A milliohm meter is a very handy piece of test equipment. Most hand-held multimeters cannot measure low resistances and bench meters that can, are usually quite expensive. [barbouri] has shared details of his milliohm meter build on his blog post, and it looks pretty nice.
When using a single pair of leads to measure very low ohms, the resistance of the measuring wires and voltage drops across the various joints become substantial enough to invalidate your measurement. The solution is to use the “Kelvin method” or 4-wire measurement. This involves passing a highly stable current derived from a temperature compensated constant-current source through the unknown resistance, and then using another pair of leads to measure the voltage drop across the resistor, which then gets displayed as a resistance on a voltmeter.
The finished project not only looks good, but is able to measure up to 2Ω with a resolution of 0.0001Ω (that’s 0.1mΩ). The project is originally designed by [Louis] from [Scullcom Hobby Electronics] and [barbouri]’s second iteration adds an improved board layout to the original project.
The schematic is essentially four well defined blocks. A 5 V linear LDO regulator powers the whole circuit. It receives power from a bank of six AA batteries, with a polyfuse in series for over-current protection. A switching regulator, or power derived from a utility supply may affect the stability of the current source. The constant current source is built around a LT3092 [pdf] — a 200 mA two-terminal programmable current source, adjusted to provide 100 mA output. Its accuracy is further improved by using a LT1634 [pdf] Micropower Precision Shunt Voltage Reference. This section also requires precision, low drift resistors to set the output current. [barbouri] spent time measuring a batch of resistors, checking their values over two different temperatures, and then picking the ones which showed the lowest drift. All of this effort translates in to keeping the current source as stable as possible.
When the 100 mA current passes through a 0.1 Ω resistor, the voltage across the resistor will be 0.01 V. Obviously, this needs to be amplified by a factor of x10 to provide the correct resistance reading on the voltmeter. This is done by the INA106 [pdf] — a precision differential amplifier with laser trimmed resistors. The dual power supply for the INA106 is derived using the MAX680 [pdf] — a dual charge-pump voltage converter that provides ±10 V outputs from a +5 V input.
On his blog post, [barbouri] shares all of the design files, including a face plate layout which will fit a standard Hammond enclosure. [Louis], the original designer, does an excellent job of explaining how the circuit functions, and follows it up with another video to walk through the updated version built around [barbouri]’s board.