Why do i have current on the voltage sensing wires of a Four-Terminal measurement ?

[SandiegoSD]

This https://www.electro-tech-online.com/custompdfs/2012/11/LowLevMsHandbk_1.pdf is worth a read, but didn't have the info I wanted.

Look at this article: http://en.wikipedia.org/wiki/Four-terminal_sensing and the video referenced at the end.

Here is Keithley's Low Resistance handbook: http://www.keithley.com/data?asset=55835

Yep I have the two Keithley books.

What were the values of voltages and currents for both measurements?

The soldered connections do contain two dissimilar metal junctions. wire to solder and solder to pin

Clipping the leads to the pin don't.

Looks like I do need to use separated clips/jaws for true four-wire setup.
If I directly put the volt meter probe on the metal pin(outside of the pin body), the reading is slightly smaller than what's measured at the end of the voltage sensing wire. So very likely the blob of solder joint and some contact resistance are picked up when using the wires like this. I'll try soldering the volt sensing wires on the outside of the metal pins, and keep the current wire as is.

 

[KeepItSimpleStupid]

It's probably not "contact resistance" per se, but the two junctions of dissimilar metals.

There is thermal gradients, but i doubt it.
There is contact resistance of solder to wire and solder to pin - Maybe, but I don;t think it's significant
Dissimilar metals - possibly.

 

[SandiegoSD]

100 times. Works for bigger signals very well with accurate gain.
Not so well for small signal like 0.3mV *100 =(the amplified signal I get at Vout) 27.8 mV..

 

[KeepItSimpleStupid]

The amplifier itself will give you a 1 to 1.5 mV output offset, so 27.8+1.5 = 29.3. If I did that right.

Besides at 0.3 mV there isn't enough significant figures. You would need to say 0.300 volts for your statement to have meaning and add some reading errors to that too. Yea, picky.

 

[ronv]

The amplifier can't drive all the way to ground - close but not all the way. I would try a gain of 1000 and see if it doesn't act better. This will amplify the offset as well, but it is pretty small.

 

[KeepItSimpleStupid]

@ronv

I have an idea that might make the circuit better. How about raising the output reference by a diode drop?

Even better, a 200 mV voltage reference or something?

 

[ronv]

I'm not real clear how the Ref pins work, but I think the temperature dependance on a diode voltage drop might be a problem at such low signal levels.

Might also be a good idea to use twisted pair for the sense lines.

 

[KeepItSimpleStupid]

I do agree about the temperature dependence, but it should not really matter.

The idea would be to take the Force and Sense pins to the anode of a diode and take the cathode to ground. The output would be taken from Vout and the anode of this new diode and not ground.

Output would be from Vout and the anode of the diode rather than ground. As the ground potential moves with temperature etc, it's still relative to the moving reference which should be OK.

I did a 4-terminal I-V converter with the AD625 which also has the output sense leads. I learned a lot with that design.

Twisted pair shielded cables would not be a bad idea especially if the cables are long. I suspect they are not.

The bigger problem with sensitive milliohmeters is thermal gradients.