Experimental Physics and
Industrial Control System

Michael Davidsaver via Core-talk <core-talk at aps.anl.gov> · Mon, 9 Aug 2021 08:05:49 -0700

On 8/9/21 4:32 AM, Heinz Junkes via Core-talk wrote:
> Hi,
> unfortunately, this will be a somewhat longer mail. I hope you manage to read it until the end:
> 
> We have succeeded in an initiative of research and science in Germany to provide EPICS as a basic system:
> https://www.physics.nat.fau.eu/2021/07/05/fairmat-lifting-the-treasure-trove-of-materials-data
> 
> It is perfectly clear to us that we need to start this at an university in order to familiarize students (here of physics) with the concept of EPICS.
> 
> For a "kick-off" I visited the chair last week in Erlangen to familiarize them with EPICS and the other stuff like Phoebus, ArchiverAppliance, Bluesky. So the things that we use as a matter of course. We had prepared installation scripts and in Erlangen they provided me with a PC running Ubuntu. This also worked so far and we were able to set up an IOC and read out a measuring device (Keithley K2000) via serial interface using StreamDevice.
> ( By the way, they still use GPIB for almost all lab setups. I thought that no longer exists ;-) )
> 
> But then I found out that the people on site couldn't really follow me. Although they are quite strongly motivated.
> But unfortunately they have no experience at all with the Unix environment. They are also not familiar with editing configuration/startup text files. And also the many ways you can build an IOC now and then also the many operator possibilities just overwhelmed them.

This has been my experience as well when doing training involving the EPICS database.
To some extent this can be overcome by having a couple of assistants to help the one
or two students who will inevitably have some problem getting started.  Often an under
powered laptop struggling to run the mandated antivirus and disk encryption.  Of course,
practically speaking this necessitates a large class size, which is harder to arrange.
(I'm obligated to mention that Osprey DCS can help with this :) )

> I think it would be purposeful to address such situations in general.
> 
> The following idea:
> 
> Provide epics base with the usual support modules (async, StreamDevice) as a Windows installation package (MSI?) so that the typical Windows user only has to "double click”.

If you haven't looked into MSI before, I would encourage you to continue with this policy.
It's a bizarre world of XML with ten kinds of GUID for doing seemingly simply tasks.

> In it an example IOC (started with procserv) which is designed to "talk" to measuring devices. Provision of a low-cost/simple GPIB Ethernet gateway ( z.B. PROLOGIX GPIB-ETHERNET CONTROLLER). Installation package Windows for Phoebus, ArchiverAppliance, Bluesky.

If your goal is a full scale setup with services, I think a VM is probably the best way to go.
We've also toyed with the idea of providing each student a raspberry pi (or similar) preloaded,
and preferably with some simple I/O.

And while not a complete solution, anyone with a recent python environment can get a minimal
softioc pre-built (Linux, OSX, and Windows).

> $ python -m pip install epicscorelibs
> $ python -m epicscorelibs.ioc -h
> usage: ioc.py [-h] [-m MACROS] [-d DATABASE]
> ...

> Maybe someone has already done something in this direction and the whole thing already exists.
> 
> I think our previous attempts with ready-made UNIX VMs or installation scripts do not address the typical Windows user and prevent the spread of
> EPICS in teaching or even controlling smaller experiments without existing IT infrastructure and support.
> 
> I would be happy if we could discuss this at one of our next core-developer meetings.
> 
> Viele Grüße
> Heinz Junkes
> --
> Experience directly varies with equipment ruined.
> 
> 
> 

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System