Intermittent VI_BUS_ERROR when commanding a Boonton Power Meter

[jmni]

I have a problem trying to communicate with a Boonton 4230 Power Meter. For the most part, my program goes along quite happily writing commands to the instrument and reading data back from it, but occasionally the interface times out with a "Bus error occurred during transfer" (VI_ERROR_BERR). This could be on a write or a read, a "listen" command (e.g. Change the channel ), a "listen-action" command (e.g. Take a measurement ), in fact any command I might send to the instrument. A worst case scenario is where the problem occurs during a measurement read when the Boonton goes into Talk Mode in order to resturn a measurement with the result that the bus hangs up.

Is the VI_ERROR_BERR the Visa representation of the GPIB EBUS error?

Everything I've read on this error suggests that if it isn't a problem with the specific instrument then it is a timing issue and that the timeout period isn't long enough for the driver to send command bytes - a hand-shaking problem.

 - The problem has been seen on 2 different Boonton 4230's

 -  I've played around with the VI_ATTR_TMO_VALUE changing it from our default value of 10 seconds up to 30 seconds without removing the problem.

Is there a lower-level, GPIB timing value that needs to be changed or is VI_ATTR_TMO_VALUE my only option? 

I have a temporary work-around that will get the interface controller to send an Interface Clear (IFC) command when the condition is detected. Meanwhile, the Boonton has a bus to itself while I search for a more permanent fix.

System Details:

 - Sun V440 (sparc) computer with Solaris v2.9.

 - NI-VISA for 32-bit Solaris - v3.1

 - NI-488.2 Software for Solaris 2.x PCI-GPIB  v2.3 

[Emilie S.]

Hi jmni:

Yes the VI_ERROR_BERR occurs when a GPIB EBUS error occurs.  This will happen when there are no instruments at all connected to the GPIB board.  A different error (no listeners error) occurs when there is just some incorrect addressing.  That being said, the first thing I would do would be to check the cabling.  Are your cable lengths within spec?  Do you have other cables around you can switch out with?  If the cables are all good, it would lead me to believe there is some issue with the instrument.  The specified timeout should not have an effect on this error as again, it will only occur when there is nothing connected to the GPIB board.

Regards,

Emilie S.

[jmni]

Thanks for the feedback, Emilie.

I've seen the problem with two different GPIB buses with 2 different Boonton power meters. In both cases the cables have been 2 meters or less. I've also been in touch with Boonton themselves but they have been unable to shed any light on it.

jmni

[Emilie S.]

Hi jmni:

Is Boonton able to reproduce the problem?  Do you have any other instruments you can try with this GPIB interface?  Any other GPIB interfaces you can try?

Emilie

[jmni]

Emilie,

The original configuration had a number of different GPIB instruments ( DMM, Oscilloscope, Waveform Generator, etc) on the same bus as the Boonton and none of them experienced the same problems. The Boonton was isolated onto its own GPIB bus so that it didn't interfere with the others and we still saw the problem. I have a test setup dedicated to another.Boonton and still see the problem. It's intermittent and it can happen on any command type. I've seen it happen within the first 20 commands sent to the Boonton and I've also sent as many as 18,000 commands to it before it occurred. There doesn't appear to be any way of regaining communication with the instrument without an Interface Clear.

Boonton have tried to be helpful but they really haven't seen this problem before. I've had some success with my work-around but it is by no means ideal.

  jmni

[Emilie S.]

Hi jmni:

Unfortunately I do not have a Boonton device here to test.  It seems that since you can communicate to other instruments that the GPIB interface and the cable are both functioning properly.  This leaves only the instrument to be not working properly.  Has Boonton been able to test with the same NI GPIB interface that you have?

Emilie