My philosophy on this stuff is a little different. I used to be trying to prove something, thank God I woke up to that! These days, valve audio for me is a substitute for a full-size steam train set. I LOVE retro-technology. I have always been fascinated by valves and they are something I can realistically play with anytime I want to. And so my designs are what I want to do; they can appear complex to some minds. I do find that what I build does sound good and I am interested in improving the sound, but always in the spirit of having fun with my train-set.

If you are interested in more detail (including schematics) for the various projects, hyperlinks (underlined) are provided further down this page OR go 'up' and you can navigate through SE, PP, Phono, SACD e.t.c.

Top, left to right; Sota Star Sapphire with Benz LP11, MK11 416 phono-stage, Sony SCD 777-ES. Between the speakers, left; hot-rod DV563a on top of Definitive Powerfield 1500*, to the right of the DV563a is the 2A3 SE line-stage (with separate PSU underneath), on the floor to the right of the sub is the 50-300B SE (with 50s installed) and lurking behind the right speaker is the 845 SE.  Just visible to the right of the right speaker is the 6AS7G PP.

*I developed a somewhat novel integration technique for this unit which is described on the 2A3 SE line-stage page.

Cables are a mix of Analysis Plus Oval 8s (speakers), Analysis Plus Solo Crystal Oval interconnect from the line-stage to the amp, Canare Star-Quad balanced from the DV563a and Cardas Crosslink unbalanced from the SCD 77-ES and phono-stage.  I use a home-brew power conditioner that is described at the bottom of this page.  Heavy rugs are hung on the walls to control reflections.  The sub is on spikes and supports a concrete slab on sorbothane pads in turn, supporting a constrained layer plywood platter under the DV563a.  The Sota is on a cabinet that is bolted to the wall, the cabinet has over 100lb of ballast in it;  On top of the cabinet is a maple platter on sorbothane, then the Sota turntable.  The maple platters are butcher's blocks from IKEA.

The 50-300B is a star with the 50s and shows up the massive 845 SE for the white-elephant that it is (it is a good performer but it does not match the 50, at 200lb and dissipating 700W of heat, it is redundant). The Hot-Rod DV563a is also a star; the SCD 777-ES cannot match it, especially in CD mode. I will be following the lead of my friend Kevin Kennedy in modifying it to eliminate a multiplicity of op-amps in the analogue path. It is EXTREMELY well-constructed and I am sure that it will out-perform the DV563a; That it does not out of the box, shows the potential for thoughtfully modified multi-format players as Hi-Fi digital sources.   

The speakers are Gemme Audio Concertis using Fostex FE108 Sigma banana-pulp cone drivers;

I bought these on impulse because I wanted single driver speakers and I am very attracted to the mechanical realisation of the back horn in these speakers. In brief, the profile of the horn is cut into several layers of wood that are then bonded together to form a beautifully solid and smooth horn.  In the middle of the picture, you can see the layers forming the back-horn profile.


These speakers have been and are, disappointing. They have a pronounced mid-bass muddiness. Taking (finally) the advice of my friend John Dahlman (who is a superb bassist and all-around musician), I put some polyfill stuffing behind the drivers and into the throat. That yielded a big improvement. However, I noticed that the span between the front spikes was only 5 9/16", far too narrow and the speakers were not stable. I have made out-riggers and increased the span to 12 1/2" and the sound is much more clear. To be fair, an out-rigger kit was available as an option; I think that the manufacturer made a mistake by not including the out-riggers with the speaker as standard. They are necessary and the speakers do not perform anywhere close to their potential without them. I am persisting with these speakers because I do find the single-driver, no crossover idea interesting, for now at least. I have a pair of Tannoy 3836 drivers awaiting in the sidings and intend to build a pair of replica Churchill enclosures for them.


This is as near as I can get to a chronology of my system:

My return to tubes started with the original incarnation of my KT88 PPP mono-blocks running from the variable output on a Adcom GCD 700 CD player.  The speakers were Energy C8s and the whole thing was a revelation!

