Measuring very small resistance change

[jjlwork]

Is it possible to read a change of 100 micro-ohms in-expensively? This 100 micro-ohms indicates a 100C change in temperature in the material I am testing and I want to be able to cut the power supply to the device if the material reaches 300C from ambient temperature so basically once it has a change of 300 micro-ohms. It would be nice to be able to measure this down to 1 micro-ohm so I can get a real temperature but I am guessing that would be even harder still. If the 100 micro-ohm  change is even possible. If it is not possible is there a way to do it crudely to say the material has changed somewhere between 300 micro-ohms and 600 micro-ohms so shut down. The device/circuit does not need to measure the real ohms of the device accurately just the change in ohms. Pointing in the right direction to research would be fine as well but I've googled the hell out of this.

[ataradov]

Have a look at a Wheatstone bridge. With appropriate design and ADC selection, I think what you want to do sounds to be feasible.

[barry14]

A critical piece of information you need to provide is the nominal value of resistance for which you are trying to detect such a small change.  Detecting a 100 microhm change in a 10 ohm resistor (10 ppm), while still not easy, is much easier than detecting the same change in a 100 kilohm resistor ( 0.001 ppm).

[hamdi.tn]

correct me if am wrong you trying to heat a piece of metal and detect the change of resistance, so i suppose initial resistance is basically on the range of milli-ohms in theory it shouldn't be that hard , a constant current flowing on the metal, an amplifier and an ADC. i will say 1A / 1000x gain
but with such amplification any noise will be amplified and may cause reading error so, need to be careful on that.

[jeroen79]

Wouldn't it be easier to just use a thermocouple or other temperature sensor?

[Ian.M]

If you are heating to 300K above ambient, and need to detect such a small resistance change, you are going to have major problems with contact oxidisation and thermal EMFs. You'll need welded contacts of the same metal (and alloy composition) going to an isothemal block for the potential wires, and to use four wire measurement to have any chance of doing this successfully.

Unless you are working on a micro-scale that makes it impossible, a type K thermocouple in a drilled hole in the metal, or a surface one insulated with ceramic wool is probably your best bet. 

[Marco]

How much current are you putting through it? If it's DC can you put an AC current on top for the resistance measurement?

[TimFox]

Measure the resistance with low or audio frequency AC excitation of a wheatstone bridge.
After a good AC-coupled amplifier, use a lock-in amplifier or other synchronous detector to get a narrow bandwidth around the AC frequency for null detection.
This will avoid DC offset, thermal EMFs, and possible 1/f noise.