Rediffusion Central Services - Jake Loddington

Photography
At first, the main application of home-built electronics to my photography was for making timers for use when printing. There were (and probably still are) two schools of thought about this: digital or analogue. My constructions involved both.
The first was made when I was still in Nottingham, and had just been introduced to relay and uniselector design. It used as a clock signal the half-second pulses from an ex-Government clockwork timing mechanism, which I think was originally intended for delayed-action fuses. This fed a divide-by-two circuit consisting of two 3000-type relays, and thence to a uniselector which stepped round, one second at a step. A group of switches selected the desired time, at which point the uniselector switched off the lamp and homed ready for the next exposure. It worked as planned, but it was far too big, it was much too noisy, and it radiated r.f.i. for miles around. So it had to be consigned to the bin.
The next was one of the best; it had been designed for 'Wireless World' by a photographer. It was a very stable and accurate analogue design (using valves), with a Miller run-down circuit as the main timing element. Two big virtues were that it had fourth-power compensation for mains voltage changes, so that a reduction in mains voltage (not uncommon in those days) gave an increase in exposure time. The other very sensible idea was that instead of being calibrated in seconds, it was switched in root-two steps, with a sixth-root-two vernier switch for fine adjustments. I adopted this timing scheme in my own later and definitive design.
For no good reason, I was still seduced by the absolute accuracy obtainable with digital circuits, and I built two decade timers: the first used Ericsson dekatrons and cold-cathode relays, and the second used thyristors in decade ring counters, with surplus C106-type thyristors.
The final design, which was in use until my interest in printing evaporated, was a direct consequence of getting to grips with TTL logic circuits. It used the 50 Hz mains as a clock, and a sequence of 7493 decade dividers. But although it counted in decades, the selector switch ranges were based on a 1, 1.4, 2, 2.8, 4, 5.6 sequence. (Again with a sixth-root-two fine adjustment.) There was no mains-voltage compensation, for the following reason.
Another worry at the time was the variation in mains supply voltage, which caused variations not only in the brightness of the enlarger lamp but also in its colour temperature. (There were quite enough variables when colour printing, without this extra one.) The best solution was to use a saturated-core stabilising transformer, such as those made by Advance, but these were very expensive. I therefore produced a small thyristor unit, in which one half-cycle was phase-controlled by a cosine-modified ramp and pedestal to give open-loop control of the firing point. This gave adequate correction over a range of mains voltage from 200 V to 260 V. (The Advance stabiliser, which I eventually managed to find second-hand, was effective from 170 V to 280 V.)
Flash photography was of great interest. I built a range of slave flash synchronisers for the local professional photographer, which ranged from an early battery-operated device using a CdS cell, germanium transistors and a thyristor, to a more modern design using a high-voltage light-sensitive thyristor. I could never believe that CdS would be fast enough for this, but it worked well.
Another device I made for the professional was a flash delay unit. When he used electronic flash with slow shutter speeds, the firing of the flash at the beginning of the exposure often caused the model to move or flinch, giving a blurred image. He needed a delayed firing switch, so that the flash would fire at the end of the long exposure (but not after the shutter had closed!). This unit was built using 555 timers, and could be switched for 900 ms, 450 ms or 220 ms delays.
I became very interested in high-speed flash, and used a Mullard 2 kV flash tube with an assortment of old oil-filled capacitors in a totally unenclosed power pack. This lethal contraption gave good service for the usual range of balloon bursting and soda syphon pictures.
Naturally I built my own studio flash outfit. It began as a series of self-contained flash heads, each containing its own flash capacitors and power unit, but the later version used a large floor-standing power pack with heavy cables to the heads. There were three 100 joule broad heads, but I never managed to build a decent flash spot.

