Experimental Physics and
Industrial Control System

Mark Davis <[email protected]> · Tue, 14 Jul 2015 11:20:38 -0400

Hi Mark,

As is probably true with most labs, we do indeed have some VME stuff,

but for the main control system we have been slowly eliminating VME when

we can. The hardware is just so much more expensive than using fairly

generic rack-mount PCs running Linux and using Ethernet (directly or via

Ethernet-to-serial/GPIB/whatever adapters) to talk to stand-alone devices.

So far the list of requirements I have compiled from searching through

the mass of documents regarding the new FRIB infrastructure has not

turned up anything particularly demanding, but of course many details

are still in flux so I am trying to be "generous" in terms of reading

between the lines and extrapolating a bit regarding what is likely to be

required once the details start firming up. As long as some particular

feature doesn't result in a drastic increase in price, a solution that

does a bit more than meet the minimum requirements will undoubtedly save

us money in the long run.

This is what I have come up with so far that is related to the choice of

controller:

 - Complex coordinated motion (2 axes minimum for insuring smooth

movement of things like pairs of jaws)

 - Fault outputs (so an external PLCs can manage the state of a group

of related devices)

   - Support for 3 limit switches (min, max, and home/reference positions)
   - Support for on-the-fly scans:

 - At minimum, the ability to generate pulses for each axis

based on current position of each axis

        - Ability to capture/buffer the data itself would be a plus

 - Support for a variety of analog or digital encoders. Currently

specified types include:

        - Linear motion potentiometers
        - Rotary Sine/Cosine potentiometers
        - Digital (optical, rotary, relative, absolute?)

 - Support for local control box (i.e. continue to track/report

position, disable movement if limit switch activated, etc)

   - Prefer models that do not require external motor drivers
   - Ethernet interface
   - Firmware version readable by control system
   - Ability to update firmware remotely

Regarding closed loop mode: While I can imagine circumstances where you

might want to disable it, I would think that any controller that

supports an encoder would also provide a closed loop mode. Is that not

so? I would think there would be little advantage to supporting an

encoder if it couldn't be used directly by the controller during

movements (i.e. isn't this question the SAME as "Does it support a

position encoder"?)

Regarding price/axis: I haven been given a specific $ amount on this

yet, but we have been using the Parker GT6K single-axis controllers

which are around $3000 each, and I have been assuming that most newer

multi-axis solutions will be significantly less expensive.

Mark

On 7/10/2015 1:34 PM, Mark Rivers wrote:

Hi Mark,

I think that to make a useful recommendation on this it would help to know more about the requirements.

- How many axes?
- Do you have VME infrastructure already available?
- Do you plan to use encoders?
- How important is cost per axis?
- What are the performance requirements?
   - Microstepping?
   - Coordinated motion?
   - Closed loop feedback if you have encoders?

For example, Tim's suggestion of the MAXv would be best suited if you already have VME, and don't need closed loop encoder feedback or very fine microstepping.  The XPS or Delta Tau, on the other hand can do closed loop feedback, very fine microstepping, and coordinated motion.

If you only have a few axes and no VME there are solutions like the ACS MCB-4B that are about $400 per axis, while many other solutions are $1,000-$1,500 per axis.

Mark

________________________________________
From: [email protected] [[email protected]] on behalf of Mark Davis [[email protected]]
Sent: Friday, July 10, 2015 9:52 AM
To: [email protected]
Subject: Stepper Motor Controllers

Hi all,

I am currently in the process of reviewing the capabilities of DC
Stepper Motor Controller/Drivers and was wondering what other labs are
currently using and what they like/dislike about them.

I have read Matthew Pearson's summary from 8/2010 regarding the Delta
Tau VME Turbo PMAC 2, Delta Tau Geobrick LVM, Newport XPS.

  From his description, I assume the Delta Tau Geobrick LV is a
significantly different device (at least physically) than the Geobrick
LVM.  At the very least LV is MUCH smaller (15.4" x 4" x 7.2").  It is
still a PMAC2 based controller, but they make a point of the elimination
of analog signaling:

       "The absence of analog signals – required for typical motion
controller/drive interfacing – enables higher gains, better overall
performance and tighter integration, while significantly decreasing
costs and setup time."

This may explain how they managed to make it so much smaller the the LVM
model (which, as near as I can tell, they no longer make).  It might
also help avoid the noise problem that Matthew mentioned regarding the
LVM model.

  From reading various other Tech-talk postings my impression is that the
Delta Tau PMAC controllers are very capable, but complex and require a
lot of effort to get set up correctly.  However, that may be mitigated
to some extent by using the synApps tpmac module (and BNL's page
regarding setup and configuration).

What other controllers are being used?  What are the advantages,
disadvantages, and problems that have been encountered?

Mark Davis
NSCL/FRIB
[email protected]

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System