microK 250 and microK 500

The objective we agreed with Isotech was simple to state and hard to meet: a precision thermometer for secondary standards work that was the best in its class, on every parameter, with the fullest set of features.

Meeting it meant building a new analogue-to-digital converter. The instruments working at or below 1 mK were either chopped-DC potentiometric designs or AC resistance bridges, and we wanted the microK to handle thermocouples as well as PRTs and thermistors, which ruled the AC bridge out. That left DC, and DC instruments of the day used integrating converters whose linearity is limited to a few parts per million by dielectric absorption. The best commercially available part managed 3 ppm. Good, but not good enough for what we had in mind, and correcting for a converter’s limitations in software was not an approach we were willing to take.

So we designed our own, working with NPL and Metron Designs, and reached a linearity better than 0.5 ppm. It is a multi-bit adaptation of the sigma-delta technique, and every instrument in the family since has been designed from it.

Two other decisions from that project have stayed with everything since. We used a substitution topology, switching the device under test and the reference into the same measurement point instead of passing a common current through both, which removes common-mode error and gives a genuinely drift-free measurement. And we switched the signals with solid-state devices instead of mechanical relays, which are the part of an instrument like this most likely to fail.

It stayed in service for nearly twenty years. Isotech reports it being used for ITS-90 dissemination, in ISO 17025 laboratories, and in work as far from a calibration bench as oceanography, climate research and the astronomical detection of exoplanets. It was retired in 2024 when the microK² family was completed.

Notes from the microK design team, 2006