on the Leak Troughline tuner. The long-awaited stereo decoder from valve genius Tim de Paravicini has arrived. Noel Keywood tests it.

Interest in Leak's Troughline valve tuner gathers apace. Along with the interest comes a revival of knowledge, which importantly includes the lost skills of repair, alignment and, importantly, improvement. The improvement featured here is a new extemal stereo decoder, using valves and transistors (silicon chips in fact), and available as a kit or made-up from Tim de Paravicini of Esoteric Audio Research. 

It brings superb stereo to the Leak an area where it is deficient. In doing so it takes one ofthe best sounding tuners not-available (you can pickthem up second hand only for around £50!) even further up the sound quality scale, right out ofthe reach of modem solid state tuners.

When the Troughline was first launched, back in the late fifties, it was a mono-only tuner. The Troughline 11 was able to feed an extemal stereo decoder with a stereo encoded (multiplexed) signal, but it had no decoder of its own. This was then replaced by the Troughline 3, still a mono tuner with a multiplexed output, but with revised styling. 

The Troughline Stereo was identical in appearance to the Troughline 3, but had a solid-state decoder shoehorned into its case. It also had an ECC88 front-end valve instead of an ECC84, which improved its sensitivity a fraction. Being a simple design using early transistors, the decoder is acknowledged to be pretty deficient by today's standards. Ours gives vague, almost phasey stereo.

The solution is to use an extemal stereo decoder. Revival of interest in the Troughline and some enthusiastic prompting from us has persuaded Tim de Paravicini to design a novel valve/transistor hybrid unit. 

The idea is a clever one. Valves are used in the mono channel to retain the Troughline's unique clanty and sweetness, since most infommation is camed by this channel. The stereo channel is more difficult to process, so here Tim has used a modem silicon chip, the popular 1310 designed by Motorola. It keeps the decoder simple and lessens the amount of specialist adjustment needed.

Tim has, just for the sake of it, designed an all-valve decoder as well, but it needs six valves instead ofthree and it is uneconomic in his view. The three-valve hybrid reviewed here will be available as a full kit for £90. This includes all metalwork and parts, including valves.

Made up, the cost will be £150. Well finished metalwork is expensive in small quantities; these are very good prices for a highly specialised, low volume kit. It is only because Tim is so fluent with valves that he can design an onginal unit like this for the sort of price a collection of chips demands these days.

The steel case of the decoder houses a circuit board which supports two ECC83 double-triode valves, the 1310 silicon chip, the power supply and all other parts. Final versions will have proper screen-pnnted legends on the front panel; ours was a bare faced prototype. There is a stereo/mono switch and a red LED that acts as a stereo beacon. 

A single phono socket on the rear panel accepts the multiplex signal from the tuner; twin sockets provide a stereo output that goes straight to the tuner input of an amplifier. As the Troughline settles into its on-tune position, the red LED ofthe decoder lights and stereo becomes available from these rear sockets. There is no power switch.

TROUGHLINE COMPATIBILITY

Because extemal stereo decoders have to be fed a 'multiplex signal' (MPX), tuners without such an output must be modified to provide one. Only the Troughline II and early (mono) Troughline 3 have such an output. The original Troughline and the later Troughline Stereo do not have MPX outputs, so they must be modified. It's a simple task for a radio engineer; he just has to find the output of the detector and solder in a lead.

Now that tuners are universally fitted with their own internal stereo decoders, extemal units have all but disappeared. As a result mono Leak Troughlines are less sought after and therefore less expensive than the later stereo version. They are easier to find and even, to some eyes,
visually more attractive than the later stereo unit; our pictures show the late fifties styling
of the Troughline II. Car boot sales, Bntish Vintage Wireless Society swap-meets, vintage equipment dealers and second-hand shops not hoovered of their best by Richard Kelly are all fertile territory for Troughline lls. 

SOUND QUALITY

Being a long term and enthusiastic Troughline Stereo user, I was hoping to be impressed by the combination we received from Tim: a Troughline II mono and the decoder. Initially, I suspected a bit of tweaking was necessary; I heard some obvious distortion on musical peaks. It was. I had to maximise separation to 40dB when working with the Troughline II (rather than from our stereo generator), and to minimise distortion. 

The input trimmer capacitor that adjusts pilot tone phase plays a crucial role here. We have the stereo generators and spectrum analyser to make this process speedy and accurate, which does help. Tim says that owners can, wth care and time, adjust accurately by ear.

After adjustment the combination was every bit as impressive as I had hoped it might be. The stereo staging was less emphatic in image sharpness than some solid state designs, but this was more a result of the nature of the overall presentation. 

The clean, well defined images of a solid state tuner I would compare to paper cutouts; the edges are sharp and obvious, but that's because the image is largely a two-dimensional one. Imagine adding the depth dimension to performers and a stunning insight or clarity so they can be closely experienced. They become embod-ied and believable, full and rounded. The sharpness of a flat image is exchanged for dimensionality and solidity; images become less superficial and more real.

