A bit further to Ulrich's point... one area where I had some problems
using serial/gpib converters in the old days, and I think that at least
early versions of the Prologix USB unit suffered from, is talking to
multiple instruments off the same GPIB controller. I never found a
satisfactory way to accomplish that with anything other than a "real"
GPIB card. (But Prologix may have addressed this by now.)
I think GPIB applications break down into a couple of categories: (a)
simple trigger-then-read from a single instrument, which is what most of
us do most of the time, and (b) multi-instrument control. The second
case is a lot harder, particularly if you need to use SRQ and other
out-of-band commands.
Of course, an alternative is to assign one USB or LAN GPIB unit to each
instrument, but that gets expensive.
John
----
Ulrich Bangert wrote:
Gents,
as you know I have tried in my utility EZGPIB to supply the user with high
level functions for data communications which keep him away from the low
level tasks of direct communication with the interface. Since EZGPIB not
only supports the Prologix products (USB & LAN) but also GPIB32.DLL based
products (i.e. the complete line of NI GPIB interfaces) a script written in
EZGPIB is independent from a certain manufacturer to a high degree.
That is the point that I wanted to add to the current discussion.
Please allow me to note that this independence has its limits that I just
happened to come across. I had tried to run a script for my SR620 counter
which was set to a gate time of 100 s. The script was a quick and dirty one
and set the timeout for GPIB communications to 120 s so that there was
enough time for the counter to answer, even if the data was read immediatly
after tze start of a new measurement. This script would not run ok on my
Prologix GPIB-ETHERNET interface although I was sure that is was ok when I
wrote it.
It took me some time to realize that I was fooling myself: I had written the
script with my NI interface which's DLL allows for some decades of magnitude
of timeouts while the Prologix provide a maximum timeout of 4000 ms. That
means: Some things cannot translated 1:1 between different manufactures. I
changed the script so that the conunter's status byte was read once a second
and the "data valid bit" is detected in the status byte. Then the counter's
result is read. This of course run ok (also on NI hardware). I will include
a warning in the manual of EZGPIB.
_______________________________________________
time-nuts mailing list -- [email protected]
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
