Re: [CAD_CAM_EDM_DRO] Role of Tachos in EMC / Servo Systems w/ Encoders

jackw19x wrote:

>My understanding is that EMC servo systems using motion controllers
>like Servo-To-Go or Pico connect encoders to the motion controller
>and tachometers, if present, are connected to the servo amps.
>
>Since EMC is the brains of the operation, why isn't the tacho
>feedback directed to EMC via the motion card rather than to the servo
>amps--the brawn?
>

The reason is that the tach is continuous-time, whereas the encoder is
discontinuous-time,
ie. sampled at some finite rate. The tach can have much higher
bandwidth than the CPU-mediated
servo cycle. When CNC controls were REALLY slow, such as 60 or 100
updates/second, the
tach was VERY important. With EMC running quite nicely at 1 KHz, and
easily capable of
more with a decent Pentium-III or so, it is a lot less important. I
have to say that the
servo amp's DC gain can be turned up a LOT higher with tach feedback to
keep it stable.
This higher gain reduces static error and error at low speeds by a lot,
but would likely
cause vibrations or outright instability without the tach.

> What use can the servo amps make of the tacho
>information? The plus side is that at least the servo amps I have are
>all set up for a tacho voltage signal, so it's one less thing to be
>interfaced to the motion control unit.
>
>

The classic servo motion system has the CNC control sensing position
ONLY, and deriving
all corrections from that, and sending a velocity command to the servo
amp. Because of
inertia, friction, spring, motor resistance, etc. the servo amp can do a
much better job of
deliviering exactly the commanded velocity if it can READ the velocity
from either the
encoder or a tach. The advantage of the tach really shows up when the
axis is moving slow
enough that there are only a few encoder counts per second. The servo
system that reads
the encoder will twitch when every encoder pulse comes in, while the
tach-based amp
will move smoothly down to amazingly low speeds. Mine starts to break
into stick-slip
movement below .01 IPM!

Jon