Resistance Measurement for Temperature Sensing

[scott_alexsk]

Hello,

I have been working on a circuit to measure the resistance change of a wire vs. temperature, as to glean information about the internal phases. The changes in resistance I am dealing with are only about 1 ohm.

Is there any standard system for grabbing such a small resistance? Currently I have a computer interface that hooks up to the usb port on my computer, and can take analog inputs ranging from 0-5 volts.

I am working on a setup w/ a difference amp, to amplify the difference between a small preset voltage, and the voltage being dropped over the wire, but is there a better way? Please note that I do not need the actual resistance, but just a good relative measurement to mark the phase changes.

Thanks,
-Scott

 

[mgb_phys]

Sir Charles Wheatstone

 

[f95toli]

Well, a bridge is the most accurate way but it is usually quite inconvenient unless you happen to have an automated bridge (e.g. something like the AVS-47B from Picowatt).
But 1 Ohm is a rather large change, a standard 4-point measurement will do.
Just pass a known current through the resistor (if you don't have a current source a voltage source in series with a resistance MUCH larger than the resistance of the wire will do) and measure the voltage across the wire.
Then you get the resistance using Ohm's law.

Btw, it is usually best to measure the current as well, so if your interface has several inputs you should measure it directly by simply measuring the voltage across the large (known) bias resistor (or,even better, a separate small high-precision resistor connected in series).

 

[scott_alexsk]

There are a few more conditions that I did not mention before that make it more difficult than that. Firstly the wire itself ideally cannot have a significant amperage (about 0.05 amps or above), because any significant amperage, if passed through the wire for anything but a short time, will change the resistance measurement, by inducing a martensite transformation in the wire I am working with. Additionally I need to amplify this signal so that it will fill as much of the 0-5 volt range, so that I can get the maximum resolution with my computer interface. Additionally many of these measurements need to be taken, so I can't do something like move around a potentiometer until I get a null-voltage.

Thanks
-Scott

 

[Averagesupernova]

I would think you could use a good quality op-amp to amplify the small voltage developed across 1 ohm.

 

[scott_alexsk]

I just want to mention also, that though the total magnitude of change is about 1 ohm, I am also looking for the very fine details contained in that range. The solution that my electronics teacher and I came up with to optimize the produced values to the measuring range of 0-5 volts, was to use a difference amp with one input's voltage set, to zero out the measurement initially (or w/ some buffer). A flexibility in the circuit is necessary because the resistance of the wire I am working with varies between 4 and 7 ohms depending on what I have done to it. I cannot work on the circuit right now, but I am interested in the best solution I can get.

Thank you guys for your help!
-scott

 

[waht]

Here is how HP did it. This is a 30 year old transistor based milliohm meter. Page 23.

http://cp.literature.agilent.com/litweb/pdf/04328-90008.pdf

Basically, run a constant current through a resistor you are measuring and then measure the voltage across your resistor.

The HP design uses an AC 1 KHz current source with some kind of a phase lock loop.

 

[zeitghost]

Use an instrumentation type amplifier... such as INA101.

Also use 4 terminal connection to the sample... 2 terminals for the current and 2 terminals for the measurement.

It's known as Kelvin connection.

 

[scott_alexsk]

Ok then... setting up the kelvin connection seems pretty standard w/ the constant current source and all that...but what would be the best way to amplify the voltage drop I get from the second set of connections? If the drop I have is a few millivolts at best, what would be the best way to dilate it to the 5 volt range?

Thanks,
-Scott

 

[zeitghost]

Amplification?

The hint was in the mention of the INA101... it's an instrumentation amplifier and you use 1 resistor to set the required gain.

It has very high Common mode rejection ratio so noise pickup is minimised.