Bethesda
is what is known as an 'Active Deflector' - the signals from the parent
station are simply amplified and re-radiated, without any frequency
change being involved.
In order for this rather dubious practice to succeed,
a number of conditions have to be met:
- A
good aerial installation.
Consider an 'ordinary' relay station. The receiving aerial, looking
at the parent station, is usually mounted on the same mast as the
transmitting aerial - often quite close. Hence, there is a lot of
transmitted signal picked up by the receive aerial. This doesn't matter,
for the first item after the receive aerial is a filter which accepts
the parent signal but rejects the transmitted one.



Unfortunately,
on an active deflector, the input filter has to be tuned to the same
frequency as the output signal! This means that output signal rejection
has to be done some other way. There are several steps, all of which
will probably be necessary:
- You
need aerials which have a good front-to-back ratio *
- You
need to install the receive aerial in such a position that the transmit
aerial is (more or less) well behind the receive aerial.
- You
need a good physical separation between the two, to take advantage
of the inverse square law.
- You
use cross-polarisation of the aerials - this can give you an extra
8dB of rejection at the relay station site, and of course benefits
viewers in the target area as well.
*
even a 'good' receiving aerial, such as the GD2000 log-periodic
aerials customarily used on relay sites, are nothing like good
enough. The only thing that seems to work is the standard trough-aerial
used for RBL and RBS links on main stations. These have a solid-metal
back and are exceedingly well screened.
If
the overall isolation between aerials is too low, the picture
becomes degraded with 'ringing' and ghosting effects. If it
is much too low, the entire installation oscillates in an uncontrolled
manner much like a badly-installed public address system.

- A
good amplifier
The original units used at this site were an adaptation of the BBC
EP7/513 transposer (colloquially known as the 'wine press', due to
its mechanical layout). This was a unit not to everyones taste. In
these units, there was often not an IF at all: the incoming signal
frequency was simply mixed to the output frequency using a mixer and
local oscillator. Occasionally, to get proper channel offsets, double
mixing was used with an extra mixer and oscillator. The important
item to note was that at all times the signal was at (some or other)
UHF TV channel. In the active deflector version (not coded /513, but
differently. /520? I can't remember) the mixer and local oscillator
units were simply omitted, and the signal blew through the unit unimpeded
at the same frequency. When you stop and think that the input to the
amp. was probably about -40dBm and the output +33dBm, you needed between
70 and 80dB of gain in the amp, all at the same frequency. Hence,
the units making up the amp. had to be exceedingly well screened to
avoid instability. I think it fair to say that the original units
barely made this requirement, and increasing age and corrosion made
things worse. The original units were replaced by more modern designs
by the time S4C came along.
- The
right site
Obviously, the receive aerial needs a good view of the parent, and
the transmit one of the target viewers, but in addition there must
be no objects in the vicinity which, illuminated by the transmit aerial,
can reflect back into the receive aerial. This implies a clear and
(preferably) down-sloping site towards the target.
In addition
to this, the viewers need to be in a favoured location. They must
be screened from signals arriving directly from the parent station,
else ghosting will occur on a majestic scale.
The
Bethesda site meets most of these requirements. The target audience
is (mostly) in a slight side-valley, screened from the parent by a
rock hillside. In addition, cross-polarised signals from the parent
station arrive at the target aerials at about 90° to the forward
beam, so there is additional rejection due to side-nulls.
However,
there are areas of the town that do get too much signal from the parent,
and where the relay is not usable. Unfortunatey, the parent is not
useable either, due to the strong signal from the relay. This is the
reason for the provision of the Bethesda North relay, which uses conventional
channel-changing tactics.

I
often wonder why they made Bethesda an active deflector. The usual reason
is that there are no available channels for a conventional relay. This
is manifestly not the case here: I think they did it just to see if
it could be done! The site is almost ideal, and such favourable geographical
factors are not very common |