Refrigerated Microwave Noise Sources

The accuracy of microwave-amplifier noise measurements is affected by 1) the uncertainty of the temperatures of the reference noise sources as seen at the amplifier input terminals, 2) the error in reading the output noise power from the amplifier, 3) mismatches between reference noise sources and the amplifier, and 4) the uncertainty of amplifier gain, aside from short-time gain drift. The influence of all these errors is evaluated which leads to the conclusion that the accuracy of measuring effective input noise temperatures below 60?? K is greatly improved by employing a refrigerated microwave noise source, preferably at liquid helium-temperature. Finally, such a refrigerated noise source that uses WR-137 waveguide is described with all relevant details for its construction and calibration. The measurements were performed at 5.8 Gc/s. In the Appendix a "universal" error plot is discussed which enables one to predict the best possible accuracy of noise measurements for any combination of reference and noise temperatures.