Comment on “A Look at the Concept of Channel Capacity from a Maxwellian Viewpoint” [Letters to the Editor]

In the feature article ldquoA look at the concept of channel capacity from a maxwellian viewpointrdquo by Sarkar et al. [1], the authors compare Shannon with Hartley-Nyquist-Tuller, but say that they are unsure (when the two formulas give different numerical values) which is the correct one to use. They should use the one that matches the physical situation. If the system is noise-limited, then use Shannon. If the system is limited by quantization, then use HNT. Their uncertainty may partly arise from the similar numbers they happened to obtain for their matched receiving antenna (pp. 30,31): I believe this was probably just a numerical coincidence. Their statements about the relationship between power and voltage depending on matching are true, but trivial. If they used the voltage across the resistive part of their matching impedance (2.77 mV) instead of the total voltage (i.e., use external matching), then their matched HNT figure becomes 15.0 B instead of 15.3 B. A severe mismatch from a very high input impedance would give them the full open-circuit voltage from the antenna (5.53 mV) and an HNT figure of 17.0 B. Their calculations do not take account of the effect of matching on noise; sometimes, a deliberate small mismatch can improve SIN ratio.