NI-6534; Biterrors in Inputstream

[at49377]

Hello!
I have the following problem.
I use the NI-6534-card, the NI-DAQ-version 6.4.3 and C++ Visual Studio 6.0.
I configured two groups, one for output and one for input. I use continuous input and continuous output. In C++ I use the functions DIG_DB_HalfReady and DIG_DB_Transfer to receive the Data from the onboard memory of the card, for the output operation I use the onboardloop of the NI-card. All of this work very well.
I send an outputpattern to the card wich is send out cyclic by the onboardloop. This outputpattern is loaded into a shift register where the data will be read and then set to different values. After that the data is sent to the inputgroup of the NI-card. My halfbuffer on the card is 10 times bigger than the inp
utpattern, so I collect 10 patterns before I start the tranfer to the PC-memory. I also tried different factors for the halfbuffer.
I know exactly what data I expect in the input-halfbuffer. When I compare the expected data to the received data from the halfbuffer, I notice that there are one to four bit-errors one behind the other. But these errors are only in different positions in the front of the halfbuffer never in the middle or in the back of it. So, the most inputpatterns are OK but in the first one or two patterns in the halfbuffer are the errors.My problem is that I don't know the cause of these bit-errors and I don't have any ideas to solve this problem because the configuration of the card and the use of the NI-DAQ-functions are correct. I hope someome can help me.
Thank you...

Andreas

[rgharrison]

Hi Andreas,

There are a couple things I can suggest.
1) You can try upgrading to NI-DAQ 6.9.3 (ni.com/downloads)
2) It is either the 6534 hardware, NI-DAQ driver or your circuit, since option 1 takes care of the driver (I have it an it works well with pattern I/O and the 6534 we just need to worry about the other two. What I would do is generate pattern output out of 1 group and acquire the pattern input on the other group. Simply wire the two together so that the card is talking with itself. If you still see the bit errors when comparing the two buffers (should be exactly the same) then you know there must be something wrong with the card. There are NI examples that perform pattern I/O with the 6534 on the website. If the pattern seems to be fine, th
en it might be your circuit that you are connecting to. If that is the case, it might be difficult to verify your circuit.

It could also be the clock you are using for input. If there is a slight offset in the clock by the time the 6534 read the data, then it might read the previous bit or the next bit in its sequence. Take a look at all the bit errors and see if they are the same as the previous or next bit. If this is the case you will have to use a more stable clock source at the input or use a shorter wire for the clock.

It is unusual that the 6534 sees bit errors, and in the same relative location. I hope these suggestions help. Keep me posted on your testing.

Ron

[at49377]

Hello Ron,

I only want to say that everything is Ok. There was an error in my program, it was not the hardware or the driver.
Thank you for your help!

Andreas