Re: [CAD_CAM_EDM_DRO] Closed Loop PC Based CNC

Hi Gary,

I don't know the Rutex drives all that well but I will try to answer
your questions.

> 1) Does the drive have a PID Loop to control following error
> calculated between comanded steps and encoder feedback?

Yes

>
> 2) Do the drives have both Proportional and Intergral Gain (linear
> error and time base error multipliers)?

You can control P,I and D. With the current version this is via 3 pots
but on the latest version it is all done by software via the PC parallel
port.

>
> 3) Do the drives attempt to reduce any following error by iterating
> to near zero or only monitor a "lost steps" count?
>

The I part of the loop will try to keep the error within 1 count. I used
some Rutex step/dir to +/- 10V converters a while back and with no input
steps they slowly cycled between two encoder counts (i.e. +/- 1 count).

> 4) If the drive cannot keep up or exceeds the "lost steps" error
> count does it shut down and drop a "drive ok" signal? Can this signal be
> used to take the other drives off line by controlling the "drive enable"
> signal to the other drives?
>

They have a fault output. I can't remember off-hand if that o/p can be
used as an input as well. I tied all the following error outputs into my
EStop chain.

> 5) If the drives fault and go down, the control may not know and
> keep chugging along like it had good sense. Is there an drive output to
> the software to prevnt this?
>

I used the fault signal to EStop the machine. The computer soon knows
about it then :-) The new version drives allow you to define the maximum
following error before a fault. The older version is fixed to 1024 counts.

> Some other desired features possible with a closed loop system are "soft
> limits" and homing based on the encoder marker pulse. I would want to
> drive the axis in the direction of the home limit at traverse speed
> until the mechanical limit is tripped, switch to a slow approach speed
> and drive until the marker pulse (Z/ZNOT) is detected.
>

Mach2 http://www.artofcnc.ca has soft limits. It also has quite a few
other really useful features. It is a bit of a steep learning curve to
start with but it is worth the effort. I don't think it currently
supports a marker pulse reliably.

> All in all however, I completely agree that this is a much better
> approach than where we are now.

I hate steppers. I have used quite a few stepper driven machines and
they have all lost steps at some point. No matter how careful you are
tuning them you will eventually find a job that has just the right
combination of moves to make them lose steps.

Les