Tektronix and Keithley

Dear all,

I have to "program" a sweep with an "exit condition" on a 4200 together with other instruments via GPIB. I was wondering how I can use KXCI commands to achieve the functionality which is called "Compliance Exit Condition" in Clarius (See page Page 6-106 / 433 in the manual):

"The compliance exit condition options are available from the Test Settings pane when you select Exit Condition.

When the source goes into compliance, you can choose the action that the system takes.

You can select one of the following exit condition options:

None: The run continues. This is the default setting.
Test/Device/Subsite/Site: The 4200A-SCS exits the test/device/subsite/site that is presently being run. If there are additional tests, operation continues to the next test.
Project: The 4200A-SCS exits the project. All testing stops."

I also know that in KXCI, a sweep is programmed with the VR and IR commands (Source Setup (SS) > Voltage Ramp (VR) on Page 863):

"VR and IR
This command sets up the VAR1 source function." (Note the missing compliance exit behaviour option)

For other (older) instruments like the AgE5270 analyzer, there is a feature called "Automatic Abort Function", accoridng to its manual:

"The automatic sweep abort function stops sweep (increasing or decreasing source output value) when one of the following conditions occurs. This function is useful to reduce sweep time and to prevent damage to the device during sweep measurement.
• The output reaches voltage compliance or current compliance
• A measurement value exceeds the specified measurement range
• An SMU oscillates

The automatic abort function is enabled by using the WM, LSM, or BSM command."

We found the KXCI command in the manual (on page 895, a.k.a. 9-53). It's name is EC ("Exit on Compliance")

EC
This command sets the condition to exit the test if compliance is reached.

Usage
EC A, where A is the Action on compliance:
* Off (do not exit if compliance is reached): 0
* On (exit if compliance is reached): 1

Example
EC 1
This command enables exit on compliance.

What is meant with test? Is it the measurement that is executed with the ME command?

I managed to use the EC command in a simple DC sweep. However, now I am wondering about how the 4200 returns results:

After the compliance occured and the test is exited, the "missing" values are simply returned as "(...) N 0.0000 (...)"
According to the manual, page 900, 'N' stands for "Normal", which is obviously not the case; and the 0.0000 is a totally possible measurement result. How can I tell then that these "values" correspond to to the situation after the exit condition occured?

The 5270 Analyzer also has an "Abort on compliance" feature, and to my mind it indicates much better that values are "invalide": The status is 'V' indicating "measurement data si over the measurement range or the sweep measurement was aborted by automatic stop function or power compliance", and the value is fixed to "199.999e+99", which is clearly an invalid measurement value.

Is it possible to achieve some indication on the 4200 as well?

Let me clarify, this is for the padded out data values that lack the signaling that they are from the measurements that never occurred due to exit the sweep for compliance?

Curious, are the last few measurements before exit showing the C or T indication for compliance?

What version of KTEI or Clarius are you running?

Hi Andrea,

I would answer your questions in reverse order:

'What version of KTEI or Clarius are you running?' We are using plain message IO (KXCI), KTE V9.1-SP4 (Windows 7)

'Curious, are the last few measurements before exit showing the C or T indication for compliance?' - Only when the exit condition is off and the SMU is in compliance, see below.

"Let me clarify, this is for the padded out data values that lack the signaling that they are from the measurements that never occurred due to exit the sweep for compliance?' - Yes, exactly. I have prepared a minimal working example with a resistor connected between SMU1 and ground:



The graph should show that I am doing a current sweep on SMU1 and measuring both current and voltage at this channel.
When I set the exit condition to false, the SMU will go into compliance at the end of the sweep and continously measure approximately the same values for I and V, and indicate that the SMU is in compliance with a C for the last two measurement points:
However, when I set the exit conditon to true, which means that the SMU will switch off after compliance occured for the first time, the SMU pads the last measurement with "zeros" but does erranously indicate them with an "N", suggesting that the last measurement point was measured with "no errors":
For better understandability, let me wrap up the measured currents in both cases

## | EC0 | EC1
01 | N 14.79650E-18 | N 6.810675E-18
02 | N 5.000287E-06 | N 5.000393E-06
03 | N 9.999900E-06 | N 9.999878E-06
...
18 | N 90.00053E-06 | N 90.00116E-06
19 | C 91.89384E-06 | C 91.89719E-06
20 | C 91.90099E-06 | N 0.000000E+00

Is this feature intended? As mentioned in one of my older posts, the 5270 would indicate the last point with a V, standing for "measurement aborted". Would an update to the software on our 4200 change this behaviour?

Hi Andrea,

did you or your colleagues have found the time to look into this issue? Could some of you comment on the status indicator of the padded datapoints?

With kind regrads,

Marcel

Hi Marcel,

Using latest release on 4200A (Clarius 1.8.1), I see same thing.

Force current into 1.1MΩ resistor. Current sweep from 0 to 10uA, 2.5uA step size.
Set voltage compliance to 5V.
Current source list of values: 0, 2.5uA, 5uA, 7.5uA, 10uA
Expected Voltage: 0, 2.75, 5.5, 8.25, 11V NOTE: bold values will not occur due to 5V compliance limit

I like your idea that any padded values from after the exit condition could have a different status than N. I can enter an enhancement request for the feature.

As a work around, I suggest you implement some logic such that if you encounter the C status, you know that compliance was detected and any data values from after that are simply padded out "N 0.000000E+00" values to make the data array the expected size.
Since the SMU did indeed abort, it is not able to report any meaningful voltages, currents or status info.

Here I monitored the SMU voltage when forcing the current sweep into 1.1MΩ resistor with 5V compliance limit.
With the DMM6500, I trigger on falling edge. You can see the timing difference of exit on compliance or not.