He wouldn't be soldering any of them together. He would have the current leads where they presently are and then move the voltage taps to a separate location on the pins. That should work for him. The idea is to move the voltage taps inside the voltage drop associated with the current injection joints and putting them to the outside of the solder cup will do that.

No, it won't.

Just be sure we are commenting on the same arrangement I have drawn up a mock drawing of what he is trying to do.

The following circuit diagram will show the electrical equivalent and why using two connections to implement a Kelvin type connection does not work. The voltage drop will also include the pin resistance, and worst of all, the variable contact resistance on both ends of the unknown resistor.

 

Ah. I see the disconnect here.

You are assuming that the pins he has in his photographs are used to connect to the DUT. If this is the case, then I wholeheartedly agree with you.

I am assuming that the pins he has in his photographs ARE the DUT and that he is trying to measure the contact resistance when these pins are mated together.

I just went back and read the first couple of posts and I really can't definitively say which is the case. Hopefully the OP will clarify this.

 

Agree. After re-reading his posts that seems to be the case.

If the DUT IS the pin/socket mating resistance then what you have suggested is 100% valid.

BTW, I've just watch a Youtube video on an accomplished machinist Mr. Robert Renz making a kelvin clip. Quite interesting.

http://www.youtube.com/watch?v=wwgO6Lg4RZM

 

Ah. I see the disconnect here.

You are assuming that the pins he has in his photographs are used to connect to the DUT. If this is the case, then I wholeheartedly agree with you.

I am assuming that the pins he has in his photographs ARE the DUT and that he is trying to measure the contact resistance when these pins are mated together.

I just went back and read the first couple of posts and I really can't definitively say which is the case. Hopefully the OP will clarify this.

Yes..mating. that's the word.

Also thank you guys for the good heads up and advice

 

Agree. After re-reading his posts that seems to be the case.

If the DUT IS the pin/socket mating resistance then what you have suggested is 100% valid.

BTW, I've just watch a Youtube video on an accomplished machinist Mr. Robert Renz making a kelvin clip. Quite interesting.

http://www.youtube.com/watch?v=wwgO6Lg4RZM

I've only watched the beginning of the video. I think his probe would be quite useful under many circumstances, but I definitely don't agree with his claim that it works under all conditions. I would love to see him try to use his probe in the application where I first learned about 4-wire measurements (and long before I heard it associated with Kelvin's name).

When I was a junior (undergrad) I did a cooperative education stint (the first of three) at NBS (now NIST) and one of the things we were doing was measuring mechanical stress effects in superconductors. So imagine taking a wire that is about the size of a pencil lead (most of them were well under 1mm in diameter). Now imagine placing this wire flat between two anvils that are about 1cm wide. Sense you need to measure the voltage drop only across a portion of the wire that is fully under stress, your voltage taps have to go between the anvils and be soldered to the side of the wire at a separation (which must be measured with some accuracy, but that can be done later) of about two millimeters. Those voltage tap wires then go up and out to a nanovoltmeter. The current probes are soldered to the wire further out and will be injecting about 1000 amps of current while the whole thing is submersed in liquid helium in the core of a 10 tesla magnetic field. Oh, and that solder joint that is connecting the voltage taps onto that wire? There can't be any solder that gets onto the top or the bottom of the wire since that will result in stress concentration points for the anvils (which are going to be applying well over a ton of force to that poor 1cm long piece of wire!)

 

Wow, who would have thought that resistance measurement can be that interesting and challenging?

With a superconductor, by definition there shouldn't be any voltage drop across the wire section but if the wire somehow becomes non-superconducting, then 1,000 ampere flowing through it would surely melt it in a flash.

Is the change between these two modes happens abruptly or gradually so that one can reduce the current to a safe level?

p.s. My apology to SandiegoSD for hijacking his thread.

 

With further apologies to the OP (hopefully you're finding this useful or, at least, interesting) there are several different models, but (at least for the low-Tc stuff I was working with) the best model (i.e., the one that matched our experimental data the best) was the E_J power law model which has the voltage increasing in proportion to the current raised to a constant (the n-value). Typical n-values (at least in those days) were in the neighborhood of 10 to 20. At an n-value of 10, doubling the current results in (roughly) a thousand-fold increase in the voltage (and a million-fold increase in the power).

So you had to be careful. We would often take hysteresis curves and walk the voltage up until we had about 5uV/cm and then walk it back down to well below 1uV/cm and do that a few times so that we could explore the results of permanent deformations on the sample.

Our current supply was derived from 2V submarine batteries and we has huge NPN transistors (150A each) in parallel, each with a ballast resistor to prevent thermal runaway in them, and a 0.1mohm current sense resistor. At the end of everything, I think the max voltage we could apply to the sample was about 0.5V. So if it went normal (i.e., quit being a superconductor), the current dropped real fast.

The fun times were when you were working with a coil sample. One time one of the other teams had a coil sample (imagine about 10 turns of a coil that is about one inch in diameter) that was running 4000A in it when it went normal. There was enough energy stored in that inductance to vaporize the coil as well as the end of the cryostat it was mounted on. More impressive, it vaporizes the liquid helium (which doesn't take much) from the core of the superconducting magnet. You now had a 74H inductor with about 120A of current go into thermal runaway. The magnet is protected by having big diodes and resistors down in the cryogen to shunt the current into but the result is a magnet dewar that looks and sounds like an old steam locomotive as it vents the vaporizing helium.

Oh, I ran across the paper the was published in Journal of Applied Physics and the two types of conductors we made measurements on had diameters of 0.7mm and 0.6mm. And only about 25% of the area of the conductor was actually superconductor, the rest was stabilizing and support material. Even at that, we had to get to well over 5 tesla and about a ton of load before we could drive it normal with 1000A of current. At zero field and zero load it probably would have taken over 10,000A to do it.

 

Thanks WBahn. That was really interesting to know.