Keithley 2000 DMM AC Read Overflow

[StevePenny]

I’ve been having problems controlling a Keithley 2000 DMM during AC volts measurement using IVI. When reading AC Volts the DMM will intermittently sense an ‘overflow’ and return the value NaN. I’m using IVI and TestStand 3.0 but get the same problem with IVI and LabWindows 7.0.

The set-up is as follows:
Device: Keithley 2000 DMM, GPIB device number 6
Device interface: GPIB, DMM set-up in MAX as GIPB0::6::INSTR
IVI Driver version: 3.02, downloaded from the NI web site
IVI Compliance Package version: 2.1
NI-488.2 version: 1.70
NI-DAQ version: 7.1.0f1
MAX version: 3.0.2.3005

For test purposes I’ve connected the DMM front panel input high and low and set the inputs mode on the Keithley to front, this should mean that I get a very low Voltage reading. I’ve then run the LabWindows/CVI IVI control sample code provided with the IVI driver. At the initialisation panel I’ve entered the logical name. At the configuration panel I’ve selected Volt AC measurement but left everything else as default. The second configuration panel I’ve left as default. At the Read panel I’ve left all setting as default. Pressing the read button up to about 20 times with about a second gap in between each press will eventually lead to an ‘overflow’ reading.

Any help/explantion will be appreciated

Steve

[gagi1]

Steve,

In the Read function, if an over-range condition occurs, the Reading parameter contains an IEEE defined NaN (Not a Number) value and the function returns a warning: IVIDMM_WARN_OVER_RANGE (0x3FFA2001).

Do you use a random signal or a well defined AC signal? If the measurement value is not within the configured range, you will get the OverRange warning i.e NaN value back.

Regards,
Vesna

[StevePenny]

Vesna

Since I posted the first message I've been able to set-up a signal generator to generate 9 Volts at a frequency of 5kHz and have connected this to the DMM front panel high and low inputs.
This set-up generates a nice clean signal but AC Volt DMM reads still overflow occasionally when using the sample DMM IVI driver application provided by NI.

I've attached a zip file that contains three files that'll give more info on the DMM set-up and read:

"IVI setup.doc" contains screen captures of the DMM IVI driver application, the first three show how the DMM is configured prior to a read.
The last two screen captures are a good read and a read when the DMM shows "OV.RFLW".

"Config Capture.txt" contains the VISA calls to the DMM that occur when the three configuration screens are executed.

"Overflow Capture.txt" contains the VISA calls to the DMM when the read button is pressed.
For each press of the read button 67 VISA calls are logged. The first 5 button presses (first 335 calls) produced good reads but the last button press (last 67 calls) returns a Nan (OV.RFLW).

[StevePenny]

Here are the results of my investigation into this problem, it’s two fold:

1. Keithley have informed me that their engineers in the US have just seen the problem I’ve been seeing and are working on a firmware update. Currently here is no date for the release of the updated firmware but I’ll be informed when this it’s ready.
Note: I’ve found that the read AC Volts overflow problem only occurs when the read trigger delay is set to zero seconds.

2. The TestStand DMM IVI configuration step always sends a trigger delay of zero seconds regardless of the trigger delay set.
In my TestStand sequence the DMM IVI configuration sets the read trigger delay to 3 seconds but as noted above this is sent a zero seconds thus invoking the Keithley DMM AC Volts read overflow problem.
See attached VISA call log captured by NI Spy. The first 22 call are made when the DMM IVI configuration step is called form TestStand the next 22 calls are made by the IVI DMM sample code configuration provided by National Instruments when DMM configuration is run thorough LabWindows. Note call 12 (IVI DMM TestStand) has the trigger delay set to zero seconds (should be 3 seconds) and call 34 (IVI DMM LabWindows) has a delay of three seconds.

I’ve found a workaround as follows:
After each call to the TestStand DMM IVI configuration step an additional step that calls the TestStand IVI Tools step can be added. The IVI Tools step can be set-up to specifically configure the DMM Trigger Delay to 3 seconds. I’ve proven that this individual call does in fact send the correct trigger delay command to the DMM thus alleviating the AC Volts read overflow problem.

Can NI confirm that this problem does exist with the DMM IVI step and when they plan to fix it. I’ve attached the TestStand sequence file that I’ve been using to investigate the problem.

Thanks

Steve

[SachaE]

Hi Steve,
I've verified this behaviour against TestStand 3.1 and the latest IVI installs.
I've raised the issue with our developers and am waiting on their reply.
I'll post back here as soon as I get something.

Thanks
Sacha Emery
National Instruments (UK)

[AllenP]

Hi Steve,

We have traced the problem down to the TestStand IVI Step Type code. Please fill out a service request at http://sine.ni.com/apps/we/niae_asc.main and include your contact information. We will have to send you an update to the step type in order to fix this problem. Please include my name in your email so it can get directed into the proper place.

Allen Piscitello
TestStand R&D

[SachaE]

Hi,
a new DLL has been generated.
Please do the following :
Shut down TestStand
Start->run
then type in (depending on your directory structure)
regsvr32 -u "C:\Program Files\National Instruments\TestStand 3.1\Components\NI\StepTypes\IVISteps\TsIviStep.dll"
rename the
"C:\Program Files\National Instruments\TestStand 3.1\Components\NI\StepTypes\IVISteps\TsIviStep.dll" to something like .dlx (just so you don't lose it)
Then copy over the attached dll to that directory.
Then start->run
regsvr32 "C:\Program Files\National Instruments\TestStand 3.1\Components\NI\StepTypes\IVISteps\TsIviStep.dll"

Now try your sequence and monitor in NI-spy. I've checked it with simulating the keithley 2000, and my NI-spy now shows the trigger delay correctly.

Thanks

Sacha Emery

[StevePenny]

Thanks for the updated DLL.
I’ve done as instructed to install the DLL, I’m using TestStand 3.0 so the directory for TestStand is “c:\Program Files\National Instruments\TestStand 3.0\Components\NI\StepTypes\IVISteps\TsIviStep.dll” not “c:\Program Files\National Instruments\TestStand 3.1\Components\NI\StepTypes\IVISteps\TsIviStep.dll”
Running the TestStand test sequence now reads back AC Volts measurements without any overflows from the DMM, i.e. the configuration step sends a read trigger delay of 3 seconds as required. I’ve also checked this with NI Spy.

I do have a problem though, if I now try to edit the IVI DMM configuration step a message is display and TestStand runs through an installation procedure eventually prompting for the TestStand installation CD. This installation can be cancelled and the IVI DMM configuration step is then display but it doesn’t look correct, drop-down lists in grey rather than white, etc. I’ve attached a screen capture of both the TestStand installation prompt and the IVI DMM configuration step display. Is this happening because I’m using TestStand 3.0 not TestStand 3.1?

Thanks

Steve

[SachaE]

Hi Steve,
I don't see this problem under 3.1, however I need to setup a machine to test the 3.0 since it seems to be getting confused with the fact I've got two installs with the same dll registered in two different places.
Please bear with me.

Thanks
Sacha.

[SachaE]

Hi Steve,
I've checked this against a single install of TestStand 3.0, and it comes up grey as per you attachment, however it doesn't ask for the install cd on my machines (tried two).
Under TestStand 3.1, it's fine, and you have access to all the fields.

I've passed the information onto our developers, however since it's an old version, I can't make any promises of a fix, since there's the workaround youy came up with, however they will look into it.
If we do get a fix, then I'll post it back here.

Sorry.

Thanks
Sacha Emery
National Instruments (UK)