Proposed in-parallel diode design

[Paul]

I have a hypothesis that the reason diodes in direct parallel
connection produce less voltage is due to the fact that parallel noise
sources decreases. Since the diodes depletion width changes with the
net applied potential across the diode, this diode model is in
agreement with measurements. Diode noise sources in parallel are
relative to the inverse of the square root of the noise sources. So
the net noise across 2 diodes in parallel is 1/sqrt(2) ~= 0.707 times
less noise. Due to the diode square law, the net rectified DC voltage
from 1/sqrt(2) less noise equals 1/2, or half the DC voltage.
Therefore, the DC voltage is relative to Rz, which is in agreement
with diode experiments.

The proposed new design for diodes in direct parallel connection is to
place an inductor in series with *each* diode. So if there are 10
diodes, then there would be 10 inductors. Each inductor will hinder
some of the AC noise *between* diodes, but it will not hinder the
noise that each diode produces on itself. The noise on a single diode
would be what is known as kTC noise, which is equal to sqrt(k * T /
C).

Paul Lowrance