Please forgive my ignorance, but I don't understand what a "noise head" is.
Are you trying to build a noise generator?
If so, for what bands?
FWIW, I found out many years ago that a very inexpensive Zener diode makes a wonderful noise generator. I built mine for my 144 MHz SSB/CW station.
If you use the older HP 342, more current HP8970A and some other
Noise Figure Meters the source of noise for tests is not integral to
the Noise Figure Meter. Rather a separate module containing a
Noise Diode is used. This allows you to connect the Noise Source
directly to the DUT (Device Under Test). In many labs techs and
engineers call the Noise Source a Noise Head as when it is attached
to the Noise Figure Meter there is a cable from the Meter to the
Noise Source which supplies power. It looks somewhat like a snake
with an enlarged head.
Noise Sources (aka Heads) can use either a vacuum tube diode or
a solid state diode to generate the Excess Noise for testing. The older
HP342 was provisioned to use either the tube or silicon noise diode
with separate connectors for each. In general Noise Figure Meters
are speced to work with Noise Sources which provide Excess Noise
rated at 15 dB nominal. The older 342 had to be calibrated at each
test frequency manually. The more modern HP8970A was calibrated
at each test frequency and the information stored in memory inside
the test instrument.
Tube Noise Sources I was told generally run out of steam in the few
hundreds of MHz range. Enter the silicon Noise Diode. Noise Diodes can
be purchased that are specified up to 110 GHz. There may be others
out there that can produce noise at higher frequencies. I have never
looked into sources with higher frequency specs as my interests were
much lower in frequency.
The Zener Diodes and some other devices such as the Emitter-Base
Junction of some transistors have been pressed into service as Noise
Generators in the past. Generally these devices were used to build
one or two sources in a shop and calibrated within the lab. Often they
did not generate a flat amplitude spectrum across the frequency ranges
of interest as well as the Excess Noise Amplitude generated was not
Enter the productized Noise Diode. These units are optimized to produce
a specified Excess Noise level across a specified frequency range of interest
with a comparatively flat amplitude response. The HP Noise Heads I have
used in the past have exhibited a Noise Level that was flat within plus or minus
0.5 dB from 1 Mhz to 2 GHz.
If you look at the specs for the NoiseCom NC302L diode you will notice it
is designed to cover 10 Hertz to 3 GHz. The supply current is specified as
6 mA with an avalanche voltage between 6 and 8 volts. Both the Noise Diode
and the Zener Diode use the avalanche phenomena but the Zener is optimized
for a specific voltage as compared to the wide voltage spec for the Noise Diode.
Why use a commercially manufactured Noise Diode? Well my lab work
in the basement at home spans from a few MHz upwards to 2500 MHz.
Being the proverbial tightwad I am not in a mood to hand HP (yeah I
know it is called Agilent now, that is an axe I have to grind with Carley
whatever the heck her name is) $4000.00 plus for a Noise Source knowing
dang well the diode is a $25.00 part with a few chip resistors, connectors
caps and a milled housing all totaling maybe $150.00. (OK I ignored HP's
Hope the info helps answer some questions.