LV crash after ca. 11 hours, last saved value "NUL" - solution ? workaround ? (USB-cards 6210 und 6361 )

[Bernysberny]

Hi,

Unfortunately, I've got a really bad problem, which cannot easy be debugged

Program (LV 2011) runs on a Intel Atom PC with card 6210 und 6361 (both USB) , Win XP without any problems in over 11 hours, than suddenly the pc "freezes"- last saved value  "NUL" ( see attachment) . After restarting the PC, LV shows the warning, that it crashed the last time 

can anyone help me ?

thx !

greatings

Berny

[Mcdan]

Berny,

Without seeing your VI, I can't really suggest much...  However, I would recommend that you look for places where resources may be acquired but not released.  This would include tasks or other references being created but not cleared, files which opened but not closed, or arrays which are continuously appended to in a loop.  Have you watched memory usage as your VI runs?

Those are the first things that come to mind.  If you can attach the code which reproduces this issue, I may have more suggestions.

Hope that helps,

Dan

[Bernysberny]

Hi,

thank you for your answer. This is the "smallest" piece of code, where i got the malfunction. I already got the hint, that i should not use the express-vi.  I will test that an write again, could be that the permanent  "open" and "close" of the card is not the best way 🙂

Thanks !

Greatings

[Mcdan]

Berny,

I can't really see how your DAQmx task is configured, but looking at your code, I would probably approach the problem as shown in my attached example.  I do the following:

1) Open and setup the data file.

2) Set up a DAQmx Task, configured to run continuously.

3) Start the task running

4) Read two seconds worth of data

5) Append it to the existing file

6) Clear DAQmx task when done

7) Close file I opened.

Note that the DAQmx task is created and started only once, and the the file is opened and closed only once.

Hope that helps,

Dan

[ADagli]

Hi Berny,

do you already have contact with the NI Support-Hotline?

If it is so, can you please tell me the SRQ vie private message?

Abdülkerim

[Bernysberny]

Hi again,

i would like to ask again:

I tested quite a lot of things  - here is the result:

The DAQ is written direct with DAQmx ( see attachment)

Now i have a different "crash"  - labview is running, the digital ports of the card are working - just die anlalog in port is "froozen"  - if i open all vi's there ist "no execution" in the DAQmx read   vi ...

Stopping the program and start again "solves" the Problem

Can anyone help me ?

Workaround ?

Greatings

Berny

[Mcdan]

Berney,

A few more questions for you.

1) Do you see this behavior with both the 6210 and the 6361?

2) Does the DAQmxRead ever timeout (as specified it should be using the default value of 10 s)?

3) Is this issue still happening after 11 hours of run?  Does the amount of time passed until you see bad behavior change if you change either the number of channels or the sample rate?

4) How many channels are in your task?

5) Are there multiple code paths being executed in parallel with the DAQmxReadLoop?  Does the code attached exhibit this behavior when run by itself?

6) Have you seen any unexpected behavior with either CPU or memory usage prior to or during this observed hang?  Does 'cycle-time analog read' stay fairly constant throughout the course of execution?

In your explanation, I think I'm missing what you mean when you wrote, 'if I open all vi's there ist "no execution' in the DAQmx read vi...'  Does this mean that the DAQmx Read VI is not executing, or that execution is stopped inside the DAQmx Read VI?

I can try to reproduce similar behavior, however I'll need the answer to a few of these questions before I can do so.

Dan

[Bernysberny]

Hi Dan,

to your questions  -  because of the time delay it's not easy to test everything 🙂

1) Do you see this behavior with both the 6210 and the 6361?

no  - just on the 6210    (see bottom )

2) Does the DAQmxRead ever timeout (as specified it should be using the default value of 10 s)?

default values 10s

3) Is this issue still happening after 11 hours of run?  Does the amount of time passed until you see bad behavior change if you change either the number of channels or the sample rate?

its just approimately 11 hours , sometime a few days  - it random i think

4) How many channels are in your task?

16  ( all of the card are needed)

5) Are there multiple code paths being executed in parallel with the DAQmxReadLoop?  Does the code attached exhibit this behavior when run by itself?

i deleted all other loops  - runs since 1 day - its not clear if that is the / one reason

6) Have you seen any unexpected behavior with either CPU or memory usage prior to or during this observed hang? 

No, that was also my first idea

Does 'cycle-time analog read' stay fairly constant throughout the course of execution?

Yes

BUT i found a way to generate this error  very fast:

I opend the windows remote desktop connection (without a monitor on the PC), connected to the PC  and than i  unpluged the LAN-cable out of the PC  - the card gets directly frozen, when i reconnect the remotedektop  (after plug the LAN-cable in the PC) the cards are running again ... 

( i tested also a reset of the card  - it works  - the cards run fine after that ;   when i open and close the remote-desktop in the standard way  - the card work also fine...

Do you have any idea why this happens  - could be the same reason

Greatings

Berny

[CDreher99]

Hello Berny,

back to your attachments CC1 & CC2, is the sample clock rate 5000.00 or 5000,00?

The "sample per channel per read" can be round about 10% from your sample clock rate.

In this case 500 or 50000

DAQmx Timing and Sample Rates

http://digital.ni.com/public.nsf/allkb/3E3D74E26B8A5B83862575CA0053E4B5?OpenDocument

Hope that helps,

Christian

[Bernysberny1]

Hi,

thank you  - the number is 5000.00   - so 5000 samples/sec will be read - that works

greatings

Berny