On 2013-11-25, jim "sjedgingN0Sp"@m wrote:


Jim Wilkins wrote:

[ ... ]

http://www.mmsonline.com/articles/ge...icron-accuracy
"The axis position feedback system uses a 0.5-nanometer scale to
reliably track axis motion commands programmable in steps as small as
10 nanometers."

Yes that machine should be able to cut 3d parts so they match
the 3d computer geometry to within a tenth.
However, the CAD model of the part can be made a million times more
accurate than even that machine can manufacture.

I didn't say there were no way to produce complicated 3d
parts that are within a tenth of the 3d CAD model.
I said that I doubt anyone reading this was doing that. You
still haven't changed my mind.




I've watched a large aspherical Germanium infrared camera lens being
diamond-turned to a mirror-like submicron finish on a CNC lathe.

Turning something on a lathe doesn't need a 3d model.
That is not 3d machining.

Turning an *aspherical* lens does need a model, and a very
precise and mathematically complex one at that, to cause it to focus
where it should. And you can't check a Germanium lens by the techniques
used for visible-light lenses -- they are opaque to visible light, so
you want it cut right the first time. Same applies to silicon lenses.
I've seen both used in various experimental Infrared cameras.

And watching someone else do it doesn't count as doing.

He did not say that he was watching someone else do it. He was
watching the *machine* do it. No *human* does it with CNC. At best
s/he feeds the data to a program which generates the motions necessary
to produce the surface to feed to the CNC machine tool.

Enjoy,
DoN.

--
Remove oil spill source from e-mail
Email: | (KV4PH) Voice (all times): (703) 938-4564
(too) near Washington D.C. | http://www.d-and-d.com/dnichols/DoN.html
--- Black Holes are where God is dividing by zero ---