On Thu, 3 Feb 2005 16:52:33 -0800, skuke wrote:

On Thu, 03 Feb 2005 16:19:33 -0800, Eric R Snow wrote:


But my point is that .001" is
pretty tight to hold day in, day out with any kind of efficient speed.

I have a Wells Index CNC knee mill. The mill itself weighs 6000 lbs.
So it's bigger than your typical bridgeport type knee mill which weigh
about a ton. The control on the mill failed a few years ago and I
bought a new control from AJAX. Even though about 20 years old and
plenty used the machine is quite accurate. So holding .0005" location
when boring holes is a breeze. I can get closer by changing the
program to compensate for position error because the machine repeats
to .0002". So an older machine can be checked for error in both
positioning, flatness, and squareness and if good enough then it might
fill the bill. I bought my Index mill mainly for complex shapes, not
for production. My next mill will be enclosed, the Index is not. Being
enclosed will keep the chips and coolant in the machine and allow
higher speeds and feeds without throwing chips and coolant all over
the place.
ERS


I agree with you that one can program out (minimize) positional errors by
approaching from the same direction to eliminate back lash issues, center
drill, drill... With all due respect, true hole position is much easier to
hold and measure than the 3 axis contouring the original poster wanted.

My experience with the knee mills (and even the Fadal to a much lesser
extent) is that to interpolate a round circle is pretty gosh darn difficult.
You wind up with some "egg" shaped hole with dwell marks at the axis' when
the ball screw must actually come to a complete stop for a moment to turn
the other direction.

Please try it: Interpolate a ~3-4" circle in a rigid piece of aluminum with
a good rigid endmill of your choice. Take a few finish passes. Then,
without removing the part from the vise (to eliminate other problems like
stress/tension relief), measure TIR with a test indicator. I'm pretty sure
you'll have at least .001" TIR.

So I think trying to hold 3 axis timing on a cheaper, less rigid machine
will be all but impossible.

...just based on my experiences. ...your results may differ.
My point was that lead screw error can be programmed out. The
backlash, once measured, is entered as a parameter in the control. One
of the first things I did after installing the new control was measure
backlash and enter it in. Unfortunately, there is some variance in
backlash over the length of the travel. But, the machine will
interpolate a hole that's .0006 out of round in the approximate center
of travel. Of course, you are right about a smaller machine maybe not
being able to hold tight tolerances. My knee mill is 3 times heavier
and aboutt 1.5 times as large as a bridgeport mill. So it is massive
enough to get good accuracy on heavier parts.
ERS