Re: serial VISA not sending output until closed

[tmh]

Just for the record - since this question came up for me recently -
the official answer is that VISA 3.0.1 does support LabVIEW 5.1. In
fact it should support any older version of LabVIEW, but that may not
be tested/officially supported. See a
href=http://exchange.ni.com/servlet/ProcessRequest?RHIVEID=101RPAGEID=135HOID=5065000800E9C6USEARCHCONTEXT_CATEGORY_0=_201_%24_13_USEARCHCONTEXT_CATEGORY_S=0UCATEGORY_0=_201_%24_13_UCATEGORY_S=0;here/a...

Re: serial VISA not sending output until closed

[Scot]

So far I have found nothing that helps.  I have tried all of the
wiring configurations given, and still have the same result.  The data
is sent to the buffer as it is written according to NI spy, but
nothing actually exits the port until after VISA is closed, even with
flushing the buffers.  Is VISA waiting for some sort of command? Is
there a bug in my software?

I could close the VISA session, but I would still want an accurate
count of how many pulses have been sent at a given time during my
scan.

Thanks

Re: serial VISA not sending output until closed

[Scot]

I upgraded my VISA to 2.6, which seems to work fine with Labview 5.1
(I couldn't find version 2.5 for Windows).

I still have the same problem.  If I use serial VISA, I can get
well-timed pulses at up to 62kHz out of the parallel port, connecting
it as given in the a
href=http://zone.ni.com/devzone/conceptd.nsf/webmain/72C6FC6CE4AD4D1386256B1800794596?opendocumentnode=DZ52058_US;Using
the Parallel Port in LabVIEW/a tutorial and using their vi.  NI-Spy
shows that the commands are sent as they are generated, but the pulses
do not actually go out the port until after the VISA session closes.
I have tried flushing the buffer after every write.  As a result, I
can get the requested pulses, but with no feedback as to when they are
complete.

If I open VISA, write a number of steps to VISA, then close again, I
get a 50ms gap between segments that I write.  I want the output to
appear seamless.

In the past I have used the outport.vi, but to get evenly-spaced
steps, I have to use the wait-ms.vi  This limits my pulse frequency to
500 Hz.  To go faster, I have used a for loop with several hundred
iterations -- this leaves an erratic pulse train and is difficult to
control and ties up the processor. 500 steps/second is also too slow
when you consider that a typical scan requires 120,000 steps, each
direction.

So if anyone knows how I can get a regular pulse train out of the
parallel port (or is their another port that would work -- not the
serial though) and know when it completes so I can continue my
program, I would appreciate some help.

Thanks,
Scot