[CQ-Contest] band-pass filtering -- a new approach?

[Maurizio Panicara]

I think that a passband filter on the TX helps no better than a notch filter
in normal conditions, I mean with a good transmitter and a good amplifier,
but may affect the overall reliability of the station since discrete
components are to withstand high power levels.

A normal PA with a simple PI network and reasonable Q will attenuate the
second harmonic more than 40 dB, probably 45 to 50 dB beeing typical, one
with a  PI-L can be at 60 dB rejection.

A good notch filter made with low loss line in the form of a shorted stub
easily adds another 35 dB attenuation, and if one want to have huge
attenuation more than a single stub can be used. Best stubs are those with
highest Q and lines with higher Q are those with air dielectric or foam. The
difference with polyethilene is quite appreciable as the used frequency goes
up.
It's proper to focus that the ultimate harmonics rejection is also functon
of the transmitting antenna that's in some cases is a good filter itself.

Incidentally, I'm not a supporter of open stubs because of the high voltage
involved at their ends.

On the other hand, filtering at RX edge is quite critical and when levels
are quite high some intersting effects take place.
During the last WWDX we had problems to receive on any band when
transmitting, on 40m in spite that we used extra RX filters that had more
than 70 dB rejection.
At first, this effect was supposed to be a transmitting problem but later,
replacing the first set of filters with another one the problem disappeard.
Apparently this was very funny because the filters working properly had less
attenuation and a worse shape factor while the first ones were absolutely
great performers under the analyzer and the tracking generator.
In practice, anyway, the second set of filters is worse in terms of absolute
attenuation but had been designed to withstand 200 W, has a good return loss
and really behaves as a linear device in all cases. The first filters were
instad typically an RX device (a popular design)  and, although theorically
able to manage a few watts, something (may be the small coil cores or some
of the capacitors) did not behave totally linear and caused the problem.

In a case like this, the biggest out of band rejection of a filter placed
after a PA does nothing because an high RF level on the fundamental is
overloading something at the RX edge. The same is valid when external non
linearities (i.e. because of rust)
do generate troubles (ghost signals) using the fundamental frequency or
frequencies.

73,
Mauri I4JMY






----- Original Message -----
From: "Guy Olinger, K2AV" <k2av at contesting.com>
To: <cq-contest at contesting.com>; "Pete Smith" <n4zr at contesting.com>
Sent: Friday, November 09, 2001 6:10 PM
Subject: Re: [CQ-Contest] band-pass filtering -- a new approach?


>
> The lack of filtering the transmitted signal will bother you quite a bit,
> and I would respectfully propose that this disadvantage will greatly
> outweigh a
> hundred watts on the output. Eg, a fraction of a db versus wiping out weak
> received signals on SO2R
>
> At NY4A CQWWCW multi this year, I'm going to add a HIGH POWER filter after
> the amp on 40m
> to cut down on the interference to 15. Probably could do the same with
> *properly* installed stubs, but that is the rub, getting that right, so
will
> cheat by spending some money for the high power 40m bandpass.
>
> By any angle I can think of, SO2R invokes all the cross interference
> problems of multi-multi except nX +- nY intermod, and except you only
annoy
> yourself, and maybe that's
> worse. I'm betting there's no way you can cheat or finess the problem.
>
> GL & 73, Guy.
>
>
> ----- Original Message -----
> From: "Pete Smith" <n4zr at contesting.com>
> To: <cq-contest at contesting.com>
> Sent: Friday, November 09, 2001 10:12 AM
> Subject: [CQ-Contest] band-pass filtering -- a new approach?
>
>
> >
> > I've been researching band-pass filtering for SO2R, and it has occurred
to
> > me to try a different approach if the peer review sounds favorable.
> >
> > What I'm planning is to build a box for each of my transceivers,
> containing
> > a receiver protection circuit (diodes and a resistor), receive-level
> > bandpass filters and a relay switching matrix to select the filters AND
> > bypass them on transmit.  This would seem to offer several advantages --
> >
> > I can use W3LPL's 3-pole receive filter circuit, which uses standard cap
> > values and delivers performance that is probably superior to any
> commercial
> > alternative except for W3NQN's expensive transmit filters.
> >
> > I avoid issues with burning out the filter components from transmit
power,
> > particularly with a messy amplifier input SWR on some bands.
> >
> > I avoid the reduction in drive to my SB-220, which really needs its 120
> watts.
> >
> > I can be fairly relaxed about the physical layout, since SWR is not the
> > issue it is on the transmit side.
> >
> > I realize that having a bandpass filter on the transmit side offers
> > advantages too, in particular reducing harmonic output and wideband
noise.
> > But on balance, this seems like a good idea, plus it'll give me a
project
> > for the winter.  Comments?
> >
> > 73, Pete N4ZR
> > www.qsl.net/n4zr


--
CQ-Contest on WWW: http://lists.contesting.com/_cq-contest/
Administrative requests: cq-contest-REQUEST at contesting.com