----- Original Message -----
Sent: Sunday, August 08, 2004 2:30
AM
Subject: LF: AR88LF on 136kHz
Dear LF Group,
I bought a rather battered AR88LF at a rally a
few years ago - This version of the AR88 has an LF band covering 75 - 200kHz,
so it is potentially useable for amateur LF. I had not really tried it
seriously, so with all the pining for good olde days that has been going on
lately, I dragged it out of the shed to see what could be done with
it.
After replacing the usual catalogue of leaky
capacitors, temporarily bypassing a broken IF transformer, and peaking up
the IF amplifier, the sensitivity was good - 0.3uV gave a good audible signal.
The input impedance on 136k is quite high, something of the order of 1kohm.
Connecting a 200pF capacitor in series with the sig gen to simulate the source
impedance of a long wire antenna only reduced the signal level by a few dB, so
it should be possible to connect the RX directly to such an antenna and get
good sensitivity without additional matching components (allthough this does
not work at my QTH - you get about 100V of 909kHz from Brookmans Park at the
RX input...)
The selectivity was poor however. The IF
frequency is 735kHz, so one can expect worse selectivity than other old RXs
with 455kHz IF. This meant that when tuned to 136kHz, large signals from DCF39
and the Datatrack beacons just outside the band pass through the IF and
intermodulate together at the detector, producing lots of heterodyne whistles
even with the BFO switched off, as G3GVB mentioned. It has a single crystal
filter, but this did not seem to make much difference. However, this proved to
be due to mis-alignment - I think every receiver with one of these
single-crystal filters I have ever encountered has been wrongly
aligned. Partly it is due to ageing I'm sure, but also I expect a lot of
over-enthusiastic fiddling has gone on over the years! The main problem
was the 'phasing' trimmer, which was a long way from where it needed to be.
After a lot of re-tweaking, the -3dB bandwidth on the narrowest position
was only about 200Hz. This proved good enough to get rid of the worst of the
whistles, however, the skirt selectivity is still not good enough really -
with the RX tuned to 136.5kHz, the best I could get was about 40dB rejection
of DCF39 on 138.83kHz, which still gives quite a loud whistle above the band
noise.
So I decided to graft on an audio filter from a
CR100 receiver, which happened to be lying around. I connected this between
the 1st audio stage and the audio output stage; allthough really just a
lash-up it works quite well. I also increased the amount of BFO injection. On
the AR88 the BFO signal is injected into the final IF amplifier using just
stray capacitance between the wiring. I added an extra bit of wire to
increased the rectified voltage at the detector from 4V to about 15V. The idea
of this was to make the BFO signal much larger than the unwanted signals at
the IF output, and so hopefully reduce the amount of audible intermodulation
products produced - It seems to work. It does mean that the AGC has to be
switched off, otherwise the BFO signal causes a big reduction in gain. The
best way to set the controls is with the AF gain set to as high a level as
possible without producing excessive receiver noise, and then controlling the
gain with the RF gain control. This ensures the lowest possible signal levels
at the detector, reducing the likelihood of IF or detector being overloaded by
strong out-of-band signals.
The end result works quite well. This morning,
when the QRN was relatively low, it was possible to hear the Loran chatter
above the band noise, with only faint whistles from out of band stations, so
sensitivity ought to be adequate for most things. The warm-up drift is a few
100Hz, but after it has been running for a few hours, the frequency only
wobbles about by several Hz, so in fact it is adequate for QRSS3 reception.
When I get round to doing a proper job, I will add a product detector
(probably using the socket of the not-very-useful noise limiter valve). Also,
the CR100 audio filter is really too narrow - with 100Hz bandwidth, it is
quite hard to get the BFO pitch just right so that the peak of the IF response
coincides with the peak of the audio response, and it has to be re-adjusted
due to drift of the BFO. A filter with 200 - 400Hz bandwidth would be better.
Obviopusly, an AR88LF is not as good as a modern
receiver. But with some simple mods it is certainly usable. I think mostly the
same considerations apply to other old equipment, such as the HRO and CR100
when used on LF.
Cheers, Jim Moritz
73 de M0BMU
