To many valve amplifier enthusiasts the phrase “Radford Amplifier” prompts feelings of admiration, or even reverence. Radford valve amplifiers designed and built around 40 years ago still command high prices – when you can find one! The 1961 ‘Hi Fi Year Book’ listed just three Radford amplifiers, yet the 1963 issue lists a dozen! So what happened to cause this explosion in activity, and a lasting reputation for the man and his products?...
Arthur Radford was born in 1914. He studied at the Merchant Venturers Technical College and Bristol University. He developed an interest in electronics at an early age, getting his radio amateur license (G6YA) in 1930. He sometimes said that the “YA” in this stood for “Young Arthur”! He also became involved in the manufacture of audio amplifiers, microphones, loudspeakers, and public address systems.
During the Second World War he spent time working on communications and military projects. When the war ended there was a flood of ‘surplus’ equipment that came onto the market, and many of the people who left military service did so with a knowlege of electronics combined with an interest in building their own radio and related equipment. Realising this, Radford set up a shop called Cabot Radio in his native Bristol.
Arthur Radford had produced and sold some valve amplifiers during the 1950’s. These were based on the well-known designs by Williamson, and Mullard. However his main commercial success in that period was in designing and selling ‘Lab Pack’ bench power supplies to colleges and schools. Thousands of these were sold as the UK changed its education system to promote science and technology. Older readers may recall the ‘Nuffield’ scheme being introduced as the new way to teach science and engineering.
Radford Labpack
The quality of the power supplies was based on Radford’s understanding of the design and manufacture of transformers. His other success during that period was in producing output transformers for valve amplifiers. In those days many people built their own amplifier, either from a ‘kit’ or using one of the published designs. Radford sold transformers for these. He also sold them to various companies for use in their own commercial amplifiers – again often versions of the same few published designs. However Radford always had a keen interest in how things worked, and in trying to make improvements. He was convinced he could make better amplifiers than the ones other people were building.
Ideally, a power amplifier would have a flat frequency response, minimal phase delay, and low levels of distortion before any global feedback is employed. In such cases, adding some feedback can ‘gild the lilly’ and improve the already good performance. Unfortunately, many of the old designs had poor frequency and phase response across the audio band. That made it more desirable to find a way to improve the performance. But the phase shifts meant that applying global feedback risked the amplifier becoming unstable. The result might be bursts of oscillation that could damage both amplifier and loudspeaker.
Often, the performance limits were imposed by the imperfections of the output transformer. Knowing this, Radford decided to develop transformers that were so good that this would cease to be the problem. At the start of the 1960’s he released the first of his ‘MA’ (Mono Amplifier) and ‘STA’ (STereo Amplifier) designs.
The STA12 was reviewed by George Tillett in the March 1961 issue of Hi Fi News[1]. At that time Radford was planning to produce a range of amplifiers with various power ratings. The designs swiftly gained a good reputation and Tillett’s review ‘highly recommended’ the STA12. As outlined in the review, the original MA/STA amplifiers were well-made and carefully optimised versions of the established Mullard 5/20 design. However, despite the positive reception, various aspects of the initial designs left Radford dissatisfied.
The graphs on the left in Figure 1 show how the distortion level of the STA12 varied with power at three frequencies. These plots are taken from George Tillett’s review in 1961. The specifications for the STA12 were that it should produce distortion below 0·1% at powers up to 12 Watts. The amplifier easily passed this requirement. However we can see that at higher powers the distortion starts to rise. In particular, the distortion level at high and low frequencies rises much more swiftly with increasing power than at 1 kHz. This reveals the effects of problems with the output transformers and the inherent gain and phase imperfections of the valve stages. The performance of the original was good, and the amplifier easily surpassed its specifications, but Radford was sure he could do better.
After the review, two significant developments occurred. One was that Hi Fi News invited Radford to produce a constructional kit version for the magazine, and provide a series of articles about it. The other was that Radford decided that his ability to make better transformers was now outpacing the rest of the electronics! Until then, the practical limitations of most output transformers cast a shadow over the limitations elsewhere in the well-known amplifier designs, giving them somewhere to hide. Radford saw that these other problems would now be exposed, and needed attention. He therefore got in contact with Arthur R. Bailey, who was a Lecturer in Electronics at Bradford Institute of Technology. Together, Radford and Bailey made a detailed and systematic study of existing power amplifier designs.
The key problem they attacked was the phase shifts in the amplifying stages at high frequencies. This cropped up as a result of the combination of the high resistances used in valve power amps and the effective capacitances of the valve amplifying stages. (The term used in engineering for this problem is “Miller Effect” – although don’t blame the editor of HFN for this as it isn’t his fault!) It was a double-headed problem. The gain of the amplifier tended to fall as you increased the frequency, and the signals were also delayed (phase ‘lagged’). This meant that the input and output would get increasingly out of step as the frequency of the signal increased. Once this had happened, if you tried to apply very much global feedback the result was likely to be instability. Hence the common designs tended to suffer from relatively high distortion at the ends of the audio band, high output impedance, and limited frequency response.
After many experimental tests, and a lot of careful analysis, Bailey came up with a neat solution. The previous Radford amplifier followed the Mullard 5/20 and many other common designs by using a double-triode (ECC83) as the ‘phase splitter’. This stage tended to have a gain that started to fall at only about 15kHz, and produced phase lags that rose to being around 40 to 50 degrees at just 20kHz. However by changing to a pentode-triode for the phase splitter, Bailey was able to extend the gain up to well over 100 kHz, and keep the phase lag produced to no more than a few degrees at 20 kHz. Figure 2 compares the performance of a typical double-triode with a pentode-triode, based on Bailey’s results. As well as providing a more uniform gain across the audio spectrum, the extended phase response made it far easier to apply some global feedback to improve the performance without risking instability or other problems[2].
