Upon first reading the article "Predict Mixer Noise Behavior" (January 2012), I did not see any significant problem in the author's mixer analysis. But the author's conclusionsand mainly Fig. 8, where the result is that system noise temperature or noise figure must be higher than 3 dB, in any caseis nonsense.
Later, I addressed the author with my comments. I argued that in the real world, the Friis equation clearly explains how the gain and noise contributions affect the system noise temperature and noise figureand more, that in my long experimental work I never saw any deviation from the Friis formula. Mr. Monzello, however, sticks to his statements, though he gave no derivation or explanation how he reached the wrong result. He is a software author who creates models for engineers to use in receiver design. His responsibility, however, lies in that his wrong statements would be read by students and engineers who search for reliable information to use.
Why would so many people around the world develop and manufacture microwave and mm-wave low-noise amplifiers if, in any superheterodyne receiver, the result will be as bad as Mr. Monzello insists? He promises that, whatever one does, the system cannot achieve a noise figure lower than 3 dB. Those who know better can wonder: In GPS and satellite TV receivers, noise figure typically is less than 1 dB and the receivers are SSB superheterodynes, with mixers and low-noise preamplifiers.
The last sentence in Mr. Monzello's paper reads: "For small LNA gains, the system noise figure will improve. In the limit, as the gain of the LNA goes to infinity, the system noise figure will be equal to the noise figure of the LNA plus 3 dB." I am sorry, but this statement is plain wrong. I think there will be more comments to Mr.Monzello's paper. This confirms my deep concerns of software models which are presented as correct, without any reference or responsibility.
Jiri Polivka
Spacek Labs
Santa Barbara, CA