Welcome to Radio Retro

Why Radio Retro?

Rick’s blog about retro techology and their restoration

Many posts record passed projects to rejuvenate or fix faulty electronic items that no longer operate as intended or suffered physical damage needing restoration. Spare parts have been sourced from private collections or accumulated from past involvement in the electronics industry.

Family History

Most of the knowledge needed to fault find to component level was gained from my father whose career began with Muirhead engineering as an instrument maker. He taught me the importance of understanding how things work by knowing how to make rather than replace. He worked in the early sixties for an electronics manufacturer Nagard,  who produced oscilloscopes and pulse generators at their factory in Belmont, Surrey. Below are some pictures of an early production line from the fifties.

The Nagard valve ‘scopes warmed the workshop and helped me discover the mystery of audio and radio frequencies. When Nagard was bought by Advance electronics my father moved to a startup Darang in Hackbridge near to the Mullard factory in Mitcham. Darang’s innovation was a series of logic modules based on discrete semiconductors and predated the TTL integrated circuit which superseded them. They also produced an early digital clock with Nixie indicator tubes that graced our living room in the 60s.

Darang Black Perspex Nixie Clock 1965

Early Influences

Growing up with the space race and Apollo landings the clock’s case in black perspex glowed magically if not always keeping perfect time. An advert appeared in Practical Electronics, October 1965 from which the photo was taken. The Digicron clock below was a later example of a more accurate metal cased version from a sales brochure.

Darang Digicron Nixie Digital Clock – 1967

Analogue Audio Design

tldr;

My interest in electronics as a teenager was to build ever more powerful high quality audio devices. Friends wanted high-powered amplifiers for mobile disco ventures so I designed and built them for fun and funds of a few pounds. At the time a low-cost silicon transistor, the 2N3055  – or dirty thirty –  was commonly used but suffered from low gain at higher audio frequencies and hence produced a characteristic harsh ‘transistor’ sound. I remember hearing valve amplifiers at gigs and fair grounds that sounded much better at high volume and wanting to understand why this was so.

It was possible to create a transistor amplifier with a clean sound at high volume with careful design – it needed a basic understanding of operational amplifiers, negative feedback and gain-bandwidth products. Some mathematics helped too. John Lindsey-Hood was a master in audio design and I remember building a low distortion oscillator from his design published in Wireless World, that made it possible to measure and test my amplifier designs.

One trick was reuse. Colour TVs at the time required fast transistors in their switched mode power supplies to avoid costly transformers. These switched at supersonic frequencies and so were ideal for audio amplification having much higher gain than the dirty thirty.

Careful  selection of passive components was also important in achieving a clean sound. I was lucky to have a store of instrument quality capacitors and metal film resistors to draw on. Here I found the difference between ceramic, polycarbonate and polyester capacitors for high-quality audio. A simple 0.1uF ceramic compared with a Mullard tropical fish C280 series polyester could make a big difference in power amplifiers. One reason why the Mullard range found their way into many classic guitar amplifiers of the period. Marshall for example favoured the mustard series which were more rugged than the fish that could easily split apart.

Noise or rather lack of it was possible by keeping values low and by using metal film resistors. One and two percent tolerance with their brown and red colour codes rings seemed best. Getting to that 95dB or better noise floor was not easy though. I tried many designs with parallel bipolar and FET input devices that should in theory have produced less noise. Differential inputs were necessary for operational amplifiers so take advantage of their simple design characteristics.

I learnt much about linear amplification by building op-amps from discrete components. Later on it was much easier to use IC op-amps that offered similar performance in a much smaller size. But a discrete design allowed higher voltage swings than the +/- 30v the ICs could stand. And if you want high power with max headroom then +/- 30v was far too low.

In the seventies valve or tube amplifiers were being surpassed by bi-polar transistor designs. Dad’s home-brew  EL84 stereo radiogram calved out of solid light oak became a stand for my ever more powerful transistor creations. I built a series of discrete ‘hi-fi’ amplifiers and FM tuners that increased in power from 15w then 25w to over 50 watts (true RMS per channel). I remember connecting the radiogram to my hifi speakers and despite the increased raw power of my transistor designs to my surprise Dad’s valve set up always sounded smoother. Why was this?

Complex active component design could amount to nothing if the connection between output device and speaker or headphone was imperfect. I remember my first attempt at hiding wiring between my hifi loudspeakers and amplifier in our lounge. It involved black bell wire that seemed ideal to tuck under the fitted carpet edge. Eight metre cable runs with gold plated connectors resulted in a sonic wasteland. Woolly bass and fuzzy top end. Why? Gale provided the answer with their multi-strand speaker cable. Having gold plated terminal posts means nothing if the cable resistance is too high. And it was with thin bell wire that may be fine for a few hundred milliamps. Bell wire measured not far off an ohm for the distances I needed and this meant poor bass response as the impedance was too high. Changing cables made a huge difference with a challenge of hiding the thicker cable. Such speaker cable has become the norm but back in the seventies I don’t remember seeing any at Olympia’s Audio Fair.

Comparing valve amplifier with transistor over improved speaker cables did not answer the question why valve sounded smoother. Luckily I has access to oscilloscopes and through analysis of the waveform when valve and transistor amplifiers drove active loads like hifi speakers it was possible to see the culprit – clipping. Valve designs I had access to sounded smoother as that’s exactly what their waveform looked like compare with transistor design when driven at higher volume. So although my transistor amplifier faithfully reproduced the signals at levels below peak power when the volume increased the output signal flat lined or clipped resulting in a harsh sound. The power in these harmonics could be heard and easily identified. Valve designs on the other hand used transformer output stages to match low impedance speakers and clipping was much smoother.

Having gained a degree in electronic engineering, I’ve spent much of my career in software engineering, chasing the elusive circuit diagram developers love to hate, never loosing my fascination with radio.