Although the ‘Mark 2’ versions of the MA/STA amplifier had various other ‘tweaks’, it was this key combination of a change from triode-triode to pentode-triode and the improved output transformers that made the difference. If you look at the right hand graph in Figure 1 you can see the distortion plots of the MA15 Mk2 as measured by George Tillett in another review[3]. Compared with the original STA12 the change is dramatic. Not only is the distortion lower, but the results are now almost uniform over a wide range of frequencies.
The improvement also shows up clearly if we compare the maximum output powers available.
Figure 3 plots the maximum available powers found by George Tillett in his reviews. Note these values are not the frequency response which is much flatter! Since one amplifier was rated at 12 Watts and the other at 15 the plot shows values relative to the specified level in each case. (i.e. 0dB corresponds to 15 Watts for the MA15Mk2, but to only 12 Watts for the STA12.) Looking at Figure 3 you can see that the Mk2 design is significantly better in terms of being able to deliver power at high frequencies. The Mk2 could deliver its rated power right up to 20kHz, but the original could only just manage 10kHz.
The issues of Hi Fi News from June to September 1962 carried a series of articles by Bailey and Radford[4] which presented their new design as a constructional kit for readers to build for themselves. The articles didn’t just give the circuit diagrams and explain how to build the amplifier. They also explained in detail how the design worked, its advantages over other designs, and included extensive test results to demonstrate how well it performed. Bailey also published a technical article[5] in ‘Wireless World’ magazine.
The HFN articles introduced the Mk2 versions of both the MA and STA since the two amplifiers were essentially the same except for their power supply requirements. In effect, an STA was a pair of MAs on the same chassis. They also introduced both the ‘15’ and ‘25’ versions, nominally rated at 15 and 25 Watts, respectively.
The articles on the power amplifier were immediately followed by another four, this time giving the constructional details, etc, of a matching pre-amplifier – the commercial version of which was the Radford SC2. However it was the four articles by Bailey & Radford on the new power amplifier, plus the review of the commercial version of the MA15 Mk2 which was also in the September 1962 issue of HFN that had a major impact. From then onwards, ‘Radford’ became a name linked to to ‘excellent valve power amplifiers’, and with good reason. The amplifiers came into existence as a result of his drive to make ever better designs, and was founded on his skill in manufacturing transformers, etc.
Curiously, although Radford is well remembered, Bailey seems to have been overshadowed and largely forgotten. Yet the four HFN articles were credited to Bailey and Radford, with Bailey’s name given priority. If you read the articles the first few paragraphs make plain that Bailey’s contribution was vital in identifying the problems, and in devising the solution in the form of the pentode-triode phase splitter stage. This becomes even clearer when you read the WW article by Bailey.
A similar fate seems to have happened to another design for which Radford has remained well known. Many people recall that Radford patented a design for the ‘transmission line’ form of loudspeaker. Yet if you look at Wireless World, then again you find an article by Bailey[6] describing the loudspeaker, mentioning it is patented, etc. Bailey also developed the early Radford solid-state amplifier designs – with details appearing in Wireless World – but these failed to gain the reputations of the earlier valve designs.
In fact, the Wireless World article by Bailey on the phase splitter prompted a series of letters – some of them from well known designers like Williamson – which debated the merits of the new phase splitter. For an electronic engineer these letters make fascinating reading. They show that good design is often a matter of personal judgement and skill, based on deciding which facets of performance matter most. (The loudspeaker article also prompted a series of letters.)
Following the HFN articles Radford remained determined to make further improvements. Indeed, even whilst the articles were being published he and Bailey made a change to the design, which was reported in the final article! Once the amplifiers were on sale he continued to work on them. This spawned a variety of models with slightly different details, and with increasing output power capability, etc. Thus 30, 60, and even 100 Watt versions appeared! And in 1966 the ‘Series 3’ versions appeared with performance improved still further in various ways.
Even after Radford retired and the original amplifiers ceased being manufactured their reputation would not fade. This has led to various ‘revival’ versions such as the ‘STA25 Mk4’ of 1984, and the ‘STA25 Renaissance’ which appeared in 1987. These differed from the originals in various ways but were aimed at giving the same high quality of performance to justify the name. This process seems set to continue as another maker has been promising an ‘STA26’ using the Radford brand name for release this year.
Arthur Radford made an outstanding contribution to audio as a result of his determination to make excellent products, combined with his skill in being able to manufacture to a level of quality many other makers would not have even attempted to reach. And as with the Sugden story examined a few months ago, Hi Fi News played a significant part in supporting the advances he made. Although in some ways a ‘swan song’ for valve amplifiers at a time when they were being displaced by solid-state designs, the Radford designs established a benchmark, showing that a valve amplifier could deliver audiophile performance worthy of respect – even into the 21st century!
2100 words
Jim Lesurf
13th Apr 2007
[1] Review of STA12. Hi Fi News March 1961 pp 721
[2] For details of the circuit diagrams, see http://www.audiomisc.co.uk/HFN/Radford/circuits/circuits.html
[3] Review of MA15 Mk2 Hi Fi News September 1962 pp 239-241
[4] A. R. Bailey & A. H. Radford A New High Fidelity Amplifier. Hi Fi News June 1962 pp 29-32, July 1962 pp 90-5, August 1962 pp 150-3, September 1962 pp 204-7
[5] A. R. Bailey A New Phase Splitter Wireless World September 1962 pp 411-3
[6] A. R. Bailey, A Non-Resonant Loudspeaker Enclosure Design Wireless World October 1965 pp 483-6
