10-28-2009 06:58 AM
Maybe that's a solution if the measuring frequency is always known. In our situation the frequency can vary in the range from 1 Hz to 10 kHz and dividing the input signal in this case isn’t always the best choice.
10-28-2009 08:34 AM
Then you should think about using another hardware, e.g. two X-Series cards, where each has four counters (new STC3 chip) and six DMA channels.
Another way would be to use an FPGA card.
Maybe you should contact your local NI office to get an offer.
Christian
12-08-2009 06:07 AM
Now we could solve our problem with following solution. We used the N-Samples mode with a 4 samples width buffer. In this mode the task will be stopped after all samples were captured. After the task has stopped we wait specified time until we restart the task to measure the next PWM signal. See also the attached file.
With this solution we could reduce the number of issued IRQ dramatically. So we were able to measure frequency up to 20 kHz on all counters in parallel on the NI-PCI6602 card. Maybe this solution helps someone else who has the same problem.