Andrea C

On the graph view, where you have data, what is name of the Active buffer.  Go to Data tab once Graph view is displayed.

Do your trigger blocks make use of same buffer name?

Hi Maia,
I'm going to ask the factory to take a look at this.
Excellent detail here and steps to duplicate.  Thank you for taking the time to put it so clearly.

Andrea

I suggest you examine the Quick Start Guide, User Manual and Reference Manual, in that order.

All those can be downloaded from tek.com:  Manuals for 2611B SourceMeter

Quick Start Guide attached to post.

The instrument is shipped with a mating connector and hood/strain relief for making a cable for the SMU channel.
This connector has simple screw terminals to receive your wires from the LED device.

Both the User and Reference have section on DUT connections with pin identification for the connector, etc.
For LED device, recommend to use 4-wire connections.

If you do not want to fabricate a cable of your own using the supplied mating connector, consider the 2600-BAN cable.
Check that the sheathed 4mm banana connectors that it presents are compatible with your LED connection requirements.
There are numerous banana adapters to clip lead and similar on the open market (Pomona, Mueller, etc.).

How are you planning to operate the instrument?  Manually or software?  Does the SMU need to coordinate with any instrument such as light meter, etc.

If just wanting the SMU to do IV curve and not wanting to write any software, consider the KICKSTARTFL-IVC application module of KickStart.
You can connect the 2611B to the PC on USB, LAN or GPIB and use KickStart.  RS-232 is not supported by KickStart and 2600B.

For the 2182A, in Chapter 11 of the User Manual, the error codes are defined.
Error -101 = invalid character
Error -102 = Syntax error
Error -410 = Query Interrupted

The first two indicate you are sending invalid commands to the instrument.
The -410 just means you sent a query type command (one that ends in question mark), but did not fully read the response before sending a new command.  Even use of *IDN? query command can cause the -410 if you do not read the identify yourself response before sending new command.

To know which command is wrong/causing error -10x, LabVIEW has some single step debug features.
Or, install the NI IO Trace tool and invoke that.  It captures the bus traffic between your LV and and the 2182A.
You can see which command causes the error condition to be raised.

When using LAN based connection to KXCI, it is normal for every config command you send to respond back with ACK\0.
Pg 2-11 in the KXCI Remote Control Programming manual shows ethernet based write and read examples.

For example, sending ME\0 command, the 4200A will respond back with ACK\0

In your code, if using LAN to KXCI, plan to do a write + read for every operation to pull the ACK\0 off the bus.

I think you are just having a lot of backlogged ACK\0 in the output queue that need to be read and this is why you do not get expected response to the SP or the DO commands.
 

In datasheet for the 2380, pg 3 Specifications, it lists the minimum operating voltage.
Suppose you have the 2380-500-15 model and using the low current range (3Amps), the min operating voltage is 0.6V at 3Amps.  It increases to 4.5V at 15Amps for the high range.

This means when operating in CC mode and pulling 3Amps, the 2380 will develop a voltage drop of 0.6V between the HI to LO terminals.  Think of the 2380 as using the output stage of a FET to control the CC mode.
This voltage drop could be attributed to the I*Rds of that FET plus other I*R drops in the path inside the 2380.

Consequently, your device that is supplying the current cannot have a voltage potential lower than this minimum operating voltage of the load.  The load needs to be able to develop at least the minimum voltage drop in order to maintain the CC set point.

In situations where you want to test your device at voltages lower than the minimum operating voltage, use an external (linear) power supply in series with your device.  One floats on the other.
Use the remote sense lines to still measure the voltage only from your device rather than device + external supply.

Otherwise, the entry of the max and min voltages during CC mode should be influenced by the I*R drop and open circuit voltage of the device that is suppling the current to the load.


 

Hello Maia,

That is very odd behavior.

I see you have the latest firmware, A14.

What if you manually operate the instrument from a reboot/factory defaults?  Does the current range adjust appropriately then?

As you increase the source voltage a little bit, does the 200pA range ever report overflow?
If you have a volt meter or scope channel, what does the 2V output do as you approach full scale 200pA and push past?

Seems like the :MEAS:CURR? rather than a :READ? will be needed.

How is your remote connection to the instrument?  Looks like you have a LAN to something converter?  Using GPIB or RS-232 on the 6517B?

What is your exact model and firmware level?  The response to a *IDN? command will tell us.

For the delay, are you talking about the stime parameter?  Or the delay inside the my_wait_complete().
Of course, if the my_wait_complete() loop rate is very slow relative to the speed of the Source-Measure sweep, then it will be slow to detect the exit condition.  In the sample as posted, it uses 50msec loop rate.

What parameter values do you use when calling the ExitCondition_SweepVLinMeasureI(smu, startv, stopv, stime, points, current_limit, lowest_range) function?
And at appx what voltage and currents are you expecting the exit to occur?

Keep in mind, during the active trigger model sweep and during the my_wait_complete() loop, attempts to interact with instrument from the PC will be blocked.  This is why the logic of an exit condition is moved into the TSP code.

As for ACS, it is doing a for/next loop and able to check for compliance at each iteration.  These are implemented as functions in the TSP runtime memory on the instrument so there is no required bus traffic to the PC for each iteration of the loop to decide about exit condition.

 

Should Sense HI also be guarded?  Ideally, yes.

In concentric conductors like a triax cable, is easy to envision the minimization of leakage currents from the signal on center pin if the next layer inner shield is driven at same potential by the guard.  No delta V = no current.
This is one motivation for guarded cable especially for the Force HI.

But guarding has additional benefit of smaller effective RC time constant if the coupled capacitance between the layers does not have any charge up requirements.  This would be the reason to also guard the Sense HI line.
The input R of the Sense HI is typically > 10GΩ

For long cables, you might find some settling time benefit by guarding depending on the speed of your IV curves, size of dV steps, etc.

Looks like your Python code was written for different product than the model 3706A.
You are sending SCPI syntax commands.
The 3706A is TSP command set only.

:SENSE:FUNC "VOLT:DC"  becomes dmm.func = dmm.DC_VOLTS

See if the attached Getting Started document helps you.