[NLRS] Parts is parts

[Dr. Gerald N. Johnson, electrical engineer]

My first thought for a filter is a three rod interdigital filter made of
UG-141 and PC board ends as used to be in the 1296 transverter in the
ARRL handbook. You will need image rejection for the receive converter.

NBFM can be copied by slope detection or by a full blown FM
limiter/discriminator IF strip, but it works great with a SSB receiver
copying exalted carrier. Just like you'd tune an AM signal with the SSB
only receiver. Most modern filters don't slope detect well because of
the quest for steep filter skirts. The higher order modulation products
are normal in FM when the deviation index gets a little larger than for
NBFM. Not a distortion product, more of a frequency modulation product
where the audio bandwidth is spread over a spectrum several times the
original bandwidth. In the most common extreme, 15 KHz of audio spread
over most of a 200 KHz channel in FM broadcast. E.g. where the
modulation index is much greater than 1. The increased bandwidth of the
FM broadcast signal leads directly to a very good post detection signal
to noise ratio, that is very good quieting. The narrower the modulation,
the poorer the signal to noise ratio and the quieting. Commercial two
way users have accepted poorer quieting and poorer audio signal to noise
ratio in order to have more channels. FM two way has been splitting
channel widths and modulation indices for all recorded FM communications
history. One more split and they might as well run SSB.

FM quieting has a threshold effect below which the signal is not
copiable but only a dB or two of signal increase at the threshold can
significantly improve the quieting and thus the S/N. This is most
pronounced at wide deviations. SSB and CW don't have the quieting
effect, (unless the receiver AGC acts to cut the receiver gain), signal
is added to the noise, but that means there isn't a threshold effect in
SSB and SSB can typically be copied 20 to 30 dB weaker than FM (also
helped by the narrower receiver for SSB and thus less noise power
present). Copying NBFM by SSB receiver will give quite a bit more range
than copying it with an FM detector. Notice that on SSB/CW we routinely
make contacts at distances FM only hams consider DX, "BAND's UP!".

The 45 MHz rock probably has three leads. Its a standard frequency for
quartz monolithic filters in crystal cans.

Yes, there ought to be a crystal to mix 45 down to 10.7 MHz. Hard wired,
not in a socket.

The 45 MHz offset is a bit wider than our band, so we have to make the
radio agile or use two radios, one programmed for transmit, and one for
receive.

I did an epay search for manuals, but found none, I did find more
radios, at least three different power levels, probably otherwise
interchangeable with yours. The owners thought they are worth something
however.

I've not looked closed at a Mocom70 PA section though I have a few
laying about that came with some more interesting 450 base stations (250
watts out of a pair of 4CX250B) and I think I have a manual. These were
really cheap, haul 'em off or they go in the trash! Though I got two
sets of PA but only one rack cabinet. There's room in one cabinet for
two PA and two power supplies. Might need to add a larger blower for two
PA to be combined.

Last year I was looking at more modern Mitrek low band for a 6m
amplifier and I planned to bias the PA for linear service. Didn't really
find any 100 watt Mitrek at a price under a buck a watt and I was
wanting a couple of them at least to combine for more power. I fell into
a Mirage 200 watt amp on epay for a buck a watt, needing no conversion.

The 2m amps will be far easier to make linear than the 900 MHz amps
because the grounded base makes it hard to add linearizing bias.
Requires a lot of current to add a negative bias to the emitters.

There has been an "improved" bias circuit for ordinary ham solid state
linears published somewhere, either NTMS or G3SUK if I recall correctly.
That would be what I'd have applied to the Mitrek PA's. The reason for
that circuit is that some of the "linears" aren't terribly linear either
because the PA bias supply is too poorly regulated in the face of large
base currents from RF drive. Motorola RF applications notes have some
better than average biasing circuits too.