My NI-6031E card acts as if the multiplexer is stuck.

Question

Dear Folks,
My NI-6031E 64 channel card acts as if the multiplexer is sticking. I see the same voltages on little groups of 2-3 channels. All the channels are connected to inputs, and there are no overvoltages. The output impedence of the sources should be low, because its the output of an op amp from some optical sensor boxes we build in-house. So all the things listed in the knowledgebase as possible causes are eliminated. I even tried replacing the cable, based on a suggestion from someone on info-labview, but that had no effect. I also ran the on-line E-series diagnostic and all checked out fine.
Now here is the kicker. These same exact systems, same sensor hardware, same cards, have run for the last 5 years in Macs with never a sign of this problem. Now we decide to migrate to PC's running WinXP, and we run into this very large snag. Any helpful ideas out there?
Thanks,
Alvin.

Answer 1

AWMoore,
I have an idea about what might be happening.
When you swtiched from Mac machines to PC machines, you changed the DAQ
driver.  If you currently are using NIDAQmx 7.5,  your
interchannel delay will be set to a minimum value.  In previous
versions of the DAQ driver the interchannel delay was set to round
robin by default which meant the interchannel delay was always the
maximum value.  The decrease in interchannel delay might not be
giving your channels enough settling time and thus causing your problem.
There are a few easy ways to test this
1.  Change the interchannel delay (convert clock rate) using property nodes, so that it is longer.
2.  Remove every other sensor and short the input to the DAQ card it was using (the problem should go away).
To calrify,  interchannel delay is the amount of time between each channel's sample.
For example (I am pretty sure you know what interchannel delay is, but
this is for other people who might be reading the posting.):
If you are using 10 analog inputs and are sampling at 10 Hz,  a
sample from each channel will be taken every 100 ms.  Since the
channels are multiplexed to the same ADC, the samples cannot be taken
at the exact same time.  So, we have to take 10 samples one after
another within the 100 ms.  In the past we would spread the 10
samples out across the entire 100 ms.  This would result in a
sample taken from a channel every 10 ms.  Now, we take the samples
as fast as possible.  This means that at the beginning of the 100
ms period we take 10 samples with a delay of something like 5 us
between each channel.  After the 10 samples are taken (this would
take 5 us times 10 samples = 50 us) the DAQ board does nothing for the
rest of the 100 ms period (99.95 ms would be left over) and then would
repleat the process for the next 100 ms period.
The new setup for the interchannel delay means there is less tolerance
for a sensor's impeadance, but it also means that the samples are taken
at much closer to the same time as one another.  Like I said
above, this is just the default setting and it can be changed within
LabVIEW using property nodes (DAQmx Timing Property Node: more->AI
Convert->Rate).
This is just a shot in the dark,  but it was what came to my mind
as a possible cause for your problem.  I realize you said your
sensors are low impeadance, but they may not be low enough.  It is
worth a try.
Lorne Hengst
Application Engineer
National Instruments