The correction factor of the National Institute of Standards and Technology's WR-15 (50 to 75 GHz) microcalorimeters has been re-evaluated. Use of modeled values for the reflection coefficient of a flat short and offset short instead of measured values results in lower uncertainties in the evaluation of the calorimeter correction factor obtained with these artifacts. The uncertainties are now comparable to that of an evaluation technique in which two calorimeter cores are directly connected to each other. Improvements in a fourth technique were also achieved with an improved measurement of the transmission coefficient of a line section. The differences in the derived correction factors are a little larger than would be expected based on the uncorrelated uncertainties. The source of the discrepancy has not been determined.Index Terms -Bolometer, Calorimetry, Millimeter wave measurement, Power measurement
I. BACKGROUNDThe primary power standards between 50 MHz and 75 GHz at the National Institute of Standards and Technology (NIST) are microcalorimeters that are used to evaluate the effective efficiency of bolometric power sensors. The microcalorimeter measures temperature changes in response to direct current (DC) and radio, microwave, or millimeter-wave power (generically referred to as RF). This response is expressed aswhere e is the microcalorimeter thermopile voltage, k is a constant that depends on the thermal properties of the calorimeter, P DC and P RF are the DC and RF power absorbed in the bolometer device under test (DUT), g is the calorimeter correction factor and g ≅ 1. g deviates from 1 due to the different response of the microcalorimeter to RF and DC inputs. The most important contribution comes from RF losses in the calorimeter transmission line leading to the DUT. Uncertainty in g is an important, if not dominant, component of the uncertainty in an effective efficiency measurement. Special measurement techniques are used to evaluate g by arranging the apparatus so that the calorimeter transmission line losses and/or losses in the input section of the DUT can be isolated from the main losses in the bolometer resistor. Previously, the correction factors of a pair of WR-15 (50 GHz to 75 GHz) microcalorimeters used as standards at a U. S. government work, not subject to U. S. copyright NIST were evaluated with four techniques that were labeled Through, Line, Foil Short and Offset Short [1]. The standard (k = 1) uncertainties achieved with the four techniques were a weak function of frequency with typical values of 0.001, 0.003, 0.004, and 0.006 respectively (see Fig. 6 in [1]). The techniques use artifacts similar to those in a vector network analyzer (VNA) calibration, but the measurement process itself is different. The largest contribution to the correction factor uncertainty in the Line technique was the measured value of the artifact's transmission coefficient |S 21 |. Similarly, the largest contributions to the correction factor uncertainties in the Foil Short and Offset Short evaluations were th...