[R-390] R-390 Power Supply update; receiver voice testing

[John Wendler]

First, thank you for the warm welcome!  I had some e-mails both on and
off-list - it is great to be surrounded by friendly and encouraging people!

Update on the power supply, code name: Crispy Critter.

I unsoldered enough connections to pull off the shelf.  The transmformer is
still good, and was putting out reasonable voltages.  Some of the teflon
wires on the shelf had burned so badly that they had shriveled (!)
insulation.  As best as I can figure, an open flame torch was applied to
this unit to install the solid state diodes.  This would account for the
condition of both teflon and enameled wires, and would also account for the
unnaturally dull finish on the tube socket pins and other solder joints.
 (Never having seen a fresh power supply, I am assuming that some wires are
enameled magnet wire.  I have been measuring the diameter with calipers to
estimate the AWG.)  Unbelievable.

================

Testing with voice signals is tough, and ultimately it is subjective.  At
one point I was involved in voice quality testing.  One test involved stock
recordings of different men and women reading particular phrases.  The idea
is to have a large number of people listen to the recordings through a
receiver and score.  Afterwards, you apply some statistics.  (Think of an
older and slightly deaf man listening to a soft and breathy soprano woman -
you want to make sure that score is diluted by others who can make out her
consonants)

I have also performed VHF / UHF radio receiver testing per TIA-603
(conventional FM) and TIA-102 (P-25)  There are procedures for setting the
voice level input at the transmitter.
In general, Analog RX sensitivity tests were defined with a 60% of peak
modulated 1 KHz sine wave at the transmitter and if I remember correctly, a
12 dB SINAD at the receiver.  Preemphasis and deemphasis per particular
curves.

I have not done formal AM or SSB testing professionally, so no direct
experience with any standards, yet.  I would research what standards exist
for AM broadcast and AM Aircraft Receivers as a starting point; I would
also examine the R-390A alignment and final acceptance test procedures.
 (Sorry, haven't gotten there yet.)

There are the usual NF / MDS / IP3 measurements.  Once you get to modulated
tones, you have to define things like percentage modulation on the
transmitter, etc.

.wav recordings into an external modulation input can be a great way to try
different tests rapidly - I did some of this, too.  Discrete tones, swept
tones, bandlimited noise, all repeated at several different levels.  Noise
power ratio tests are another way of getting at broadband distortion.
 Don't forget that the TX output has to be clean and in-channel, and your
levels into your recorder should not be driving distortion or falling into
the noise floor there, either.  Watch out for ground loops and noise
coupling; audio isolation transformers can be your friends.

Standardized testing is useful because everyone can independently pull the
standard and try for themselves.  In this kind of testing, it is most
important that everyone agrees to what the standard is, and much less
important that it accurately characterize some rare confluence of
conditions.

Per an earlier comment - QRN and QRM as filtered by the operator's brain is
the ultimate test.

Probably more than anyone wanted to hear...

73 de N5CQU