That there were perfectly established left and right signals, and therefore perfect stereo, was obvious whilst listening both from the aerial, from test signals and music transmitted from our stereo generator. In fact, I constantly had to check the Troughline from the generator to confirm
my suspicions on what we were hearng off- air. 

The difficulty with assessment from broadcasts is that a good tuner reveals a whole host of problems, especially hiss, source quality and - of course - the increas-ingly strong amounts of programme compression used. So if the tuner sounds bad - it may be because it is good!

John Vizor of Mission was passing through so we asked him to listen and comment. He noted the Troughline was clearly, if disconcertingly, revealing the character and hiss level changes ofthe
varous sources belng used by the radio station (Radios 2 and 3). The quality of the microphones being used was also audible as they changed. All this was the sign of a great tuner; he said he hadn't heard anything so revealing for years. 

When we started this short listening session with our own transmissions from CD John immediately wrote off the sound as "too good"; "you can't get anything like this in practice" was the comment. "It sounds like a high quality source such as CD. It doesn't sound like radio." But even listening off-air he felt the Troughline was way ahead of what is currently the norm.

In use this combination had such a fulsome, atmospheric, yet clear sound, it hardly seemed possible. The radically different nature of the electronics and, of course, the relative simplicity of the circuits in contrast to modem tuners was and is, with the Troughline, highly apparent.

There are a few practical qualifications I need to make for the sake of balance. On peaks, I felt I could hear some coarseness creeping in and this was almost certainly due to distortion. The
Troughline lI's detector stage produces around 1% (second and third harmonics) at 50% modulation and around 3% or so at full level ( 100% mod.). 

This can be heard, but it doesn't really seem very significant in light of all the other attrbutes. Audibility depends upon modulation level too; the compressed Rock stations were potentially the ones to suffer, but their sound was so poor in any case that it was hard to tell what was going on. Radio 2 and 3 sounded beautifully clean and clear, because they don't hit full level and, therefore, don't suffer this distortion. In effect the slight distortion of the Troughline was disguised by the nature of radio station output.

Radio 3 was silent (ignoring transmitted -hiss from tapes, etc) dunng the day but became noisier at night due I suspect to interference breaking through. This is a result of the mediocre selectivty of the Troughline; it has a very short, simple IF strip. Sound quality on Radio 3 was beautifully fulsome yet clear all the same. Breakthrough depends upon location - in ones quieter than ours (Central London) this will not be a problem.

With Rock music the tuner was very obviously lucid at spectrum extremes. It offers clear, solid bass lines, rather than the somewhat anaemic suggestion of a bass line delivered by modem solid state designs. Similarly, I noticed sweet and finely differentiated treble of a sort not available from modern tuners. But this is an obvious and acknowledged weakness of modern tuners.

For me, this Troughline set-up offered the best sound I have ever heard from radio. It did require quite a lot of specialist adjustment and knowledge to get it all optimised, but Tim can tackle this for customers. All the same, I wouldn't compare this tuner with any other, it's vastly superior.
Luckily for present day manufacturers, the practicalities of owning and using a Troughline, with ts creaky analogue tuning, will dissuade most from binning their tranny. It's just a pity that they'll never hear what they are missing from radio.

I've been a long-standing devotee of radio, as readers will know by now from my comments in tuner reviews. I've not a lot of patience for people who dismiss the airwaves - and all that wealth of music - as a third-rate source. I was about to listen to another tuner when I was hijacked to hear Tim's Troughline.

I caught - and uttery unmistakeably so - the end of Joshua Bell playing Sibelius' Violin concerto. His tone simply couldn't be mistaken: the Troughline reproduced it so remarkably truthfully I recognised it within seconds, even though it's nearly a year since I last heard him live. 

The listener's seat was simply the best in the hall, as it should be. Quite stunning was a few minutes of Hamish Milne's Annees de Pelerinage, with great weight and solidity to the piano and the full scale and ambience of Birmingham's Pebble Mill. It was like being there - and I have been.

Equally effective was playing CDs through the Troughline, pretending to be a radio station. It was diffcult to believe that there was a tuner in the middle at all. Simon Rattle's Mahler ~ was expansive and dynamic, a strong, heady balanced mixture with a broad sweep of orchestral colour, wide and deep, with superbly defined orchestral detail. 

Tuming to Norrington's authentic-instrument Beethoven Sixth, it had realistically dry cellos and basses, recognisable gut strings and a good round, full sound. String tone generally was extremely sweet.

If I would query anything, it would be a trace of discernible distortion in orchestral tuttis, where the brass took on some harshness and orchestral sections became less clear-cut and somewhat muddled in texture. All the same, thinking of instrumental textures, Melvyn Tan's Fortepiano was so gloriously real and dynamically well-defined it defied belief.

