Re: Alternatives to softGlue? (re-configurable TTL I/O line logic controllable from EPICS)

[Andrew Gomella <andrewgomella@gmail.com>]

Hi Mark,

Thanks! This definitely makes the Zebra currently the best option.

-Andrew

On Sat, Nov 14, 2015 at 2:42 PM, Mark Rivers <rivers@cars.uchicago.edu> wrote:

A PowerPoint presentation by Quantum Detectors in 2013 listed the price as ~$1,600.

If still valid, that compares favorably to the cost of the IP-20X plus the cost of a PCI-E Industry Pack carrier card.

Mark

________________________________
From: tech-talk-bounces@aps.anl.gov [tech-talk-bounces@aps.anl.gov] on behalf of Andrew Gomella [andrewgomella@gmail.com]
Sent: Saturday, November 14, 2015 12:36 PM
To: Mooney, Tim M.
Cc: tech-talk@aps.anl.gov Talk
Subject: RE: Alternatives to softGlue? (re-configurable TTL I/O line logic controllable from EPICS)


Hi Tim,

This definitely looks like what we are looking for, and I somehow missed this option. You mentioned a good point about one of the downsides of the IP-EP20x and the care that needs to be taken with field wiring. Would be nice to avoid this extra work altogether like with the Zebra module. Definitely will be considering this though I am curious about price now.

Thanks,
Andrew

On Nov 14, 2015 1:00 AM, "Mooney, Tim M." <mooney@aps.anl.gov<mailto:mooney@aps.anl.gov>> wrote:
Hi Andrew,

You might also want to look at the Zebra:

http://quantumdetectors.com/zebra/http://quantumdetectors.com/zebra/

Here's the user manual:

http://controls.diamond.ac.uk/downloads/support/zebra/1-13/documentation/TDI-CTRL-TNO-042-Zebra-Manual.pdf

I haven't used this, but I saw a presentation on it.  I don't remember what it costs.  I don't think it's less expensive than softGlue on a Hytec IP carrier (which I also haven't used, though I plan to).  Anyway, it looks like a very good product.  I think Zebra does a much more comprehensive job of handling field wiring than the IP-EP20x used by softGlue.

We're also looking at CompactROI and FlexRIO as solutions for things that softGlue can't do.  Very impressive stuff, I think, but I was dismayed at the price.

Tim Mooney (mooney@aps.anl.gov<mailto:mooney@aps.anl.gov>) (630)252-5417<tel:%28630%29252-5417>
Beamline Controls Group (www.aps.anl.gov<http://www.aps.anl.gov>)
Advanced Photon Source, Argonne National Lab

________________________________
From: tech-talk-bounces@aps.anl.gov<mailto:tech-talk-bounces@aps.anl.gov> [tech-talk-bounces@aps.anl.gov<mailto:tech-talk-bounces@aps.anl.gov>] on behalf of Andrew Gomella [andrewgomella@gmail.com<mailto:andrewgomella@gmail.com>]
Sent: Friday, November 13, 2015 7:20 PM
To: tech-talk@aps.anl.gov<mailto:tech-talk@aps.anl.gov> Talk

Subject: Alternatives to softGlue? (re-configurable TTL I/O line logic controllable from EPICS)

Hi Tech-talk,

Quick one-line summary of my entire email's question: Are there any solutions out there for re-configurable high-speed TTL digital circuit logic besides soft-glue(and associated hardware), that does not require VME or IndustryPack?

Basically we want a relatively low cost way to program logic for fast real-time synchronization/precise timing of TTL signals based on some counter inputs and other simple binary logic conditions, and is easy to modify/re-configure (leaving out hardware circuits as an option), yet also allows EPICS control/monitoring of the TTL signals. (EPICS control of the logic itself like what is found in softGlue is a huge plus, but not necessary)

If you want to hear my detailed thoughts/explanation on this (warning: this is new territory for me):

Mark Rivers recently introduced me to Soft Glue [1], a module I had not known about previously. Our lab (small table-top instrument control only) has pretty light requirements compared to most of the posts I see on tech-talk and correspondingly we have no VME crates and currently run our experiments entirely with "Soft IOCs" Given the asynchronous nature of our experiments this has worked well and kept our instrument control specific hardware/software costs minimal.

However we now are looking into more tightly synchronized experiments which would require real-time coordination of TTL I/O signals. Soft Glue seems like a very promising option, especially since Mark informed me that it was possible to buy PCI-e cards from Hytec [2] which allow IndustryPack modules to be run from a Linux desktop system. It also seems you can run several of the features similar to what is found in the popular SIS3820 [3].

I realized that there has been a recent boom in low cost FPGA boards devices being available to consumers and am curious about the feasibility of adapting softGlue to use one of them (or another solution similar to softGlue). (Amazon even sells a $23 development board [4] featuring the same FPGA as the one used in the IP-EP200, of course this isn't a fair comparison and who knows what kind of reliability this board has).  Many of these boards have usb/ethernet inputs, which would allow direct control of the device (ignoring the possibility of having the IOC run on the board as well). After various techtalk searches I found some discussion of implementing something similar to this on tech-talk[5], and several discussions of implementing IOC's entirely in FPGA code.

Our other main option would be using high speed PLC's with ladder logic (I received several suggestions from tech-talk regarding this route [6]) Ultimately, It seems PLC's are more commonly used for "slow control" and safety features, at least within the EPICS community. I would also rather avoid ladder logic altogether. The Galil RIO options suggested to us seemed like they could work well. Though I am not sure what kind of speed to expect from these devices.

I am also aware of CompactRIO, but would prefer to avoid it due to cost. Along the same lines SimuLink seems like it would give similar end result to SoftGlue, but lacks all of the EPICS interfacing, and also costs a decent amount for software alone. Though I did find some options from "United Electronics Industries" that can function as SimuLink targets and also have EPICS support. [7] Though I do not think the logic circuits can be dynamically programmed like how they are in softGlue. It seems ZedBoard + SimuLink[8] might be a powerful option, but would require a decent amount of work to get EPICS interfaced with the inputs/outputs.

Along these lines here is an interesting project where a Zedboard was incorporated with TANGO: http://zedboard.org/content/zedboard-tango-control-system-gateway

Any advice or thoughts on this topic are appreciated. Currently it seems softGlue + associated hardware is our best bet.

Thanks,
Andrew

[1]  http://www.aps.anl.gov/bcda/synApps/softGlue/softGlue.html
 also : https://www1.aps.anl.gov/files/download/Committees/InterCAT_Technical_Workgroup/2011/20110616mooney.pdf
[2]  http://www.hytec-electronics.co.uk/DataSheetHtml/6335.html
[3]  https://subversion.xray.aps.anl.gov/admin_bcdaext/softGlue_examples/
[4]  http://www.riorand.com/electronics/riorand-trade-ep2c5t144-altera-cyclone-ii-fpga-mini-development-board.html
[5]  http://www.aps.anl.gov/epics/tech-talk/2014/msg00261.php
[6]  http://www.aps.anl.gov/epics/tech-talk/2015/msg01365.php
[7]  http://www.ueidaq.com/ueinet-sim-upgd.html
[8]  http://www.mathworks.com/hardware-support/zynq.html