Skip to Main Content
This paper addresses, from an experimental perspective, the long-standing question on whether the 1/f noise of microwave semiconductor devices should be considered cyclostationary in compact models. By using a simple instrumentation setup and basic mixer concepts, it will be shown that the external equivalent current noise sources of such devices simply cannot be stationary. To demonstrate our ideas, a through experimental analysis was carried out on purely-resistive bridge circuits made up of microwave varactors and transistors under large-signal operation whose DC current component is kept constant. When computing the voltage noise power of specific circuit arrangements, errors in excess of 2 0dB may be induced if the stationary concept is adopted. The difference between the stationary or cyclostationary concepts will be simulated with the aid of the simulator ADS from Agilent.