially when doing heavy stock removal by turning, I have to leave
the part
spinning in the lathe for 10 minutes or so to cool before making the
finish cut!
Otherwise, the part will mike out fine just after that cut, but will be
several
thousandths small after it has cooled. You really have to watch workpiece
heating when making precision fits.

So, be aware that simple positioning accuracy is not all there is to
making a precise
part.

Jon