quicker to walk round to the person concerned and give him a tape of the data than to
go to the bother of setting up a CB communications link).
As already proposed, straight transmission of data is vulnerable and likely to
attract the attention of jammers, etc, if it is obvious. So a more sneaky method is
required. One method would be to transmit the data as a sub-audible Pulse Code
Modulation signal underneath a conversation. This would require the construction of
extra encoding and decoding equipment to attach to the computer, and would cause the
generation of extra sideband energy which in turn would require a reduction in
deviation if the modulation index were not to be exceeded. However, it has the
advantage of being undetectable unless someone is looking for it. Because of the very
restricted bandwidth available to the PCM signal the data rate would be necessarily be
slow. That means that who ever is talking on the speech part of the carrier had better
have a lot to say as the data transfer could last from minutes to hours.
Another method would use a duplex mode of transmission where the two computers
would work a kind of handshake. First, it should be noted that most CB rigs use Binary
Coded Decimal channel change logic, so the channel switch could be removed and the
channel change circuitry fed from the computer’s PIO device instead. The transmit –
receive circuitry must also come under computer control from a port. The system must
use a synchronized clock, and when this was discussed some time ago, it was suggested
the the Rugby MSF time transmission be used for the purpose. At that time the only
circuit around (short of designing ones own) was one published in Wireless World about
three years back. This had the disadvantages of being both complex and expensive.
However, Radio and Electronics World, April 82, have published an interesting design
which should suit. (The R & EW article is good reading because the MSF receiver feeds
another board which decodes the signal, and it uses a Z80 to do it. Food for
thought??) Anyway, here’s how the idea works:
The programs in both computers are set to a standard time point for the data transfer
to commence, say on the hour.
At that time both computers look up the same number from identical tables of random
numbers between 1 and 40 and select the channel chosen.
The sending computer switches the rig to transmit and sends a burst of data, staring
with a numbered header, for less than one second.
At the next second time blip from the MSF receiver, the receiving computer switches to
transmit and acknowledges the receipt of the data, correct, otherwise or not at all.
On the next second time blip, the two computers look up the next channel number in the
random table and change channel.
The sending computer now knows one of three conditions:
- Data received correct
- Data received but incorrect or no data received
- No acknowledgement received from the receiving computer
In condition 1, the next block of data is sent and the sequence repeats
In condition 2, the same block of data is sent again and the sequence repeats
In condition 3, as in condition 2
So it can be seen that the data will eventually be sent, albeit slowly, by the
use of all the available channels for a minimum period only. Minimum interference
would be caused to other users, and any attempt at jamming would be futile as the
random channel table would be sufficiently long to make following the signal
impractical. Also, because of the short duration of the data and the constantly
changing frequency, it would be very difficult to pinpoint the actual location of the
tranmitter without using broad band direction finding equipment. This method of
transmission would probably be slower than the sub-audible PCM method, but is simpler
and could be used overnight for instance. Don’t forget, computers are patient beasts
and will keep trying until the job is complete.
So what is my opinion? Well I don’t think it’s worth trying, round here at any
rate. Does anyone else have any sneaky ideas which would pass unnoticed. I suppose the
only answer would be to take the RAE exam and get an amateur ticket. Although that
won’t help much, amateurs are restricted to Baudot code at a maximum of 50 BAUD, and
then plain text only. Mind you, it would be an interesting exercise to think of a way
of cramming eight bit data into the five bit Baudot code, some kind of four level
shift I suppose.