Early Multimedia
Although I was a confirmed print man, I became interested in colour slides, and, inevitably, in slide-sound or audio-visual. The first thing was to buy two identical projectors (Rollei P35s), and to fit them with low-voltage high-current triacs in the 24 V lamp circuits. My first crossfade unit was a simple twin ujt device, manually operated from a common slider control.
I then considered the design of the tone-operated Animatic 'Convar', produced a suitable pcb, and went into production. Many people in clubs all round the North-West made these units, having signed a non-disclosure agreement.
One disadvantage of the tone-operated design was the extreme audibility of the control tone (1 kHz - 4 kHz), and this made cassette tape recorders quite unusable. Good-quality reel-to-reel machines had to be used, and I converted a second-hand 'Ferrograph' to take a third head, to give stereo sound as well as the control track.
My most vivid memory of this device is the succession of occasions when Lesley and I were visiting camera clubs in the North-West, and we had to resort to the back-up machine, because the Ferrograph let us down. I would try it out at home, and it would work perfectly. We would then take it to the next club on our circuit, and when we ran the test sequence, there would be erratic triggering of slide changes. It took me a long time to recognise poor power supply regulation as the problem. At home, playing the tapes at low volume, the supply rail voltage was stable enough, but in a large auditorium, at full volume, the Class B stereo amplifier hammered the supply sufficiently to upset the control circuit.
Another disadvantage of the 'Convar' system was that a single f.m. tone was used to control two projectors, and so only one could be fully on at a time. (Much scorn from the people who used mechanical crossfades.)
I designed and produced a prototype of a machine called IMPACT (Independent Multiple Projectors Automatically Controlled by Tape), which used pulse-width modulation to give fully independent control of four projectors. This worked and was demonstrated, but never really took off.
Decline and Fall
By this time I had realised that the only sensible way to control multiple projectors was via D-A converters from a microcomputer. So I booked myself onto a course at Blackpool Tech. called 'Microprocessors for Engineers', and the rest, as they say, is history.
My first computer was built from a kit by Newbear. It was called a 77-68, because it was built in 1977 (can it really be 33 years ago?), and it used a Motorola 6800 cpu. It had a clock speed of 1 MHz, with 256 bytes (yes: bytes) of RAM, no keyboard, no storage, no monitor, no nothing. It had 8 switches for addressing, 8 switches for data input, and 8 leds for output. It had to be programmed in machine code using binary, because assembly language was considered effeminate. But I did use Assembler later, of course.
Later, when I had added a couple of 4K RAM boards and a 300 baud tape interface, I bought a "tiny" BASIC in the form of a 4K hexadecimal dump, and spent a weekend keying it in. (Actually I spent Saturday morning keying it in, and the rest of the weekend debugging it.) By this time I had connected a home-made serial keyboard and a monochrome monitor.
Producing hard copy was a problem. About the only printer available for amateurs at that time was an Epson MX80 (I think), and this was beyond my means. So I scrounged from a cousin who worked for BT an old Creed teleprinter (was it a Model 5?), which had a footprint about a metre square. It used a 5-bit baudot code, so I had to write my own program, and blow it into a 2708 EPROM, to convert ASCII into baudot. Caps only, of course.

Proper Computers
I resisted the urge to buy a Sinclair ZX80 kit, and I certainly couldn’t afford a Commodore PET, but I did actually order a Sinclair Spectrum. But the delivery times were so delayed that I cancelled the order and bought a Dragon 32 instead (also a 6800 machine). This had its own ROM BASIC, and was a proper machine at last. Its I/O facilities were limited, though, and I wanted to make it do things, so I bought a BBC Model B (6502; 32 K) and used this for several years. A wonderful machine (still going strong in many schools). BBC ROM BASIC was semi-structured (although it still had GOTO), and it became possible to write respectable programs.
At that time I was teaching computers at Queen Elizabeth’s Grammar School at Blackburn, and the decision was taken to change from BBCs to Nimbus computers made by Research Machines. To keep up, I bought my own machine, and had to get to grips with the IBM PC technology on which the Nimbus was based. So I learned DOS, but was still programming in BASIC.
My own Nimbus had two 3.5” floppies but no hard drive, so I bought a naked drive (10 MB; £200) and built the HDD interface board myself.
I felt the need to use a more powerful structured and compiled language, so I taught myself to use Borland’s Turbo Pascal, and many of the control experiments I devised for QEGS were written in Pascal. But I never got as far as C, let alone C++.
When I retired from QEGS I sold them my Nimbus and bought a PC (and a laser printer) from an ex-pupil of mine, who was then running a small computer shop in Blackburn. (The Epson laser cost nearly £2000: my present Samsung was just under £50.)
That was all in 1991. Since then I have tried to keep up with Bill Gates, going from DOS to Windows 3.1, and at last reaching XP. I will give Microsoft another year or so to get the bugs out of Vista, although for my own requirements, there seems to be no need to move from XP.

Jake Loddington