This is another of my hacks (the fourth in a row...) for the AOR AR3000
scanner. This time I tampered with the front end board, in order to
remove an annoying click which is heard throughout HF, and is due to
crosstalk between digital command lines and signal input. This click is
heard a) on the line audio output on the DIN-8 connector, which is
unmuted, and b) on the loudspeaker, once my "demuting" mod is carried on
(sorry for the self reference, but...). My mod does not eliminate the
click completely, but almost. Before the mod, a click is heard whenever
issuing a command (this includes turning the tuning knob), throughout
the range 0.1-50 Mhz (remarkably strong below 10 Mhz); afterwards the
click is almost cancelled above 1Mhz, and easily overwhelmed by signal
above 0.5 Mhz. I didn't find a way to do, quickly, better than this. In
this way, the tuning wheel may be spun giving "almost" the feel of an
analogic vfo... You'll enjoy the whhhooshhh on the unsquelched output
during free scan... Also, the click noise won't mask weak signals while
scanning. The other limits of the front end of the AR3000
(intermodulation, blocking, birdies) will now be fully appreciated. One
day I'll buy a real shortwave radio.

The click is due likely to leaking of the digital signal on the command
board, and ultimately to the pcb tracks layout. Sometimes I have the
impression that the AR3000 must have been conceived as a high end
receiver, but projected in a rush with a lot of mistakes - check the
superetherodine conversion pattern, for instance, and you'll find out
that all the range 940-2036 Mhz is sent trough the same prescaler
filter, despite that this implies NO image rejection on some frequencies
and despite that steering signals for additional SHF filters exist on
board...

What I've find to work is the following:

-one signal has to be routed via a shielded wire rather than via a long
 pcb track.

-two capacitors can be added between two digital lines and ground. True,
 you're not exactly supposed to short a CMOS output to ground with a
 4.7nF capacitor, you'd rather make an RC filter to the next input... it
 works so don't bother.

What I'm describing here refers properly to my old 1989ish AR3000. I
have also the schematics of the 1991ish 3000A: the circuit seems exactly
the same, so I suppose the problem.

The mod is easy but tampers on SMD - shall I insist once more that I
TAKE NO RESPONSIBILITY for how you may screw your scanner up?

Just a bit of theory: command signals for the front-end are fed
_serially_ by the cpu, and parallelized by three 4094 shift registers
(IC1-3) connected in cascade. When the cpu wants to change something on
the front end, it sends a string of command bits, a clock for moving
them across the registers, and then a strobe pulse to make the outputs
active. We'll put a capacitor on the clock line, we'll route via
shielded wire the overflow of IC1 which goes to the input of IC2, and 
we'll add a capacitor on that line too.

Operation: 
----------

-disconnect the power cord, unscrew the and remove the covers, and go to
 the front end, which is the upper board. You may want to unscrew it in
 order to solder the capacitors on the lower side, though you can do all
 the job on the upper side, without unscrewing. If you unscrew it, you'll
 have to plug off J1 (the antenna connector), J1 on the lowermost board
 (the front end output, which is soldered on the front end and called
 there J4).  Labels are printed on the board. You'll also have to be
 gentle when turning upside down the board, and to be careful not to
 deform the coils.

-the track to replace with a shielded wire is the one which connects pin
 9 of IC1 with pin 2 of IC2. This track runs mostly on the upper side of
 the board. The upper part starts on the upper board just beside pin 4 of
 connector J5, and runs mostly parallel to the backpanel until past the
 central screw, close to the relais, where it sinks back to the lowest
 side (no drawings, please). This has to be cut in two places, as close
 as possible to the board crossing points. Scrap them with a sharp tip.
 Solder a shielded wire between the two points. The shield can be
 grounded practically anywhere, as the ground track covers most of the
 side. The wire can be fastened to the board with a drop of hot melt.

-Solder a 6.8nF capacitor between this track and ground. I found it
 easiest at the relais end. (it could be up to 33nF, more than that
 would prevent normal functioning).

-solder a 4.7nF capacitor (up to some 15nF) on the clock line, i.e.
 between pin 1 of J8 and ground. I found it more confortable on the SMD
 side, just at the pin of the connector, but the clock track is also
 available on the upper side. It sinks to the lower side just beside pin
 9 of J8, between J8 and the electrolitic capacitor C80, and is
 surrounded by a convenient ground track.

-rescrew, reconnect connectors if you disconnected, close the unit and
 enjoy.

A final note: I'm indebted to  Henry Laviers <hl1@acpub.duke.edu>, who
sent me a copy of the service manual two years ago, and to Marc Gauw
<marcgauw@simplex.nl> who forwarded me the schematics of the 3000A
(which should be available and orderable from aor-uk, just to prevent
requests to us).

I'd like to very much hear comments or to know if anyone has tried other
hacks on the same box. Please email me.

Enrico Segre, segre@polito.it

PS: check out also my previous 3 hacks for improving the audio bass
fidelity, narrowing AM to 6kHz, and removing the cpu muting - they
should be in the same directory of this one.