# Kelvin “4 Wire” Resistance PCB Design Questions

I am designing a PCB that will utilize 4 wire measurement for 40 or so test points.

A little background here - day to day I troubleshoot PCB's to component level, I'm an electronics technician and not an engineer and I'm okay with that. 95% of the time I can find an issue with a circuit based off of how the DVM measures resistance and it's behavior. My boss wants me to automate my process using LabVIEW; so I need to make a PCB that does just 4 wire resistance testing. I have an existing pogo pin block and a PCB that mounts to it, I just need to re-spin the PCB board and write the LabView program.

Are there any special PCB design details I need to consider when doing this very basic design?

Please be gentle; this is my first engineering project.

• Put your high-impedance sense element as close as possible to the sense connection. – Ignacio Vazquez-Abrams Jun 5 '14 at 16:46
• Oh god for everyone's sanity don't use lab view. Use a saner language, like python or something. – Connor Wolf Jun 6 '14 at 1:50
• To be honest I really wish I could; but that was squashed by my supervisor :(. He spent half of last years budget to get everyone in the department trained in using it "except me, I wasn't here yet" and now he in his own words "wants to see a return on the company's investment". It's an awful programming language. – MEAM Jun 6 '14 at 12:49

The nature of the Kelvin measurement is that you have a separate current path and measurement path so that no current (except leakage and bias currents) flow through the measurement conductors.

Thus, the PCB layout (unless the design itself is faulty) is remarkably easy, and I don't expect you'll have any troubles, since series resistance hardly matters.

simulate this circuit – Schematic created using CircuitLab

In the above schematic, the exact value of the resistors R1, R2, R3, and R4 hardly matter, provided they are reasonably low. R4 affects the common-mode voltage the instrumentation amp sees, R1 and R2 affect errors due to input offset current (and noise) a bit, but really the PCB layout is not very important until currents and voltage drops start to become significant wrt the common mode range of U1. So you want to keep the voltage drop across R4 reasonably low (not a problem typically unless you use very thin traces and/or very high currents).

• It's also a good idea to make R1 and R2 about 10k, with a 0.1 uF to ground at each input of the IA. This prevents common mode noise from getting into your high-gain amplifier. If you're feeling paranoid, match both the resistors and the capacitors. 1% is good enough. Use good layout - be careful where your load currents run, because even a few milliohms of unintended current path can give you errors. – WhatRoughBeast Jun 5 '14 at 18:51
• If you follow what @WhatRoughBeast suggests, adding resistors to the line resistances R1/R2, you can put another (bigger) capacitor between the +/- input lines. His suggested values don't have much roll-off at mains frequencies, so matching is likely not critical, but you can put a big fat 1uF cap between the +/- inputs of the instrumentation amplifier without harm (slows down the response a bit, that's all) and it helps to kill noise. All caps preferably film types (or NP0 ceramic). Resistors can be metal film precision types. – Spehro Pefhany Jun 5 '14 at 19:26
• Does trace thickness matter? I have a few measurements that are 1 ohm most are in Meg ohms – MEAM Jun 5 '14 at 20:45
• Not really. If the circuit is good it won't matter much, if it is bad it won't help much. – Spehro Pefhany Jun 5 '14 at 20:53
• Maybe I should have also explained that the measurement equipment I am using for this project is National Instruments. Is the circuit provided above along with all the comments such as R1, R2 and caps to ground for common mode noise to obtain the measurement as accurately as possible without NI equipment or should I also be considering these components on the PCB that I am re-spinning? – MEAM Jun 9 '14 at 19:06