Some sources of error in microwave milliwattmeters

An approximate analysis is made of the factors that influence the intrinsic accuracy of a heated filament used as a transfer device for measuring power at microwavelengths in terms of a low-frequency calibration. It is shown that instruments incorporating such filaments are usually subject to errors of significant magnitude if the filaments are longer than about one-tenth of a wavelength, and that the errors are usually smaller in resistance milliwattmeters than in thermocouple instruments used in similar conditions. Experiments are described which show that serious errors, due to the inefficiency of the transformer used to match the power source to the filament of the instrument, can occur in addition to those described in the analysis. It is concluded that filament instruments cannot, in the present state of development, be relied on as standard transfer instruments for measuring microwave power in terms of a low-frequency calibration.