Tuesday, 30 December 2014

The Legacy of Tube Preamplifier Designs – The Driving Output Stage I

Hello to everybody,

a last posting for this year is made with the intention to bring credit to the leavings of important figures which did die within the year, in the early beginning Ken Shindo and in autum Jean Constant Verdier. Ken Shindo is one of the few in the western hemisphere known japanese designers who did design a wide range of exceptional balanced and elaborated sounding tube amplifiers. As the most people might know, in Japan the evaluation of historic western tube equipment was paragon. They started to inspect, research. refine and outperform the classic examples made for the early professional broadcast or cinematographic use like WE. RCA, Altec, Klangfilm and others, while the western world got flooded with their cheap commercial electronic mass products from Pioneer, Sansui and others in the 1970ties years for home entertainment. Their particular asian appropriation of mastership thinking, their diligence and finally their conscientiousness made at the other end unique products possible, which collect more experiences in that field of expertise and are far ahead of other made product in the world of tube audio.

This article will be some sort of summery to my evaluations which I made within the last weeks in order find out about some basic influences of particular design topologies important with tube amplified output stages. It came into close inspection by a typical side accident which happened in November. I had a chance to listen to a original Kaneda preamplifier in a very elaborated audio chain, where a Line Magnetic LM22A/555W/597 horn system was driven by a Shindo power amplifier and a Monbrison preamplifier. It happened that we were talking about Kaneda preamps and it came to the point that I instantly got offered the chance to listen to a sample. Me was told that the (japanese) transistors inside normally love a long warmup time, – in every day use they never will be switched off… The Kaneda preamplifiers got popular in Europe through Jean Hiraga and the parisian Maison de L'Audiophile in the late 1970ties and early 1980ties, who brought this design topology from Japan. In these days the preamps got mostly sold as kits and have been highly praised as perfect alternatives to high priced tube amps mainly coming from the US in these days.

In order to get the Kaneda warmed up to best performance, we did wait for some hour before we switched over from the Monbrison. As we did before we did listen to digital streamed music material. This means we did listen to a fraction of both amplifiers, so only the output stage (line) and the power supply from each unit have been involved. During the warming up we did listen to the Shindo with a variety of program material, before we switched to the Kaneda unit.
To be honest, I would not have believed before what happened. The difference between both units was tremendous wide, more than that. Coming from the Shindo, the music was natural in a flow with all dynamic attributes right in place and loads of micro dynamic information (commonly named as air). With the Kaneda the music flow was stiff, almost grey and grainy, without any particular micro dynamic detail. It might be have been the temperature of the Kaneda, but even with a imaginable top condition it would not be assumable to difference out to a natural completeness shown by Shindos Monbrison. I was staggering and noticed that only one tube with its PSU was able to outperform and boost the Class A battery supplied Kaneda into a different world. Again a perfect example that Shindo did better than perfect know what he realizes. And in difference to the most other tube amplifier designers, he knows exactly how it has sound finally. His superiority was not alone the knowledge and experience with parts, components and topologies, it was mainly his aim to get the right decisions into order for a precisely aimed audible result, always having a perfect dynamic natural performance in mind.

After that unforeseen demonstration of superiority I went home deeply impressed how well Shindo implemented his output stage in his design. The most here reading, do know that I am into the merits of directly heated triodes in small signal stages since a decade. My long term approved preamplifier uses the Aa Weitverkehrsröhren, a design with some little restrictions (high output impedance) but so far best audible general performance I have used. I have been at a point were I did not want to question this design again. But this experience did do. It pointed out to the most questionable part, the output impedance of the driving line amp is tremendous important part of the complete sonic result. OK, thank you Ken Shindo for your advice, I will take another lecture as homework about this topic.

A typical professional US broadcast preamplifier as made by several companies like IPC, RCA, Pultec or similar with high mu input  stage 12AX7/ECC83 and paralleled 12AU7/ECC82 transformer coupled output stage in order to archive low impedance output (600 ohm or lower).

