Hi Mike,                                                             1/2

     About "89.3 FM" of June 26:

MR> ...UHF dx from Burlington Vt. to Montreal...  ...excellent in winter
MR> time but is very degraded in summer...  ...returns at night.

     I'll confess my expertise in fields like electromagnetic dispersion
is rather limited, maybe my opinion that FM reception is better at night
happens to be innacurate and involves a more complex compound of factors
as weather combined to interference from daytime activity.  In the years
when i had a simple amplified antenna installed on the 3rd floor balcony
of a building at the corner of Henri-Bourrassa and St-Laurent, channel 3
seemed to propagate as differently from 97.7 Mhz as the later FM channel
could differ in behavior compared to channel 57 (which hardly gave me an
impression of being remote at all while snow usually burried channel 3).

     That's as much as i can share about dispersion!!!  When it comes to
rejecting strong adjacent signals, on the other hand, i recall searching
thru the PolyTechnic's `IEEE' collection for days and, one thing leading
to an other, i retained some lessons in the process:  no matter how good
a pre-amplifier is, stay clear of trouble by rejecting strong signals in
order to avoid an overload situation, for example...  ;-)  One other tip
which had its roots in the era of RADAR development came to me as a pure
revelation:  TV "L-C" filters are a bit innefficient compared to helical
resonator designs, especially when working with the upper ~VHF~ spectrum
up to the ~UHF~ frequencies;  an article discussing ideal geometries and
which was based on an abacus showed that High-Q "helical" filters can be
built out of simple copper HardWare bought in local stores - i could try
to figure out the ideal set of inductor/capacitor values for my TV "L-C"
filter until the end of days and never get better performance!  One more
lesson was about how to optimize/reduce the lossy components in a tuning
capacitor:  forget the ones with low-grade dielectrics, pick saphire for
a compact design - silver plating/solder being the finishing touch.  The
typical hobbyists had to learn how to handle a gas torch but the rewards
were far greater, an "L-C" design at FM frequencies most likely required
to be based on toroidal High-Q ferro-magnetic cores, which also involved
finding ideal geometries but what catalog makes this a simple chore?!...

     Well, Joe doesn't have to go thru this, anyway;  MFCo did the works
already and i read their literature...  Lets now refer to it right here:

          "Whenever the desired and undesired FM stations are adjacent
     channels (separated by only 200 Khz), even cavity type traps are
     not usually effective.  In this case, the objectionable FM station
     can be suppressed by the phasing method, which preserves the
     desired, adjacent channel.  A ``co-channel'' eliminator is used.
     This device consists of a variable phase shifter in series with a
     variable attenuator to permit altering both the phase angle and
     amplitude of the offending FM.  As shown in the diagram, a separate
     antenna is pointed at the unwanted FM station.  Its signal is run
     through the ``co-channel eliminator'' and into a tap on the line
     connecting the main FM antenna and receiver.  By adjusting the
     phase and amplitude control on the ``co-channel eliminator'', the
     unwanted FM station can be suppressed at the receiver with minor
     reduction in the adjacent FM station signal." - "Taming The Strong
     FM Station"/"TV Receiver Overload", an unsigned 4 pages
document...

     I thought of joining my ~ASCII~ drawing version of the diagram they
used in their paper but i guess we can live without it.  A few words are
probably enough to describe the setup:  two antennas with some different
orientations, one which acts as the interference reference while the 1st
one captures the desired station.  Both signals get combined once we got
adjacent carrier signals properly prossessed.  That's it...  An optional
BandPass filter can be inserted right after the interference antenna and
the processed signal may go thru an amplifier, if needed.  The offending
content is nulled out in a directional coupler, partially or completely.

     It's been awhile since i contacted Microwave Filter Co., my catalog
and literature samples are dating back to the spring of 1988 but sharing
the folowing information may help Joe in getting an idea of what's ahead
and how much money he should expect to spend over this if he chooses to.

     Lets begin with a simple situation where the undesired channels are
at least 2 Mhz away.  In 1987 Joe could have bought MFCo's Pre-Tuned and
Sealed `5KFM-(Mhz) Selected FM Channel Trap' for 20 $ U.S. - two of them
since there are two stations to reject and MFCo asked for minimum orders
of 25 $, anyway.  The interfering carriers would be suppressed by 60 dB,
no expensive "phase out" setup appears to be called for at this stage...

     Their `3634-FM Sharp Tunable 88-108 Mhz BandPass Filter' would work
fine if the adjacent channels are located something like +/- 600/700 Khz
appart from the desired FM channel and an "insertion loss" of about 5 dB
would be acceptable.  The 3 dB bandwith would be between 200/400 Khz and
30 dB suppression of adjacent channels could be garanteed - depending on
the order:  standard vs custom-built.  The filter sold from 417 $ to 483
$ U.S. in 1987;  it could be factory-tuned for a slight 20 $ supplement.

     The more costly alternatives based on "phasing out" can still wait.

                                  ;-)

     Now, the piece of resistance...  This setup doesn't include the set
of components required around the filter like the pre-selection BandPass
filter and processed-signal amplifier (both are optional), nor what MFCo
called a `DC-8' directional coupler (a "balanced quadrature network", as
i recall - nothing to do with a directional antenna) and which is at the
heart of the concept.  Everything but the additional directional antenna
could be integrated into one complete a$$embly when needed, i believe...

[The present message concludes next...]
--- Platinum Xpress/Win/WINServer v3.0pr5a