There's no doubt at all that the Paravicini'd Troughline stands very high in the ranks of tunerdom. It demonstrates, with frightening effectiveness, how - franky - lax tuner design generally has become over the last twenty years. The Marantz 10B beats it - but then that is such an over-the- top design I'd be stunned if it didn't. No wonder people have gone off tuners since. Pile the pressure on, demand a good radio listening seat. Take on a Troughline.

At its introduction the VHF/FM stereo system we use in Britain had to be mono compatible, since all tuners were mono-only. This meant that as a prerequisite, transmitters had to continue to transmit a normal mono signal (L+R). The U.S. Zenith/GE system of stereo transmission was eventually chosen for the UK. This transmits the neccessary mono signal and puts stereo information onto a separate difference signal.
 

Figure 1 At the radio station, left and right signals from the stereo microphones are combined in the encoder to form the mono channel and the difference channel (L-R). The latter is put onto a carrier at 38kHz to separate it from the mono. The modulation is double sideband (DSB) with a suppressed 38kHz carrier. The resultant composite signal, which contains a 19kHz synchronising pilot tone is then put onto the VHF radio 'carrier' signal and transmitted.

Figure 2 Tim de Paravicini's stereo decoder for the Leak Troughline channels the important but easy-to process mono (R+L) signal through valve circuits. The stereo 'difference' signal is more difficult to handle but contributes less to the audio signal, so it gets demodulated by a '1310' silcon chip decoder. The resultant sum and difference signals are then combined in the valve output stage to get normal Left and Right channel stereop outputs. This hybrid approach retains the valve sound, but makes for a simple, cost effective decoder having an excellent measured performance

The composite output from the tuner is amplfied and buffered by the first valve (VIA). A trimmer capacitor phase-equalises the multiplexed signal.

The output from this valve (about 1V at 50% mod.) goes to two places: the mono is filtered and sent to both L and R output stages. The 38kHz double- sideband component that contains the
stereo 'difference' information plus 19kHz pilot tone goes to an MC1310 integrated- circuit stereo decoder. This has no feedback in the audio part but has an accurate four-quadrant multiplier to act as a linear product detector to extract the difference audio from the 38kHz double- sideband signal.

Its outputs are buffered and summed by a dual op amp to provide a pure difference signal. For phase matching this is filtered by an identical network to that in the mono channel. It is then subtracted by V2B from the mono to provide R channel output. The signal is also phase inverted
by V1B (unity gain) to be subtracted fnom the mono by V2A to provide L channel output (-ve plus -ve = +ve).

This method was chosen to provide tube purity for the majority of the audio (i.e. that in the mono path) and the minimum adjustment for alignment. The input trim cap and the 1M + 1M resistor allow tailoring of the unit to just about all valve tuners.

A well adjusted tuner and decoder should provide in excess of 40dB separation from 30Hz-3kHz and 30dB at 10kHz. Distortion of either mono (L+R) or difference (L-R) channel is about 0.2%.
Frequency response should be sensibly flat to 16kHz. Output is about 1V for 50% mod. Pilot tone at 19kHz should be about -50dB and output impedance around 2kohms, which is adequate for most cables.

An all-valve circuit would have needed another six tubes to have reached the same standard of performance. I have developed it in parallel, but it is not an economic proposition. There are no
diodes in the switching path for purity. Tim de Paravicini

Whatever the nature and context of Tim's designs - and some of them, like this one, are pretty
unusual - measured performance is not sacrificed. Even though this decoder is a hybrid using valves and solid sute devices it possesses a fine per- formance.

Equalisation (de-emphasis) and pilot tone filtering have been handled fluently. Flat frequency response up to 16kHz, shown in the analysis, and deep pilot tone (19kHz) suppression of -72dB are the result. These characteristics become those of the Troughline in practice, since an external decoder bypasses the tuner's own equalisation and filtering circuits.

Channel separation was wide right across the audio band, measuring 40dB at 1kHz and I10kHz - an excellent result. Pilot tone phase has to be adjusted, using the input trimmer, to get this separation optimised. I found the trimmer had a coarse action; it has to be twiddled with care for best stereo separation.

Distortion was second harmonic only in nature and was almost certainly generated by the valves (I tested the decoder in isolation from the tuner). This harmonic is subjectively undetectable as 'distortion'. It lightens timbre only, but only in quantities greater than a few percent. Since the
decoder produced around 0.2% distortion in itself, it will not tinge or add obvious distortion in
practice.

Finally, there's always the potential annoyance of hiss, something that concerns fans of Radio 3 in particular. It was extremely low at -81dB. Noise from the tuner, not this decoder, will predominate in use.

Much as I expected, in spite of its unusual nature, Tim's decoder has a superb measured performance. It equals that of good modern tuners in nearly all areas. There's just a bit more
second harmonic distortion from the valves, but subjectively this is of no consequence. It is a
worthy addition to the Troughline, bringing modern sundards of stereo performance, whilst
retaining the essential character of its valve circuits.