10-30-2013 10:17 AM
The ENOL appears in the log because our software is probing for multiple devices on the GPIB bus, and in this test case, only one is present. The problem, if any, will be at the end of the Spy log, not the beginning. And, it is the computer that hangs, not the device. I never have to reset or power-cycle the device, and there is no error indicator on the front panel.
Ben
10-31-2013 03:35 PM
Hi Ben
After going through the 3458A Multimeter User’s Guide I have looked through the calls that have been caused your software to hang.
http://cp.literature.agilent.com/litweb/pdf/03458-90014.pdf
Both the PRESET NORM command and the TRIG HOLD command suspend readings from the 3458A Multimeter. Is it possible that you are calling a command attempting to read the buffer after the PRESET NORM or TRIG HOLD commands have been executed? This would mean you are attempting to call data when data is not actually being sent from the 3458A Multimeter. I can see this being a problem especially if you do not have a timeout set on the read call.
10-31-2013 04:24 PM
Sam,
We do have timeouts set on read calls, and we do poll for interface ready (using ibrsp) before reading.
Keep in mind that the meter is not hanging on a read, the host computer (the Linux OS) is hanging after a write.
And, the NISPY log would indicate if there were a ibrd call with no response from the meter, wouldn't it?
And, also keep in mind that the same sequence of meter commands works correctly under the newer Linux OS (CentOS 5.4) with the NI GPIB 2.9.0 driver package.
Ben
11-04-2013 02:17 PM
Good afternoon Ben,
I just wanted to run a double check on the drivers you are using for your system. You listed in your first post that the driver you were using was NI-488.2 v2.5.4b1 which corresponds to our NI-488.2 2.5 - Beta 64 Bit – RedHat driver. I searched through the readme of this driver and it is only compatible with Red Hat Enterprise Linux WS 5.
After searching through the NI-488.2 drivers, I found a driver compatible with Red Hat Enterprise Linux WS 3 and WS 4. Although CentOS 4.8 is not a supported operating system in our Linux drivers, the driver below is a driver that would support Red Hat Enterprise Linux WS 4.8. Based upon the readme files I think the driver below would be the most compatible with your system compared to the NI-488.2 v2.5.4b1 you currently have installed.
NI-488.2 2.5 - 32 Bit - Linux
http://joule.ni.com/nidu/cds/view/p/id/329/lang/en
11-06-2013 11:18 AM
Sam,
Thanks for pointing this out. It does seem like we should be using a 32 bit driver, since this is a 32-bit OS.
So, I uninstalled the previous driver package, and installed the one that you suggested. Here are the installed versions, as reported by rpm -qa:
nipali-1.11.0-f0
ni4882i-2.5.1-f0
nispyi-2.4.0-f0
nikali-1.4.0-f0
Unfortunately, my test case stil had the system hang, at the usual place (I did not run nispy, but it appeared to happen at the usual place).
I'm going to thank you for your help on this problem, and suggest we close this thread. My management has agreed that we can drop support for this particular configuration (calibration with GPIB meter under Linux 4.8), so it's no longer a critical issue. My main reason for pursuing it this far is to try to determine the exact nature of the bug, and to determine whether it is perhaps also a latent bug in the NI-488.2 2.9 drivers, which we rely on heavily for use with Linux 5.4 and beyond.
Sincerely,
Ben White
11-07-2013 03:23 PM
Hi Ben,
Thank you for the update and for all of the insight you have given to Sam and me. I apologize that we could not come to a resolution for this issue, but it sounds like you have a proper plan moving forward. If, in the future, you get more insight in to this issue please post your findings so that it might help someone with a similar problem.
Regards,