50 Years: 1948 - 1998
The Leak TL/12 was introduced in December 1948, and therefore celebrates its 50'th anniversary in December 1998.
For those not familiar with the Leak TL/12, it is:
The following text is an extract from an historical note written by H.J leak in late 1951 (the full text is available in the 28-Page "Point One" booklet listed below.)
To set the context of the following extract, it should be realised that Leak had been making a feedback amplifier known as the Type 15 since 1945, and that this amplifier had similar low distortion (0.1%), but was costly to produce, and was a four-stage circuit.
"In 1948 the problem of costly production was surmounted when we abandoned our original 4-stage circuit of 1945 for the simpler, more sensitive and completely stable 3-stage circuit of the TL/12.
The basic TL/12 circuit was evolved in 1947 by H. J. Leak for use with feedback cutter-heads, and the circuit was so stable that feedback magnitudes greater than 40db could be obtained over the loop containing the three stages, the output transformer and the electro-mechanical circuit of the cutter-head. We realised that with this new circuit we had the opportunity of offering a superlative amplifier, but with our high-grade labour it was essential to use the most advanced techniques of manufacture and to ensure quantity production if the price was to be kept low. We knew that the very large orders necessary for quantity production would come only if the amplifier gained world-wide acceptance from three main classes of users:
1. Professional audio engineers representing the communications industries (broadcasting and recording companies, telephone, telegraph and cable companies, Public Address and diffusion operators) and also those engineers in the distributing branches of the radio industry whose main interest is high-fidelity reproduction.
2. The non-technical music-lover or gramophile.
3. The knowledgeable amateur enthusiast (the "audio hobbyist" of the U.S.) who had previously built his own amplifiers for reasons of economy and/or personal satisfaction.
With these requirements before us further development work was carried out on the basic circuit and on the physical details, and the TL/12 in its present form was launched in December 1948, together with the RC/PA pre-amplifier.
A review of advertising literature from the time reveals many desirable features in the TL/12 that are sought after today in high-end audio amplifiers. Below is a listing of the key features of the amplifier:
From today's perspective some of the most desirable features come from the choice of components within the amplifier:
The amplifier uses a pair of KT66 valves in the output stage that are connected in a triode configuration with individual cathode bias. This enables the amplifier to use un-matched valves in the output stage. A neat side-effect is that the R/C cathode network also acts as a feedback compensation network at low frequencies to control the stability of the amplifier.
The first stage is a pentode voltage gain stage, and the second stage a long-tailed pair phase splitter and driver. This configuration was the basis for all Leak valve power amplifier circuits that would follow for the next 20 years. (Several other amplifiers released after the TL/12 have used a circuit very similar to this - one of which is the Mullard 5-20.)
To achieve the distortion figure of 0.1%, the main feedback loop encompasses all three stages and the output transformer. The loop gain is 20 (ie: 26dB) so the distortion present in the amplifier without feedback would be 2%. In addition Leak published stability margins for the TL/12 emphasising their claim that the application of heavy amounts of feedback cannot be done without due consideration to the stability of the circuit.
A circuit diagram is available for the TL/12.
A web page with extensive information on restoring the TL/12 is available at this site. It discusses resistors, capacitors, iron-cored components, connectors, and valves/tubes.
As part of the promotion for this amplifier, Leak produced a 28-page booklet titled "Leak 'Point One' Amplifiers" which fully referenced classic technical publications of the period. This booklet has been used as a reference for much of this Web page.
The booklet is rare, but is now avaialable from
this web-site. The original graphic-design has been reproduced as well as
possible for Web-distribution.
The TL/12 was manufactured from December 1948 to about 1955. During this time several variations were made to the physical aspects of the amplifier, however the circuit was always the same. Some of these variations are:
A TL/12 cost a lot of money.
In January 1949 the advertised price was 25 Pounds and 15 shillings, and the pre-amplifier cost an additional 6 Pounds 15 shillings. To put this into perspective, we need to know how much a typical wage was in 1949 - and that was probably well under ten pounds. So a TL/12 would have cost several week wages.
If the TL/12 was to be produced today, 50 years on, it would probably still cost severeal weeks wages - and then times two for stereo!
One of the most detailed technical reviews of the TL/12 amplifier appeared in "Radiotronics" Vol 20, No. 9 of September 1955, published by "The Wireless Press for Amalgamated Wireless Valve Company Pty Ltd" of Australia. The authors were F. Langford-Smith and A.R. Chesterman (Langford-Smith is well known for the classic reference text "Radiotron Designers Handbook").
In due course it is intended to add the full review to this web-site. The following extract is taken from the opening paragraph of the 9-page report:
"This amplifier has for many years been regarded by the authors as one of the world's best. It has a triode output stage with all the advantges which this confers when driving a loudspeaker load. It has extermely low non-linear distortion - the measured value at 12 Watts output was less than the specified figure of 0.1% THD" - Langford-Smith & Chesterman, 1955.
The review goes on to test the TL/12 under every conceivable condition including: square waves, shunt capacitance to cause oscillation, stability when shock excited at 5kHz, partially capacitive load tests, pure capacitive loads, absolute stability, frequency response, phase shift, transfer characteristic, harmonic distortion versus power, harmonic distortion versus frequency, intermodulation distortion, ratio of IMD to THD, effect of dummy loudspeaker loads, effect of un-matched valves, hum and noise, output resistance and damping factor, and distortion analysis for individual harmonics up to the 13'th using a wave analyser.
Hopefully the full report will appear here soon.