Re: Test Peice for 3D finishing

Hugh,

The ball-bar you mention is being used by more and more shops to
"prove out" their machine(s). People will come and calibrate your
machine using this setup. Newer types can be used in other ways to
check axis linearity,x to y to z squareness, etc.

Before this system, laser interferometry was/and is used. This is
STILL the most accurate way to prove a machine , BUT it is a very
much more time-consuming and "picky" process.

On a side note, I was at a trade show back in '92 and a company had
THE "answer to my problems of machine setup and proving!" It was
about a 6 inch cube which projected(with extreme accuracy, I was told)
3 laser beams at right angles to each other. It sold for $30,000!!
I asked for a demonstration and was told, "We'd love to, BUT THE
BATTERIES ARE DEAD!!!?"

Gotten a lot of humorous mileage out of that one over the years :-)

There are variations on this "project a beam, and measure to it"
technique which pop up from time to time.


Of course, one typical way to prove a machine is to use it to make
test parts, in much the way referenced in this thread. If the part
ck's out, the machine is ok (for that , and similar TYPE parts) Use a
different test part for other req,s.

Ballendo

>The simplest I've read of (and one of the only) is a bar connected
>to the spindle and to the table with ball joints. The bar has a
>sensor to measure extension/compression. The CNC is then directed
>to "cut" a sphere (or circles) about the ball joint on the table.
>Any deviation from spherical is an extension/compression of the link.
>By monitoring the sensor you get a one dimensional indication of
>the 3D (or 2D) motion of the CNC. This could be run for different
>rod lengths and speeds.

>What other methods are used in industry to determine CNC performance?