PXI

cancel
Showing results for 
Search instead for 
Did you mean: 

20 microseconds lost writing data

I have a signal generator cranking out a 10 kHz signal and split twice to feed four inputs on a PXI.  The PXI builds and array of 25000 sample points at 250 Khz, and sends them to the host computer using tcp/ip.  The data is presented on four graphs and streamed with tdms functions to a file.  When I opened the file and examined the data, I found the synchronazation was great, all four signals matchted. The definition of the period of a 10 khz wave worked out with the number of sample points collected.ogether.  At each splice point in the data, multiples of 25000, there appears to be a discontinuity of .00002 seconds.  I examined the data around point 25000 and 50000, and they both appeared to be missing 20 microseconds, or about 4 sample points at 250khz.
 
I believe the problem lies on the host side.  I built an array of arrays on the PXI side for the sole purpose of examining splice points.  I didn't notice any.
 
Should I try to write to a binary file instead of using tdms?  Should I change the size of the array to match sectors on the hard drive, and if so how do I know how many numbers that equates to?
 
Some advice would greatly be appreciated.  I will need to collect data with this in a couple of months.
 
 
 
  
0 Kudos
Message 1 of 2
(2,830 Views)

Hi Jason,

Thank you for contacting National Instruments.  That seems to be an unusual problem you are running into.  Writing to a binary file instead of a TDMS file might be a worthwhile change.  The process of writing to a TDMS file requires a bit of conversion and therefore it might be faster to use write to binary.  What hardware devices are you using to acquire the data within your PXI system?  Is this system connected up through a switch on an open network or directly through a crossover cable? 

If there is a disconnect between the information on the host and the target, it might be related to the medium by which the data is transferred.  If it is an option to use the Real-Time FIFO's for your transfer method it should prevent any loss of information.  A good resource for establishing this type of data transfer can be found here.

I hope this helps!  Let me know if you are still experiencing these problems or there is anything else I can help with or clarify.  Have a great day!

Jason W.

National Instruments
Applications Engineer
0 Kudos
Message 2 of 2
(2,806 Views)