Andrea C

It seems Windows has not properly associated the DAQ as a USBTMC device.
Until that is resolved, all the NI VISA based communications will not succeed.

In Device Manager, can you have it update the driver for the entry?
Or delete that entry and scan for changes.
You need Windows to associate the detected USB device as USB Test and Measurement device.

Sometimes, if a driver is not present when the plug-n-play device is first connected to Windows, Windows will make generic other device association.  With driver installed, need Windows to take a fresh look and make proper associations.

Your description of connections seems correct.

A multimeter does not typically allow adjustable source current for the ohms measure function.

To achieve what you describe, you need a current source and then use the dmm to measure voltage to have insight about the R.
A SourceMeter such as 2450 can be the current source.  It might also serve as the voltage measure, but check the specs against what you need.

The 2002 is 8.5 digit instrument whereas the 2450 is 6.5 digit.

Simply reboot the instrument will apply the factory default settings.
Also sending *RST command will restore default settings.

If no errors reported during power on, there are additional self-tests you can run from the menus.  Check the reference manual.

As for the negative number:  how big and what measure range is in use?  By how much does it exceed allowed error for the range?  Is best to have no test leads when checking the current measure offset error (open ckt).

If the error is just a little past expected values for the specifications, calibration might fix it up for you.
But if the error is very large, it may indicate a malfunction of some sort and require service.

With only two SMU channels, you will have to move cables for each step of the VDP measurement.
But for each step, one SMU forces current and other SMU forces 0A and measures V.

Is easier if you have 4 SMU channels, one for each contact point in the VDP test structure.
As needed for each step, each SMU can be current source, LO return or voltage measure, so no need to move cables.

KickStart will require a gpib connection to 2401.
The PC gpib interface should be from Keithley or National Instruments.
Use the utility software from the gpib interface provider to verify communication with 2401.
If works, then KickStart will be able to use it.

The +ILK is shorted to -ILK with a physical wire.
If using one of our 37xx-ST screw terminal accessories, we make this connection for you.
But if custom wiring to the DBxx mass termination connectors, keep in mind you need to either wire them to a interlock switch or jumper them if you want to close backplane relays on the 37xx card.

For just closing channels, either send a list of channels to be closed, or create a pattern and close/open by pattern name.

-- some TSP commands for channel closure

​​​​​​​reset()

channel.close("1001,1911")

print("******************")
print(channel.getclose("allslots"))
print("******************")


channel.open("allslots")

channel.pattern.setimage("1001,1911","patternA")

channel.close("patternA")
print("******************")
print(channel.getclose("allslots"))
print("******************")

channel.pattern.delete("patternA")

ACS inston a 4200A-SCS that contains 4225-PMU with 4225-RPM will have access to the NBTI library/Project.

The optional WLR library of ACS is targeted for the 2600A or 2600B SourceMeter products.

For HCI and TDDB using 4200 SMU, did you explore the Stress/Measure Cycling of the Clarius software?

It might be fast enough, but will be close.

You most likely will need settings on the DMM to speed it up like fixed measure range, small NPLC, etc.

Given the duration and number of readings, you will need computer to manage a loop of measure and transfer the readings for the 3 hours.

Check out example in appendix B of 2010 manual on the scan card usage.

For the three hour task, I advise you setup a scan list of your 5 channels.  This will have lowest overhead/time needed to switch+measure.
Have your PC side software loop:  init the scan, bring 5 readings, store to HDD, pause if loop rate needs it, repeat.

I ported the commands into a small Python program and timed how long the 100 point repeating average at 10 NPLC requires to return the one value.  Using 60Hz power, it is requiring about 53 seconds.

Even the small Python sample needs the VISA timeout set to long duration to account for the 53 seconds to respond.

The web interface for Send Commands is evidently not waiting long enough for the response to the :READ?
You could try sending command to read the defbuffer1 contents after you know the reading is available if important to read it in the web interface.
Command:  :TRAC:DATA? 1, 1, "defbuffer1", READ