NI-5102 device not found on some boot cycles

[Digitizer99]

I've got a PCI NI-5102 installed in a 4U rack-mounted PC with other PCI cards installed (two PCI-e  x1 cards and 1 other PCI-32 it) running Windows XP.  On about 10% of boot cycles the MAX software is unable to locate the 5102.  On these faulty boot cycles the Scope Soft Front Panel application is also unable to detect the 5102 and will only allow me into simulation mode.  Re-booting resolves the issue. 

Using previous posts here as direction, I've set the NI Configuration Manager service to 'Restart the Service' on failure.  This did not help. 

When the fault occurs all the NI services that were present when working are still present and report no issue.  During the fault condition the device number (always 1), memory range, and interrupt request number direct memory access numbers are the same as the working mode.

Any additional ideas from this forum is appreciated.

[lab_boy]

Hey

Try upgrading the scope driver to the most current one. That might help

Good luck

-lab

[dansch]

Hi Digitizer99,

I was curious to know if you upgraded your driver version yet and if the problem is still occurring. I would like to know what version(s) of the NI-SCOPE driver you have tried. Also, is this a new system, and, if not, did this problem just recently begin to happen? Has something changed in the system at all prior to this issue occurring? Does the same problem happen with that card in another system, or does a different card in this system still cause the problem to be seen?

When you reboot and the problem gets corrected, does Windows have you reinstall the card or does it just show up in MAX as if nothing happened? When MAX does not recognize the card, is it still being detected in Windows Device Manager? Does uninstalling the card and then rescanning for changes detect and install the card in Device Manager? Does it then work again in MAX?

Finally, are you actually unable to see the card in MAX or is it there and you are just getting an error when trying to access it? If so, what is the full description of the error you are getting? I am curious of this since you mentioned the information in MAX is the same in a fault condition as it is when it is working. Thus, I am confused if it is still showing up, but maybe with a red 'X' instead of the green icon that is normally present.

Thanks, and I look forward to hearing back soon.

[Digitizer99]

Hi Dan,

I have not upgraded my NI-SCOPE driver yet (currently running 3.3.0f0).  I have 13 teststands that are all configured identically as described in the original post.   They all exhibit this problem - which leads me to believe the problem is systematic and not a corrupt driver on a given teststand.  The failure rates do seem to vary station-to-station and users with administrative accounts seem (???) to have a lower failure rate than those without but I haven't got quantitative numbers to back that up.

This is not a new problem.  Its been there since we bought the PCs and configured them as described in 2008.  

After reboot, the card shows up in MAX as if nothing happend.  No reinstall required.  I have not caught an instance of the failure to look at the Windows device manager but will on next occurance.  When the failure occurs, the NI-5102 still shows up under Devices and Interfaces ->Traditional NI-DAQ (Legacy) Devices.  When I select it, the test panel icon is grayed out and thus the Soft Panel application only allows simulation mode to be run.  I don't see any errors listed.  Where should I look for the error?

I do not see the red 'X' when the fault occurs only the green icon.

I'm running version MAX version 4.5.0f0.

Al F.

[Digitizer99]

Hi Dan,

I caught the error today and looked at the windows device manager.  During the failure the device manager still showed a 'NI Data Acquisition Devices' with a 'NI 5102 for PCI' under it.  I was able to bring the test panel up in MAX this time and it showed a -10401 error code.  Perhaps you can elaborate on what that means.  I then tried a 'Scan for hardware changes' in the device manager and that allowed the NI-5102 to be found and respond nominally.  This is a improvement over rebooting to clear the fault but would still like to get to root-cause if possible.  Incidently, users without administrative privs. can not 'Scan for hardware changes'.

Thanks for your help thus far.

Al F.

[dansch]

Hi Al,

Thanks for all the info; that is quite interesting behavior, especially that you can scan for changes and Windows will then detect and install the board. The -10401 is a Traditional-DAQ error indicating that software is trying to access a resource that is not present on the system, which makes sense due to your system not recognizing the device at these times. 

At this point, I would be interested to know if you can provide some system information for this rackmount PC. You can get this information by running (CTRL+R) "msinfo32". This file will give us an idea of your system (it may take some time to save) and potentially shed some light onto what is happening.

Regards,

[Digitizer99]

Hi Dan,

I'm unable to get you the system info due to company security practices.   Sorry.

I did find other behavior that you may find interesting.

I have a power cycle now where when a non-administrative account is used the NI5102 is not found but when a administrative account is used it is found.  I've tracked it further to a system event log entry where when the administrative account launches MAX and expands the 'Traditional NI-DAQ (Legacy) Devices' a "The Nidaq32K service was successfully sent a start control." event is recorded.  No such log entry occurs when the non-administrative account is used to attempt the same activity and thus the NI-5102 is not found. 

Can you explain further the mechanism for starting these services?  Can they be started (or 'sent a start control') manually?  I do not see this service listed on the 'Services' listing under 'Administrative Tools'.  My guess is that there is some activity that is prohibited without admin privs.. 

I've found the registry entry for the Nidaq32K service and is 'DependOnService' Nipalk, nistck, nimdsk, nimxdfk, nimdgk, and niorbk.  Could it be that one of these dependent services is not started and thus the Nidaq32k service won't start?  Other Nidaq32K service registry entries are as follows:

ErrorControl = 1

Start = 2

Type = 1

Al F.

[dansch]

Hi Al,

That is some interesting information. Have you determined that the issue always occurs on a non-administrator account as opposed to an administrator account? If this is the case, I do know that some Traditional-DAQ and other legacy applications have problems when not run as an administrator. There are a couple of KnowledgeBase articles which give details on this issue.

KnowledgeBase 4S79F0XU: Why Can't I Open Legacy Applications on a Non-Administrator Account?

KnowledgeBase 3CQGAFHS: How Do I Give Non-Administrators Full Control Over National Instruments Prod...

My suggestion would be to try the steps in the second KB linked above to see if it helps to remedy the problem. Of course, if the problem is intermittent even when logged in as an administrator, we may yet still have to do more research. Please confirm this behavior (whether it is truly random on both accounts) and results of following the steps mentioned in the KB. I look forward to hearing back soon, cheers!