[CQ-Contest] Bandpass Filter

Bob Henderson bob at 5b4agn.net
Mon Jul 9 05:52:17 PDT 2012


An interesting thread.

A single tri-fed C31XR has been used with significant success in both SO2R
and M/2 scenarios at P3F for 8 years now.  The single tower installation
supports a C31XR, an XM240 and an 80m inverted vee dipole. The tower itself
is shunt fed on 160m.

More recently a second tower with additional antennas has been added but
configuration of the first tower remains broadly as it was initially
configured.

In any mult-transmitter installation it pays to carry out some kind of
cross coupling audit before battle commences so as to avoid unwanted
equipment damage.  Such an audit can become very demanding where precision
is sought and in these cases will be well beyond the measurement capability
of most contesters.  Where precision is required this may best be left to
the techies among us.

As contesting amateurs what we need are some broad metrics which will
safely allow us to configure an SO2R or M/M station without risking
equipment damage or suffering from cross band interference.  We want a KISS
approach to the issue, consistent with the tools available to the typical
contesting ham.

An antenna cross coupling audit which measures the power dumped into a 50R
load using a simple shack power meter does not come even close to providing
scientifically valid measurements BUT it nonetheless provides a useful
measurement metric and one within the capability of any ham to carry out.
Using this approach when I installed the C31XR 8 years ago the following
results were found and as I recall published to this reflector at the time.

*20m driven then 15m -32db and 10m -40db**15m driven then 20m -12db
and 10m -25db**10m driven then 20m -16db and 15m -32db*


-12dB dumped into a 50R load attached to the coax feeding the 20m yagi when
the 15m yagi is driven translates to 100W cross coupled when the drive
power is 1500W.  It is easy to undermine the validity of this measurement
on the basis that the receiver switched to 20m and on the end of the 20m
coax will not present a 50R load to the 15m cross coupled RF.  It can
further be undermined by drawing attention to the fact that neither will
the 20m DE source be 50R so far as that 15m energy is concerned.  With
neither source nor load impedances valid we might as well give up and stay
single op.

I don't think so.

Seeing that your 1500W on 15m can dump 100W into your 50R load at the end
of the coax feeding your 20m yagi is very useful information.  Impedances
invalid or not, this is a pretty solid clue that if you are daft enough to
attach your 20m receiver to this coax you can wave it good bye.  It will
reply with smoke signals.

CONCLUSION

While the measurements detailed above lack scientific validity they
nonetheless provide a vitally useful insight.

A simple metric I adopted early on and which has served me well is:

"No more than 4W cross coupled from any one antenna to any other one using
the above procedure."  *** This value assumes external BPFs will be used
between transceivers and associated amplifiers.

Coaxial stubs are the approach I recommend is used to ensure the 4W figure
is achieved.

When it comes to coaxial stubs location of the stub is critically important
where maximum attenuation of unwanted is required.  Much has been written
about this elsewhere so I won't dwell upon it here.  However, it is
questionable whether peak performance of a stub need be achieved.  An
average ham with a small shack will likely have concerns over just how much
coax cable he has room to stack up there.  My point being that an
imperfectly located stub will easily manage the 14dB rejection required to
drop the hostile 100W down to 4W.  The highest rejection which can be
achieved is clearly desirable.  Throwing away benefit without good cause is
indeed folly but let not the pursuit of the perfect become the enemy of the
well good enough.

Stubs serve two puposes and your investment in them should take account of
this.

1.  Management of hostile levels of cross coupled fundamental energy.
2.  Suppression of harmonics.

BANDPASS FILTERS

Initially at P3F we used Dunestar 600 multiband BPFs between each
transceiver and its associated amplifier.  We became dissatisfied with
these for three reasons.

1. We found that even low levels of cross coupled RF could occasionally
cause welding of the contacts in the small filter select relays.

2. We found some filters exhibited high insertion losses.

3. We had occasional capacitor failures.

As a consequence of our experience a 6 band unit based upon 3 pole
Chebyshev filters (aka W3NQN) was designed with plug-in filters providing
for a failing unit to be replaced or repaired while leaving the remaining 5
filters available for use.  Instead of the cheap ceramic capacitors used in
commercially available units, custom manufactured transmitting mica
capacitors were specified.  These use very high grade mica and are
certified iron free.  This is very important in that the presence of iron
which is found in most standard grade mica capacitors is the root of
resistive loss and consequently component failure.  The unit was also
designed to support an internal plug-in BCD band decoder and a set of
antenna relay matrix drivers so avoiding the cost and clutter associated
with external decoder/drivers. Detail concerning this BPF unit can be found
at www.5b4agn.net

The combination of stubs and external BPFs has provided for a safe and
interference free set-up compatible with SO2R and M/2 use.

SO2R WITH A TRI-BAND YAGI AND SINGLE FEEDER?

Is this possible?  Of course it is.  Combatants at WRTC have been doing
this for some time.  Entrants at the last WRTC in Moscow used BPFs and
tri-plexers I built to facilitate this.

Aha I hear you say but they used only 100W radios.  No amplifiers.  Indeed
so but it is entirely reasonable to consider a tri-plexer to achieve this
at the 1500W level.  Never mind the distinctly scary level of cross
coupling in a C31XR with three feeders, how about 2 x 1500W via a single
feeder to a TH7, TH11, KT34 or whatever tri-bander?  I do this with the
TH11 we have at P3F on the second tower.  Isolation is such that 1500W on
20m results in an only just S9 2nd harmonic on 10m.

The message I guess is you shouldn't let the science scare you away.  We
are amateurs we have to be pragmatic in our approaches if we are to achieve
very much at all.

73 Bob, 5B4AGN


More information about the CQ-Contest mailing list