Showing results for 
Search instead for 
Did you mean: 

How to encrease operation speed when measure with Keithley 6430


To fix the problems with emoticons:


Go to your Discussion Forum profile settings, My settings, Preferences, Display.

There is an option called "Emoticon Type" it is probably set on Use default (smiley).  Change it to None.

Message 21 of 24

I agree, there are a few strange things happening there.  Why is the sourced current not reaching the right set point during the sweep?  I suspect you need to look for poor contacts and also supply a delay time in the measurement to allow for settling.  I supplied an updated version of the program with a few more knobs to turn.  (Auto-zero; use ON or Once), I'm using NPLC caching and setting the averaging to 0.1 power line cycles (PLC) to speed up measurement.  The commands are 


Smiley FrustratedYST:AZER:CACHSmiley FrustratedTAT ON;
Smiley FrustratedENS:CURRSmiley Very HappyC:NPLC 0.1;


Let me know if they cause errors.  I can't test the code. Smiley Wink



Thank you. You know a lot about this device. I checked measure_R_cds.via. It works without errors and quick. I'm half way to the goal with your help.

Nest step it is a program for measurement IV curse.

@cstorey wrote:



Averaging won't help if your setup is poor.  You would be building errors into your measurements.  An SMU is a complicated beast!  Its already doing a lot of averaging for you as it auto-ranges to get the best measurement it can.  Taking repeats of the same thing won't really help.


But, why not supply more current and reduce the noise?  You are supplying 1uA to a 1kOhm resistor which means you are measuring 1mV.  That's a pretty low signal level and susceptible to noise pickup unless you are careful about shielding and cable setup.  (Resistors can be noisy!)


My approach to resistance measurements is to sweep current, usually in log steps over 2 decades, and measure current and voltage.  I try to make sure the measured voltage is in the 100mV-1V range where meter error is minimal.  I then fit the measured results on a linear scale (R=V/i) and extract the resistance using a linear regression.  I use the residuals of the fitting process to determine quality of devices/measurements. 


This particular meter also has a built in Auto-Ohms measurement method.  It does a logarithmic current sweep for you and just returns a single resistance value.  Very fast and in my experience very accurate for devices that you know to be ohmic.  (The only reason I do a sweep is to determine if a device is ohmic.  If I am sure of that then I usually use the Auto-Ohm method to look at device-device variations.)


See how the new version work.

I use 1kOm resister only as a test sample. Real sample will be a Josephson Junction. It is a combination of superconducting films.  The measurements will be below 4.2 K (very low temperature). The typical current for my samples is betwenn -100 to 100 microA

Hmm. I try to improve connection to resistor. Now it looks like four wires soldered to the resistor with one end, and the second is soldered to  banana sockets. It is poor, I agree.

I have got again strange result for sweep_IV_v1.1.png 



0 Kudos
Message 22 of 24

Hmm, are you sure you are using the second version?  Mine looks different..







0 Kudos
Message 23 of 24

@cstorey wrote:

Hmm, are you sure you are using the second version?  Mine looks different...



sweep v2.png


Yes. I have some warring when open your .vi. So I repeat it.



0 Kudos
Message 24 of 24