Combining four-probe measurement with Wheatstone Bridge

[ali8]

I have an almost squared thin film with 4 contacts along the corners. Using 4-probe measurement, I can eliminate the contact resistance. On the other hand, Whetstone bridge can be used to record resistance noise as a function of time (because the balanced bridge will have zero output voltage, so any fluctuation is considered noise). The problem is that Wheatstone bridge requires a 2-terminal sample, or a 4-terminal one with the resistances of the two halfs being equal, which is not the case here.

Is there any Wheatstone-like bridge that can incorporate 4-probe measurements as well?

 

[chuckey]

You could loop two adjacent corners together, this has "folded your film in half" so now you have two terminal device. If you are interested in the noise that this film is producing then a four terminal measurement will led to inaccuracies, because you have admitted that the contact resistance will be a significant part of the overall resistance and will develop its own noise.
Frank

- - - Updated - - -

If you are trying to use a Wheatstone bridge approach, how are you going to compensate for the noise generated by the other arms of the bridge? What is the typical value of your "folded" film resistor? Over what bandwidth are you going to measure the noise?
Frank

 

[ali8]

So you are suggesting two parallel bridges, one for each two corners? This is two 2-terminal measurements, still has contact resistance.

The other arms consists of high quality resistors with minimal noise (other than thermal noise). My device resistance is hundreds of ohms, contacts are thousands of ohms. Bandwidth is two to three decades below 1 Hz.

 

[chuckey]

Because of the very high contact resistance, I can only think of one method that is to use the device its self as the wheatstone bridge. Earth one corner and feed a constant current into the diagonal opposite corner. Now there will be a volt drop down the diagonal. If you measure the volt drop from the free corners to earth they should be the same. So if you use a high impedance differential input op amp with each input to each "free" corner, the static DCs balance out leaving noise. So if you trim out the op amps unbalance, via a very long time period filter, when the long time filter is removed anything that is measured is noise and instability. How you interporate the results is up to you!
Frank

 

[FvM]

I don't think that making assumptions about symmetrical contact resistance isn't a good idea. Instead I would use electronic compensation means for the 4-wire circuit.

I believe that the accuracy of a high performance DVM in 4-wire circuit will be comparable to or better than a wheatstone bridge.

 

[ali8]

@Chuckey Your method doesn't work because the film is not uniform; its resistance is 10 times higher along one direction than along the other. Also, with regard to the technical details you provided, I don't think I'm going to use DC. I use AC Lock-In amplifiers that can measure the signal more accurately. They also have 24 db/octave RC filter, plus I do additional DSP in computer. When they measure voltage, they provide high input impedance so external op amps should not be necessary.

EDIT: if your method is to source current along two opposite corners and measure along the other two, then this is not an Ohmic measurement, You should only source and measure along parallel pairs, not crossed ones.

 

[FvM]

A DVM can reduce a resistance measurement to a ratiometric measurement with a reference resistor, and zero offset and thermoelectric errors by switching the current excitation. In so far it also cancels most drift effects. An AC measurement may still bring advantages in a certain resistance range, athough it involves it's own parasitic effects like circuit inductance and capacitance as well as skin effect.

An AC measurement could use an (uncalibrated) current source, a reference resistor and a two-channel differential amplifier.

 

[chuckey]

"Whetstone bridge can be used to record resistance noise as a function of time", so you want to measure the ABSOLUTE value of resistance as well as the noise? So you have this dodgy film with high contact resistance that completely swamps its value with four terminals on it. So what are you attempting to measure? Have you got defined current contacts and voltage contacts like one gets on proper current measuring shunts?
Frank

 

[ali8]

"Whetstone bridge can be used to record resistance noise as a function of time", so you want to measure the ABSOLUTE value of resistance as well as the noise? So you have this dodgy film with high contact resistance that completely swamps its value with four terminals on it. So what are you attempting to measure? Have you got defined current contacts and voltage contacts like one gets on proper current measuring shunts?
Frank

No, with Whetstone bridge the voltage output is ideally zero, since we choose the variable resistance to be equal to the sample resistance (2 Terminal). We are not using the bridge to determine the resistance, we do know the resistance, we simply record any deviations from zero. We can't do that this time because the contacts have such a high resistance.