Kelvin Contact Tester/Jig

Guest

I'm currently constructing circuit tester that would test contct kelvin off-line. Would like to know if there's anyone already using such kind? It would be of great help if anyone there could tell or share.

Thanks

Guest

Hello.

May I know what do you mean by testing kelvin contacts offline? I find it quite funny to create a jig to test kelvin contacts offline. Can you elaborate? Also, why are you trying to create a kelvin contact tester offline?

Sebastien

Guest

Hi Sebastien,

What Im saying about kelvin contacts are just a kelvin contactors which is just a part of testsite. This part is removable. Off-line testing means checking the integrity of the contacts that it would certainly pass even with-out testing it on actual production set-up. Of-course, the tester or jig is a simulator like as what of production set-up. Ever use a DMM with force and sence capability testing? We have that kind of DMM, But gee, it requires to person to be able to handle four probes, plus, there's still limitation. The idea and objective of the tester that Im constructing is simply of "PLUG AND PLAY" "WHAT YOU SEE IS WHAT YOU GET". Get it? Tester should force current and sence voltage as kelvin contacts being test.
This kind of tester would surely help as part of the predictive and preventive maintenance and problem isolation.
I asked because if somebody is already using such kind. It could be easier for me to create that kind of tester because of someboby's help.
Of-course if this forum are intended for helping others.

Thanks........

Joined: 31 Aug 2026 Posts: 16

Hi Spongebob,

I've actually made this kind of project six years ago(with my previous company). I believe what you'are trying to do is to check if the resistance of the kelvin contact/s(between the force line to the sense line) is very close to zero ohms. Basing from experience, dirty contactor filaments would cause the resistance to increase especially if the meeting point between the force line and sense line are the dirty ones, since this would add to the total resistance. This might cause some offset in the readings which would result to huge failure rate, or even cause the setup to fail.

I recall that I used LED's to determine if the kelvin contact of a certain pin is OK. By plugging in the kelvin contactor, the LED's representing a good kelvin contacts will emit a green light, while those LED's that represent a "bad contact"(resistance between force and sense line > 1ohm), will emit a red light. Within seconds, you will then know which pin contacts are good or bad, basing from the LED's.

Basing from your elaboration, I can sense that you understand the basic principle of force and sense. The project I did actually followed this principle. The jig/circuit actually consist of:
1. current source = this supplies the constant current across the load(in this case the kelvin contact)
2. Instrumentation amplifier(In-amp) = this measures the voltage developed across the load.
3. comparator = checks if the voltage developed across the load is within your desired level. In turn, you're actually checking the resistance.
4. Multiplexer = switch to connect the in-amp to the pin(the one that is going to be checked).

The circuit works as follows:
1. By forcing a current into the load(say 1mA), a voltage will be developed across it(V=IR). The current is constant regardless of the value of the resistance. This means that the voltage across the load increases when the resistance between the force line and the sense line increases. If the resistance increases to 1ohm, then the voltage developed would be 1mV.

2. You will then need a circuit to measure the voltage acrosse the contacts. An instrumentation amplifier will do this job. An in-amp outputs a voltage that is directly related to the difference of its input times the gain. The gain is adjustable.

Vo(in-amp) = (+IN - -IN)*Gain

An in-amp actually works similarly as with the differential amplifier. However, an in-amp have an infinite input impedance compared to the differential amp.(recall that the non-inverting input of a differential amp have a series resistor connected), hence, it is recommended over the differential amp. to prevent "loading" down the voltage across the load. Note that you need to route/wire the inputs of the in-amp as close as possible to the kelvin contacts so that its inputs will only be exposed to the actual resistance of the kelvin contacts.

3. Once you've measured the voltage across the contacts, you will then need to check if this voltage is within the desired level. As an example, let say that you will set the maximum resistance to 1ohm. Anything that is greater than 1ohm contact resistance is considered a bad contact. This would result to a load voltage limit of 1V(assuming Gain=1000).

Vload = (1mA)*(1ohms)*Gain = 1V
where: Gain=1000

This means that if the measured voltage from the in-amp output is greater than 1V, then the contact resistance is bad.

You will then need a comparator to do this task. You can set the inverting input of the comparator to 1V(threshold), and you can connect the output of the in-amp to the non-inverting input.

Hence, if the kelvin contact resistance is > 1ohm, the in-amp output(non-inverting input of the comparator) would be greater than 1V, which would cause the comparator output to rail to its +supply, since the inverting input is set at 1V(threshold).

If the kelvin contact resistance is < 1ohm, the in-amp output would be less than 1V, which would cause the comparator output to rail to its -supply.

4. To check the other pins, you will need to use a multiplexer to connect the pin being tested to the in-amp. Hence, it is connected between the contact lines and the in-amp input.

As far as I can recall, I made a program to carry out the switching of the mux and checking of comparator output. We utilize the printer port of the PC for the program and the hardware to communicate.

I hope I've provided you enough information. Good luck in your project.
_________________
Asmoth72

Joined: 20 Jun 2026 Posts: 20

Great post, Asmoth!

Guest

Thanks Asmoth

Posted: Mon Jun 20, 2026 9:25 am Post subject:

Joined: 31 Aug 2026 Posts: 16

Yes, you have to fed the current source to the mux first before leading it to the contactor pins. The mux is needed so that you can force the current on one kelvin contact at a time. This way, the current will only flow on the path of the kelvin contact that your currently evaluating.

With out the mux, you won't be able to force the current one contact at time, hence, the forced current will be divided by the number of kelvin contacts that the current source is connected.

In summary, you will need a total of four mux:
a. Two on the force line = one for routing the current source(force line), one for routing the force line side of the contactor to an in-amp.
b. two on the sense line = one for routing the sense line to gnd(current source), one for routing the sense line side of the contactor to an in-amp.

Again, you must rout the in-amp input connection(through the mux) as closed as possible to the kelvin contacts, so that you will only be considering its resistance.
_________________
Asmoth72