MATLAB Central Newsreader - Data Acquisition

Data Acquisition

Ian Nauhaus — Tue, 15 Jul 2008 00:10:21 -0400

I am trying to read fairly large chunks of analog data from
my National Instruments board (PCI 6110). I look at the
'MaxBytes' using the 'daqmem' function, and I have plenty of
space for what I am trying to acquire (as I define in
SamplesPerTrigger). But when I do start(ai), it gives me in
error saying that its too big.

Furthermore, if I set the SamplesPerTrigger to 'inf', it
will actually accumulate bytes to the expected value in
'MaxBytes' !!??

Summary: Why is it that when I try to preallocate a specific
number of samples (using SamplesPerTrigger), which
corresponds to a number much less than MaxBytes, that I
can't trigger the acquisition? I've also played with
BufferingConfig, and that doesn't seem to help.

Re: Data Acquisition

Sherryl Radbil — Tue, 15 Jul 2008 12:07:02 -0400

Post the minimum amount of code needed to see the problem
along with the actual error messages.
Sherryl

Re: Data Acquisition

Vijay — Fri, 25 Jul 2008 18:26:03 -0400

"Sherryl Radbil" <sherryl.radbil.dontspamme@mathworks.com>
wrote in message <g5i3t6$at3$1@fred.mathworks.com>...
> Post the minimum amount of code needed to see the problem
> along with the actual error messages.
> Sherryl

Hi Sherryl (& Ian)
This issue was addressed in service request 1-5REK26
which I submitted earlier this year (containing sample
code). It might make sense as either an external bug report,
or as an addition to the 'vendor limtiations' documentation.

The issue was that DAQmx now imposes a limit on finite
acquisition lengths (below 32Msamples), which did not exist
with their prior Trad NI-DAQ driver. It certainly applies to
the S-series boards. I can't recall for sure, but I don't
think it appies to E- or M-series boards.

Regards
Vijay

Re: Data Acquisition

Ian Nauhaus — Fri, 25 Jul 2008 20:11:03 -0400

Thanks Vijay.
I had actually experienced this problem with the traditional
NI driver as well. Not so for you?
Another thing, after playing with it some more, I realized
that if I go above a certain finite number in
SamplesPerTrigger that it will let me acquire. i.e. its not
just limited to low values and INF, but rather there is a
specific finite window in which it won't let you acquire.
--
Ian

"Vijay " <iyerv@janelia.hhmi.org> wrote in message
<g6d5rr$iqu$1@fred.mathworks.com>...
> "Sherryl Radbil" <sherryl.radbil.dontspamme@mathworks.com>
> wrote in message <g5i3t6$at3$1@fred.mathworks.com>...
> > Post the minimum amount of code needed to see the problem
> > along with the actual error messages.
> > Sherryl
>
> Hi Sherryl (& Ian)
> This issue was addressed in service request 1-5REK26
> which I submitted earlier this year (containing sample
> code). It might make sense as either an external bug report,
> or as an addition to the 'vendor limtiations' documentation.
>
> The issue was that DAQmx now imposes a limit on finite
> acquisition lengths (below 32Msamples), which did not exist
> with their prior Trad NI-DAQ driver. It certainly applies to
> the S-series boards. I can't recall for sure, but I don't
> think it appies to E- or M-series boards.
>
> Regards
> Vijay
>