Inventions
In the mid-seventies the Company I worked
for sent round a circular.
It explained that within the Industry
there was a tacit understanding that one could design things
without too much consideration of whether it embodied someone
else's idea.
In order to get away with this our Company
was supposed to pull its weight and invent things off its own
bat.
Given an equitable balance of patents
between the three largest competing Defence Companies of the
day no-one would bother if they saw one of their ideas getting
used by one of the other two.
The note went on to explain that we
hadn't been bothering to patent our ideas lately and would we
all think of something that we could register.
"A Patents Expert would be visiting
us soon and would talk to those with ideas and explain what was
required".
It so happened that I was leading a
small team of engineers in the design of a data multiplexer for
an Air Traffic Control system and the design used a few novel
techniques.... eminently patentable!
I duly filled in a form.
Not so much for the glory of getting
a patent awarded to me but because, for each patent awarded,
the inventor would get paid the princely sum of £30.
This may not sound much when you consider
the worth of the Zip Fastener and the like but in those days
a good weekly wage was £50.
Because one invented something in order
to do a job, for which one got paid, one couldn't go it alone
and patent anything yourself, so any reward was better than nothing.
I called in the team and asked them
what we should do.
At the outset of the Project I had carried
out the basic design and the team was now developing it into
a workable product.
The first patent was general.
As part of the design activity we had
designed a general purpose microprocessor on a plug-in circuit
board.
It used ordinary TTL integrated circuits,
random access memory chips (RAM) and read only memory (ROM).
There were a total of ten of these used
in a multiplexer in two types.
One type was used to read data input
by Air Traffic Controllers, from things like tracker balls and
sets of keys and switches.
The second type gathered the data together
and sent it to a number of large computers.
The set of tasks for each was contained
in ROM and could be modified during development until it worked
properly.
The first patent was for a "Microprocessor
Controlled Multiplexer".
The second novel idea was associated
with the architecture of the multiplexer.
Communication between microprocessors
was by a tri-state bus. This enabled any devices wanting to engage
in the transfer of data to be easily coupled together. In effect
by just joining their input/output wires with the system bus.
There were 256 such devices and when
things were working all well and good but if one device failed
or acted in a rogue manner it was very difficult to know which
one it was. For example if one bit of the bus was permanently
dragged down to ground potential.. how did one discover which
device was responsible?
The answer was to insert a small resistor
in series with each bus connection.
The resistors for good devices would
have a different potential across them to the resistor of the
rogue device.
By carefully analysing the noise immunity
for the system we came up with a value for a suitable resistor
and ordered these in dual-in-line form consistent with the design
of the devices driving the bus.
The second patent was therefore for
the design of a bus system which could be quickly tested and
put right if it failed.
I note that this has recently been "invented"
again by a large processor manufacturer, 25 years after our patent
was accepted.
The third patent was a method of asynchronous
signalling between remote computers.
This used normal start-stop serial code
with a variable number of stop elements.
Usually a single stop element is used.
Sometimes "one and a half" elements are used and sometimes
two. Our design used both one and two stop elements. The latter
for marking specific bytes in a sequence.
For example if you wished to transmit
24 bit data blocks using a standard design of chip (a UART),
the 24 bit word would be divided into three equal parts (or bytes).
The first two parts were given a single stop bit and the third
two stop bits. When the UART detected two stop bits it signalled
the receiving device which was able to re-constitute the 24 bit
data blocks.
The design was able to detect errors
and worked as reliably as a normal 8 bit data transmission system.
The chief advantage over other potential
techniques was speed.
Our system worked in real time and needed
to provide 100% reliable information to the Air traffic Controllers
without delay.
We also had a fourth patent-able idea.
At that time there was no easy way of
inserting very large chips into sockets, or for that matter,
extracting them later.
We had developed a simple device for
these purposes. It was so useful the patent expert decided not
to disclose it as our competitors would have copied the idea
and gained a worthwhile advantage.
How to proceed?
Well although we got £30 for each
patent we had to spend a not inconsiderable time preparing a
paper describing it, and having done this, liaise with the Patent
Expert until it was finalised.
The answer lay in the wording of the
note that had been circulated.
Each patent could be requested jointly
by two people and both would get a full £30.
So in a stroke team morale was given
a boost.
For each patent I nominated a second
engineer.
For £30 he could write up the
proposal.
So from our four ideas I got £120
and four other engineers got £30 each.
If you know how to interrogate the UK
Patents Database you could look up the three that where awarded
patents.
I'd have imagined that someone has probably
re-invented a "microprocessor controlled multiplexer"
by now; somebody certainly has re-invented the bus fault detection
method and maybe someone has even worked out a new communications
technique using variable numbers of stop bits.
The Company I used to work for no longer
exists and I got my £30s so who cares?