First of all I did work out theoretically all possible tubes which might be superior within such a design (output transformer with 600 Z secondary). For the beginning I did look up what will be available from my shelves, – what will be necessary to get. I had several output transformers for that reason lying around (different types of Lundahl LL1660, LL1680, its original UTC LS27 and a pair of Tango permalloy core 20K:10K/600) and for practical reasons I started with tubes I already own. The most output transformers are limited at 10K to 20K primary impedance when set to a realistic 600 Z output with reasonable DC in primary (5-20mA). A rule of thumb says that you need to calculate three to five times the internal resistance of the tube to find the best working value of the output transformer (example a 3K Ri of a given tube will match a 12K transformer primary side). A variant of design is to incorporate a additional anode choke (getting the OPT transformer off dc in its primary, but reintroduces a coupling capacitor). For that variants I will have a selection of Lundahl LL1667 air gapped chokes for 5 to 25 mA DC by hand.

Other than a producer of commercial amounts of units I can set my interest to the best components available, I don't have to take questions of availability or single prices into account when I decide. All my tube amplified life I was a great fanatic of the RCA red base series of tubes (5691/6SL7GT, 5692/6SN/GT, 5693/6SJ7GT) which were originally made for the cruise missile control units and are extremely mechanically rugged versions with double heater consumption of the standard octal versions (600mA does not make them a universal replacement). Even their prices have always been literally skyrocketed and so you rarely will see any commercial made audio amp designed with them (Bruce Moores Luminiscence), but I believe they are some of the very best made tubes ever. In all these years these tubes have outperformed everything similar in all my preamplifiers, 12AY7, ECC82 or 802s, Telefunken, Mullard, – everything!
Here now the 5692 with internal impedance of 8K is a candidate for paralleling their two units to half their internal impedance. As well a true legend from the German Post, the C3g with 3K ripple in triode mode is a ideal contender for such implementation. Western Electrics 417A with 1.6K might be a true choice as some other space age tubes like 5687, 6463, 6386 have to be taken into account. But as well real power output tubes like 6F6G, EL84 and similar types will do the job. All named tubes are indirectly heated types, but there are a few DHTs as well which might outperform the late high-tech stuff? The Ce/Cas Weitverkehrsröhren with 4K ripple, the british pedants 3A/110-B and 3A/101-B and of course the Western Electric tennis ball battery tubes are matching perfectly, but you need to find them since they are exceptional rare. As there are little better known small power triodes like 801A, 2A3, etc., but they make output transformers necessary with higher DC capability in the primary winding or the use of the parafeed design together with a choke load.

To make such a comparison meaningful I know that the power supply might take an important part at the final result. So I arranged three different interchangeable power supplies within each setup.

First my standard PSU with AZ12 (mesh plates), all oil paper capacitors and several chokes in the HV rail and a selenium rectified LV supply with chokes is a complex and huge unit. All Transformers are hermetically sealed oil filled types. The anode supply is regulated and stabilized with two tubes as three active units.
Second a industrial type PSU with Selenium HV rectifier and highly regulated and stabilized with four tubes and AC low voltage supply is a alternative solution.
Third a standard supply without regulation and stabilization, just choke and capacitor filtered. LV selenium rectified and RC-filter rail is another option.

With all supplies I have the chance to adjust the anode supply widely, so that any tube from my range can be optimized with anode voltages between 150 to 230 volts. So I started to install identical connection plugs to each unit, which gives a flying change during evaluation a chance.

The results of my comparisons are nothing less than remarkable and expressive, they gave me a deep insight into the effects, requirements and influences of each topology. The optimal working point of the tube itself, the adaption and matching of the output transformer, the influence of different generated anode voltages and as well different filament supplies. And finally the tube itself and its character in the final soundstage. As little shortcut to the end it can be said, the tube itself has a highly overrated position to the final result.

Read on soon, Volker