Re: [CAD_CAM_EDM_DRO] Re: gearmotors,encoders

Mariss Freimanis wrote:

>
> Just did a little experiment, I like to have data backing up what I
> say, otherwise my mouth (or fingers in this case) gets me into
> trouble.
>
> I took a 36" length of steel tubing (1" OD, 3/4" ID, uncertain
> parentage regarding alloy) and clamped it in a vise at the far end. I
> superglued a mirror sliver at the other end and trained a HeNe laser
> on it. I put a torque wrench to the same end and cranked it until the
> reflected laser beam on the wall indicated a torsional angle of 0.1
> degree (equal to about 1 count on a 1,000 line encoder). The wrench
> indicated 47 to 50 in/lbs torque, or about 800 in/oz.
>
> P.S. Resonant frequency of this mess (tubing + wrench) was 180 Hz. I
> taped a differential large-area silicon cell on the wall where the
> laser beam hit, AC coupled a 'scope to it to see the frequency.
> Seemed kinda high.

Oh, very GOOD! Thanks for the measurement! Well, this is REAL
bad news for the far end encoder crowd! Now, I can say with authority,
DO NOT DO THIS! IT will DEFINITELY not work with modern servo
amps. I have rolloff starting at 1 KHz on my servo amps, with some
power gain up to about 6 Khz. So, this thing would ring like a bell
on my servo amps, and any others with high bandwidth. My ancient
Allen-Bradley 7320 CNC control had a positioning loop bandwidth
of only 2 Hz, and presumably was designed to connect to servo
amps with low bandwidth also, perhaps 100 Hz or so. This is fine
for a 20-ton bed mill, but totally inadequate for a Bridgeport or lighter
machine. Cutting forces can deflect the machine way away from the
desired position before such an amp could catch up!

Any reasonable servo system for a Bridgeport or lighter machine should
have at least several hundred Hz bandwidth on the velocity loop, and
that puts you right into the area where the leadscrew could ring in the
torsional mode!

Actually, the 180 Hz seemed kind of lower than I would have expected,
depending on the moment of that wrench. But, it clearly indicates the
far end approach is bound for trouble!

The P&W machine that someone mentioned used that design was
probably used with a low (or no) bandwidth servo system.
Didn't the original "tape-o-Matic" use a pneumatic control?
Or was that the original "Cint-i-Matic"?

Jon