variation in resistance measurement using DMM

[R.V.]

yes, i can use 1mA in another standard but .1mA is one of the standard for test. when i perform test on 1mA variation become very small but exist.

Does Offset compensating ohms works in voltage measurement?  i think it works only when we are measuring resistance through single dmm in 4 or 2 wire method

 i cannot use directly 4071 for resistance measurement, because in DMM we cannot set the voltage & current value of source & i have to perform the test on specific values of voltage & current of power supply.

[Brandon_G.]

Hi RV,

If you wish to calculate the thermal voltage with the DMM in voltage mode you can calculate this in software.  Basically what you will need to do is provide 2 different test currents from the SMU and take2 different voltage measurements.  Once you take these 2 different measurements at 2 different test currents you can solve for the system of equations below to find your resistance and thermal voltage.

V­_1measured  = V_thermal + i₁ *R

V_2measured = V_thermal + i₂ *R

Solving for R and V_thermal we get:

R = (V­_1measured - V_2measured)/ (i₁ - i₂)

V_thermal = (V­_1measured * i_2  - V_2measured * i₁) / (i₂ – i₁)

With OCO, we would typically set i2 = 0 but since the SMU can only output a minimum of 2% of the range, you can use a positive a negative test currents to do the calculations (+.1 mA & -.1 mA).

Thanks!

Brandon G 

[johnsold]

Because your signals are so small, you need to be very careful about making the measurements.

Any thermal emfs will be a big problem.  The two ways to avoid thermal issues are to keep the temperatures of everything in the measurement system at the same temperature (probably within small fractions of a degree) or to use the same material for all conductors.  You cannot control what materials are used in the instruments.  Unless they were specificlaly designed to generate very low thermal emf, they likely have several different materials: copper wires, possibly tin plated, solder (an alloy of multiple metals), perhaps steel with cadmium or nickel plating on screws, and ???  Similarly unless the equipment was designed to be isothermal at the connection and internal measurement points, a few degrees of temperature difference is likely.

Are your measurements synchronized with the power line frequency?  Some precision DC instruments have that capability so that any line frequency interference picked up by the wiring can be averaged out of the final result.

Do you reverse your polarity periodically to account for offsets?  This is a standard metrology technique for high precision or small signal measurements.

Lynn

[R.V.]

i use offset compensated ohms programmatically by apply two different current +i & -i then i got first reading 2.5mOhms then it decrease to 1.5mOhm in third reading & got stable within the limit of  + - 0.2mOhms. The same pattern i saw for different pattern but after  3 to 4 Hours, voltage drop starting varying up & down & some time i got NaN output.

[R.V.]

when i measure the resistance of same wire using the 4 wire method using single DMM then i got  3mOhms on the other hand using OCO programmatically i got 1.5mOhms. i think i am calculating more Vt than normal.    

[johnsold]

I do not have any other ideas at the moment.

Lynn

[R.V.]

ok

[Brandon_G.]

Hi RV,

I had a few questions and some thoughts on your setup I'd like to find out more about...

1. Are you measuring resistances of 80 -140 mohm or 2.5 - 1.5 mohm?  In your earlier posts I had heard 80 - 140 mohm, but in your most recent post it looks like your measuring a few mohms.  Can you clarify this?

2.  Just to verify, are you using the maximum digits for both your voltage and current DMMs ( 7 1/2 for voltage, 6 1/2 for current) and are auto-zero and ADC calibration enabled?  Also check to see your line frequency is correct (either 50 or 60 Hz).

3.  When you look at your voltage and current measurements is it just the voltage DMM fluctuating in value or is the current changing as well?

4. For each calculated resistance, are you taking both a -i and +i reading each time?  I would just like to confirm that both +i and -i measurements are done each time instead calculating thermal voltage once and using this as a constant throughout subsequent measurements.

5.  How much time do you allow for settling when you set your current setpoint to when you start taking a measurement?  If we're beginning the measurements right after setting our current and don't allow for settle we could be integrating our voltage and current measurements during the transition period which will throw off our measurements.  I would set the DMMs to digitizer mode and see how much time you need to get a stable value.

6. Since we are dealing with such low voltage levels it would be a good idea to use twisted shield pair wires for the DMMS.  You can connect the shield to the PXI chassis ground on just one end of the cable to prevent ground loops.

7.  You may also have self heating in your switch itself.  If you source .1 mA of current through the switch itself and take a voltage measurement across the switch points, how does it drift over time?  Does it have sharp changes or is it more gradual?  Let the current run through it for several minutes and see if it stabilizes.

8.  I might suggest using a precision resistor similarly valued to your DUT with a low temperature coefficient for troubleshooting purposes as well.

Since we're reading such small voltage levels it will be real challenging to get steady measurements but hopefully these troubleshooting steps will help reduce it and give more repeatable results.  Let us know what you find.

Thanks!

Brandon G

[R.V.]

Hi Brandon,

First of all i want to say that now i am getting reading within the range of 0.5 mOhms for all different values of current except for 0.1mA current. Now i implement OCO by giving two different current i1 & i2, in which i select i2 =0(means i  did not initiate SMU ).

1. i am measuring the resistance of a thin wire having length of 100mm & cross section 0.5 sq.mm. Before using OCO in program, i was getting reading b/w 80mOhms to 140mOhms but as i use OCO in my program i m getting very low reading.

2. i m using two DMM both at 6 1/2 digits, one for measuring current in the circuit & second for measuring voltage in the circuit. now i am not using any function for auto-zero, ADC calibration & line frequency.

3. it is the voltage which is fluctuating but current remains the same in the circuit for each time.

4. for each calculated resistance, i m taking current +i & 0. i calculate thermal voltage each time for the calculation of resistance, i am not using a constant thermal voltage through out the measurement.

5. i took reading just after closing the corresponding switch pins & do not give any time to settling the current. 

6.i do not have twisted shield apir wires.

7.i will do this.

i try to answer your questions, now i want to ask some questions

1. how can i set the DMMs to digitizer mode to see how much time i need to get a stable value.

i will use these troubleshooting techniques in my program & hope i got more stable results.

Thanks

[R.V.]

Thanks to both of you johnsold & brandon  to provide me right direction to complete my project, now i am getting more stable readings with very less variation.