0148 Third Year Engineering
By opting out of the more specialised branches of Physics, I was hoping that my third year would be rather simpler, and more to my liking. I was very conscious that I had only just scraped through my second year in physics. In practice electrical engineering, which I had chosen for my specialist area, was indeed more to my taste, and I put in a lot more effort -- dropping most of my outside activities. Unfortunately, it was still not enough, for I still ended up with third-class honours!
Electrical engineering was, in fact, also pretty much mathematics-based at the college. Thus I found myself studying the theory of control systems largely in terms of equations. Indeed, I can barely remember anything but controls systems, for my life seemed to be dominated by positive and negative feedback.
On the other hand, the experiments were more interesting. Indeed, one of them nearly gave me a Nobel prize. I was working with on analogue computer, essentially a collection of capacitors and resistors which simulated - in electrical form - the physical problem we are investigating. That investigation was about the classical three body problem. When you have three bodies influencing each other, through gravity or even just springs, the interaction between them becomes very complex -- and not easy to a predict. Hence, by setting up the oscillations and varying resistors, to vary the distances between virtual springs and weights of the bodies, the idea was that we could see how various of these oscillations combined together produced one single oscillation. Normally all of these should have been sinusoidal oscillations.
Perhaps it’s my fate to stumble into the unexpected. Thus one of the set-ups I ran resulted in random motion. It was nothing like nicely sinusoidal but the pattern was literally random; switching backwards and forwards in a totally unpredictable way. I notified my supervisor and he came to look. He checked all the various resistors and the other values; only to a find nothing wrong -- though he was still convinced that must be something wrong with equipment. So he went and got more academics to look at this experiment. It was running for several days, as they tried to work out what could be wrong with equipment to produce such an oddball result. Eventually they just gave up and tore down the setup. They never were able to explain what was happening. Some years later I realised the explanation, for someone else had found a similar problem not long after me. But he had persisted, and tried to find out the explanation. The explanation was chaos theory, which earned him a Nobel prize!
For my end of year project, I changed direction once more and went into experimental psychology. My fellow expatriate from the physics department asked the professor what he wanted him to do. The professor simply asked him to make a Galton Board, used to demonstrate statistics, for his lectures. This involved knocking in lots of nails -- albeit very accurately -- into a wooden board; demonstrating the skills he had learned in woodworking during the third year at school! Indicative of how marks were awarded was the fact that, for this, he got a first. My own work, which was more intellectual and much more demanding, got a third.
Indeed, that whole episode was rather fraught. The work I carried out was - in theory - in conjunction with one of the engineering students, but most of it was done by Pat and myself alone. Essentially what we were trying to establish was the impact of delay on communications. Thus, the equipment essentially consisted of two sound-proof booths, linked by equipment which -- using a tape loop - introduced a variable delay on the line. The reason for doing this was that it was at the time when communications satellites were being positioned in geostationary orbits. Nobody then knew what the impact of the delay introduced by the 50,000 mile round trip would be. The new technology would dramatically improve communication capability between countries, but it would also inevitably introduce a delay in communication. As the geostationary orbit was 22,400 miles above the earth the two-way trip, at the speed of light, took around a third of a second. I doesn't sound much but in fact, as we proved, it was about to be very disruptive of communication.
Thus, people speaking on the telephone, or even face-to-face, are used to responses coming back immediately. When there is a third of a second delay their conversation virtually breaks down. It sounds a small difference, but conversation is a very complex process and is very susceptible to the time interval. At its worse, one person starts a conversation and the other person also starts – perhaps a little after. The first having heard the second then stops, at which the second also stops. Then both the first and the second try again; and this disruption of normal conversation carries on for evermore. We did, indeed, find that the delay in travelling to the geostationary satellite meant that this factor which made speech conversation very difficult was introduced – though, fortunately, it had no impact on computer to computer communication. Even so, my finding had immense practical significance, since it resulted in the PTT's having to lay numbers of fibre-optic cables across the ocean, for at least one leg of person to person telephony.
I developed a very nice technique for quantifying the impact of these delays. Thus Pat and I talked for hours on end, between the two booths, with the delay gradually being changed. Our conversations were recorded on a tape recorder -- a very good tape recorder -- and I then analysed them. The most suitable form of analysis was quite simple. The occurrence, over a given period of time, of frequently used words - such as 'the', 'and' or 'I' - was consistent. Indeed in our experiment the occurrence of these -- over a sufficiently long time, so that any random blips were averaged out -- was plotted. These graphs very clearly showed the stress the speaker was experiencing. It was these plots which showed that a third of second produced the highest stress rates, and that was the reason why the fibre optical lines had to be laid, for at least one leg, to halve the overall time delay to less than a sixth of a second.
Maybe if I had left it at this I would have got something better than third class for my mark on this project. But, as usual, I got fascinated by the subject and read up all the literature I could lay my hands on -- at the expense of my revision for the finals exams (another silly mistake!). I found that there were a number of physiological brain functions which also seemed to have an impact at around a third of second or, more often, three cycles per second. Hence, for example, strobe lights cause epileptic fits at three cycles per second.
I worried about these facts, like a dog gnawing at a bone. Eventually, deriving from the delay mechanism we used on our equipment and combined with my knowledge of computers (and electronic delay lines), I came up with a theory that suggested that the brain had functions similar to a delay lines. It compared inputs into the ear with those that had occurred a third of a second before. This was very easy to imagine in terms of inserting delay lines. As such this enabled the brain to identify what was conversation or some other form of regulated noise and what was random noise!
The result was an inch thick project report, much of which was concerned with the backing for the theory. I don't know if the Professor read it at the time, since he gave me the third - as I have already said. What was rather surprising, however, was that he retained the project. The rule at Imperial College was that only the very best of the first-class projects were retained by the College -- all others were returned to students. Even more suspicious, about a decade later the Professor came out with a similar theory himself! I don't think it has ever held me back, and perhaps getting out of physics and electrical engineering was the best thing I ever did, but still rankles that I was only given a third for it! It took me another forty years to get my PhD!
hits
