Experimental Physics and
Industrial Control System

"Rod B. Nussbaumer" <[email protected]> · Tue, 20 Mar 2001 08:24:20 -0800

Eric Norum wrote:
> 
> Do you use non-EPICS applications to communicate with GPIB devices
> through the Agilent E2050A GPIB/LAN adapter from Solaris or Linux
> hosts?  If so, how do you do this?  I have not been able to find any
> libraries that provide this access.  National Instruments provides a
> Linux shared library to provide C-callable subroutines to communicate
> with devices through their GPIB-ENET adapter, but Agilent provides
> nothing comparable.
> 
> This causes a problem.  If I want to use non-EPICS applications (Tcl/Tk,
> standalone C programs, etc.) on my Linux/Solaris machines I have to use
> the GPIB-ENET adapter.  If I want to use EPICS applications I have to
> use an E2050A.
> 
> I could use the excellent EPICS driver from Benjamin Franksen as a
> starting place for implementing a library matching the National
> Instruments API (ibfind, ibdev, ibrd, ...) but I'd like to avoid that
> much work if I can!

I have been working on device support for the NI GPIB-Enet.  It will
be based substantially on the C library (such as it is) that we purchased
from NI.  Optimistically, it will be ready to deploy when our ISAC
accelerator starts up agian in about a month.  Right now, it is
at a very rudimentary stage of completion.  It will not include support
for such things as SRQ servicing, in the short term.  There is some 
question about licensing/distribution, due to the use of the NI code.

We have only been using the 10 Mbps device, and it is unknown whether
the code will apply to their newer 100 Mbps units.

Up until now, I have been using the devices from Labview on Solaris,
and relaying information to EPICS using Labview CIN's, calling the
Ezca library functions.  This has not been a very satisfactory means
of deploying the GPIB-Enet, and I am anxious to get the EPICS support
completed.  I have some serious questions about the robustness of
these devices, and we have had some discussions here about other
means of implementing GPIB instrument support; in particular, using
PCI GPIB controllers in desktop PC's, likely running Linux, and
some form of channel access.

You may like to discuss this further, offline.

	---   rod.

-- 

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 Rod Nussbaumer, Programmer/Technologist    E-mail: [email protected]  
 TRIUMF --- University of British Columbia,  Phone: (604)222-7449
 Vancouver, BC, Canada.                        FAX: (604)222-7307
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System