R=2882Ohm at 4.2K

a = {-0.99436994914, 1.18186319148, 0.01906207066, -0.02320427440, 
  0.00190397877}
T[R_] := Exp[-Sum[a[[i + 1]] (Log[(R - 2200)/1000])^i, {i, 0, 4}]]

The still thermometer was recalibrated during Run 52. The U02844 thermometer (typically used as a tower thermometer) was mounted on the still plate near the still thermometer and used as a basis for a new still thermometer calibration. T versus R data was taken during cooling, and when the fridge was at base I took more data points over the standard operating range by heating the still (for the higher temperatures) and switching to one-shot mode (for low temperatures).

Figure 1 : Still thermometer recalibration, operating range	full range

As can be seen in Figure 1 on the left, the older calibration (based on simply shifting the calibration of a similar thermometer, see below) is around 0.15 K too high in the range of interest, which is enough to be misleading since it makes it seem as if the still is operating at 1K Pot temperatures. (If that were the case, there would be 4He in the circulating lines and the pressures at G1 and G2 would get much higher than normal.)

The new calibration curve is

T_still = 1.655 * (R_still/1850 - 1) ^ -1.367

with T_still in Kelvin and R_still in Ohms. This is accurate to around 0.01 K and valid from 0.7 K to 2.0 K. (Likely it is still reasonably accurate at somewhat low temperatures.)

The function above covers the range of interest, but as can be seen in the right half of Figure 1 it is not valid at temperatures above around 3 K. For completeness I define a second curve for higher temperatures, up to 300 K. Below is a piecewise-defined curve covering the full range:

R_still >= 3241.5 : T_still = 1.655 * (R_still/1850 - 1) ^ -1.367 // 0.7K - 2.5K
R_still < 3241.5 : T_still = 0.9626 * (R_still/2242 - 1) ^ -1.147 // 2.5K - 300K

The higher temperature curve is accurate to within about 5% over its full range.

The data points for the recalibration are shown below. The first two temperatures are inferred from the environmental conditions; the rest are cross-calibrated from the U02844 thermometer.

These three thermometers are not calibrated. Since they all are RuOx thermometers, I just scale them to the calibration curve of H35 based on their 4.2K values. Which means make the replacements

Still:     R->R*2882/3049 <==== OBSOLETE, see recalibration above.  
Coldplate: R->R*2882/3083
MC sensor: R->R*2882/3155

in the H35 calibration formula.

U02844 thermometer calibration