Re: [CAD_CAM_EDM_DRO] Why closed loop?

vrsculptor wrote:

>Rant on:
>
>On servo based systems why would you ever want to do closed loop
>control through the controller? The step direction solutions: 1)Gecko
>reading encoder, driving motor directly -or - 2) Rutex drive reading
>encoder and providing +/- 10 volts for servo amp all seem much safer
>and easier to implement than someting like EMC with a servo to go
>card. Why use software to do something that is available in hardware
>at lower cost? The servo to go, closed loop approach, looks like the
>worst of all worlds.
>
>

What is the following error of the Gecko or Rutex servo systems? Can
you tell?
No, in fact, you can't, without adding external instruments. With the
Gecko, you
know you are within +/- 128 encoder counts (steps) because it doesn't
fault, but
you have no idea how close to the edge you are at any moment.

With either of these systems, you lose position any time there is a
fault or emergency
stop.

With EMC and true closed loop operation, you retain position even in a
crash or e-stop
situation. That can be important if you just machined off your
reference point. That's
not a good way to plan your work, but sometimes it is is the only way to
do something.

Also, the computer can monitor the following error at all times, and
make the following
error limit proportional to velocity. This allows you to set a VERY
tight following
error limit up to the maximum cutting velocity, but allow that limit to
expand for
rapid feed rates.

This also allows you to switch between manual machining with DRO
functions, and
CNC mode, which some people still feel is important.

Finally, you can use the position feedback to graph machine performance,
so you can
check that the machine's motion accuracy is remaining as tight as you
initially set it
to be.

With open-loop control, you are blindly trusting an autonomous system to
put the machine
where you want it, with no assurance that it is really THERE. You do
have an
assurance that it is "close", but no better.

With my true servo system, I have encoder resolution of .00005" (Yes, 50
u-In) per
count in X and Y, and .000025" in Z. For the X and Y, a Gecko would
fault at
128 times .00005" or .0064", which is a LOT of positioning error. I
generally
have following error at rest of 1 encoder count, and at 60 IPM it runs about
.0015". And, I can re-verify these errors any time I want to.

Jon