"DoN. Nichols" wrote:

In article ,
Richard J Kinch wrote:
DoN. Nichols writes:

I know that a servo motor powered system run by Windows would
scare me to death. The thought of a BSOD (Blue Screen of Death) just
after a rapid move command was issued, with no system ready to stop it
at the right place is not comforting.

That's pretty unlikely as the continuing pulses require a running system.

Note that I said a *servo* motor powered system. With stepper
motors, you are right. But with a servo, there is a card in the
computer which has a collection of D/A converters. You write to one of
them, and it outputs a voltage. The voltage is proportional to the
desired speed, and is fed into a servo amplifier, which drives the servo
motor at that speed.

Write the speed value to the D/A, and it starts putting out that
voltage, (no need for the CPU to keep remembering to generate pulses),
and continues with that voltage until a new value is written. If it
happens to be rapid move to the position to start a cut, you will likely
have parts of an endmill bouncing off the walls. (That happened to me
once, thanks to programming a rapid move through the workpiece without
remembering to raise the spindle to clear it. And it used to be a solid
carbide end mill. :-)

A servo at top speed can be a *lot* faster than a stepper, which
runs out of steam at high step rates. The machine which I used to use
occasionally at work (a clone of a Bridgeport retrofitted with ball
screws and servo motors by Anilam, the maker of the controller) was
capable of 200 IPM rapid moves. And modern machines are capable of more
than that.

In my limited experience a runaway machine happens far more often due to
operator error (many times) than system problems (once or twice).

Yep -- I can agree with that. But the problems with a servo
system thanks to the computer going non-compos-mentis at the wrong
moment are more severe than with a stepper system.

Any CNC system needs limit switches and a panic button for such events.

The limit switches do shut down the servo amp, and ideally lock
the rotor as well. Whether it does it soon enough at 200 IPS remains to
be seen.

At 200 IPS, by the time you even *think* of reaching for the
panic button, it is all over. :-)

However, if it is a normal cutting speed, the panic switch can
indeed save you.

Enjoy,
DoN.
--
Email: | 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 ---

I thought that most of those internal servo controller cards had their
own onboard CPU and did not rely on the host machine for processing
power in the servo loop. The onboard CPU would handle the move based on
higher level commands from the host (move to x) vs. requiring the host
to generate a step pulse stream as would be required for an external
step/dir servo drive like the Geckos.

With this type of arrangement, if the host system crashed, the servo
drive would simply complete the commands already in it buffer and stop.

Pete C.