The measurement
A thermometer with no sensor calibration to lose.
Johnson noise thermometry reads thermodynamic temperature from the random electrical noise that thermal agitation produces in any resistor. Working with NPL, we developed the first version practical enough for use outside the laboratory.
Resistance changes. The temperature reading does not.
The sensor below is held at a constant temperature while its resistance is varied to represent the effects of contamination, oxidation or transmutation. The raw noise signal responds to that alone, because resistance is in the Nyquist relation.
The temperature readout stays constant as the sensor’s resistance changes.
Why resistance changes do not change the temperature reading
–– °C
temperature · stable
noise RMS
resistor → differential amplifier → measured noise
The rest of this section
- What it isStart here if temperature metrology is not your fieldThe idea in plain language, without the maths. Why a thermometer that does not rely on sensor calibration is fundamentally different from one that does.
- How it worksThe measurement principles and equationsNyquist, the ratio method, cross-correlation between two amplifier chains, and the frequency-domain separation of calibration tones from the noise itself.
- Development statusPublished results and programme milestonesA clear timeline of the programme, with every milestone, date and figure linked to the supporting paper or project record
- EMC performanceImmunity testing for industrial useElectromagnetic interference is what stopped every previous attempt at a practical noise thermometer. Here is how ours behaved when an accredited laboratory tried it.
- Probes and sensorsWhat this means for probe designBecause we measure the resistance rather than rely on a fixed resistance-temperature relationship, the probe no longer carries the temperature calibration. It still needs a suitable resistance and a sufficiently short cable.
News and programme milestones are posted on LinkedIn.