[Microwave] Building a noise source/head

[Dr. Gerald N. Johnson, electrical engineer]

Keeping the capacitor really small (leadless disk or tiny chip)
physically and with a decent value for the low frequency end works
according to the articles I've read. The rub with the solid state diode
source is that its not predictable for noise and its rarely consistent
from one band to the next. Its reasonably consistent from one day to the
next.

For use with low noise preamps, one runs the diode at 25 dB excess noise
or so, then uses a precise attenuator to hide the impedance changes of
the diode and its mount with frequency and with bias changes, as the
diode is switched its impedance changes. 20 dB of attenuation is a good
start at that impedance change swamping. Sometimes more is needed. The
at home calibrating the noise source probably takes something like
comparing the detected NF for a test preamp, to that measured with the
cumbersome hot/cold loads technique that is precise and calibration can
depend pretty much on the thermometers at the loads, PROVIDING the Z of
the loads match closely at their operating temperatures. There's not
enough noise difference to swamp those impedance variations with a large
attenuator. One ends up with different currents for each frequency, or
different noise heads.

I have an old coaxial vacuum diode noise head, but its limitation is the
termination.

Most preamp inputs have either a series matching capacitor or a shunt
inductor to eliminate the DC from the gate. Some have several of each.
The PI matching network is not very good in the real world of many
signals. Its basically a peaked low pass filter and accepts EVERYTHING
below the band of interest and as the device gain is nearly always
greater at lower frequencies the PI matching network soaks up pagers and
cell phones and all that stuff to keep the preamp saturated. With the
growing uses of RF, VHF, and microwave, our preamps must have
SELECTIVITY at their inputs, or they are useless for anything but NF
contests.

73, Jerry, K0CQ

-- 
Entire content copyright Dr. Gerald N. Johnson, electrical engineer.
Reproduction by permission only.