Sears_Triode_60.jpg (78533 bytes)                ppptriodefront.JPG (79864 bytes)


KT88newtwelve.JPG (95934 bytes)So, I got the bug.  I redesigned the KT88 PPP amps using a cascaded long-tailed-pair topology with power supply regulators seemingly sprouting from every available space (and not-so available space) on the chassis.  I installed Sowter DAC 8347 transformers into the Adcom using 25ohm i/v resistors which results in around 560mVpk, so I needed some extra gain.  This was provided with a PP 6SN7 line stage, the (PP) line output transformer being  Lundahl 1660s.  By then I had realised the limitation of potentiometers as volume controls and came up with putting series resistors in each phase in the line stage and hooking an Alps pot across them to form a shunt attenuator.  I use the variable shunt attenuator method still.  It is convenient, sounds good (as good as my switched attenuator, I think) and lends itself to remote operation.  It does have the drawback of variable input impedance.  I minimised the consequences of this by ensuring that the sources (CD, phono, tuner) have low output Z (well, less than 2k anyway) so that the variation of load presented to them has a minimal effect.

AR TT.jpg (165858 bytes)At this point, Kevin Kennedy gave me an AR turntable with a Merrill platter.  And so my thoughts turned to vinyl.  I completed the EL34 triode / UL amp with a phono stage.  I was astonished to discover that all I had read about the superiority of vinyl over polycarbonate is true.  This TT came with an Audiocraft unipivot arm supporting a OL-2 (AT?) MM cartridge which Kevin felt was well past its' best.  Well maybe, but as a vinyl neophyte, it sounded good to me.  However, this was wiped out one sad day when I managed to drag a cable over the TT, breaking the cantilever off the cartridge - thus the guards at each end of the arm in the picture.   I replaced it with the controversial Sumiko Blue Point Special*.  All this started a love affair with vinyl expressed in a number of somewhat derivative MM and MC phono stage designs.  The most recent of these being Calrad hybrid cascode MM redesign.  I have found that the high output cartridges perhaps not surprisingly, sound best working into low capacitance.  Simply using low capacitance cables is insufficient, attention to low Miller capacitance is also critical.  *The Sumiko is sometimes criticized for a confused mid-range and sibilance.  I always enjoyed the dynamics and resolution coupled with good tracking qualities.  Coupling to the cascode seems to retain these qualities while greatly diminishing any sense of confusion and sibilance.

845_Rebuild_010.jpg (93767 bytes)I next embarked on a single ended voyage.  (Actually, I do still play with PP designs.)  Always being a one to start at the end, first up was an 845 effort.  I wound the output transformers, it was quite an odyssey!  It is also extremely heavy at more that 200lb.


Next up is a 50-300B SE design:  As with the 845, I wound the output transformers.  This employs a novel and effective regulated current topology that permits safe operation of 50s in Class A2. I think it sounds great!


Sota & 416 MK11.jpg (146378 bytes)My friend Walter Clay introduced me to the WE 416 planar triode with the potential for multiplexing 19,000 signals which surely must find application in modern recording set-ups with ever increasing numbers of channels..  (Err, just in case, I am joking here - It is a telephone repeater tube used where the level of multiplexing can reach 19,000.)  It is a high Gm, low Ra device with consequent high Miller capacitance, thus I saw potential for it in a low output MC phono stage.  The first of these is the "Adcom", so called because I built it in a recycled Adcom power amp box.  This was followed by the MK11 416 planar triode MC a similar design to the MK1 but supported by individual shunt regulation for each stage.  The TT is a Sota Star Sapphire with vacuum hold-down, equipped with a AQ PT6 arm and a Benz Micro LP11 cartridge.  The table it is shown on (prior to my move) is braced and (was) bolted to the wall. The TT is stood on a concrete paving slab encased in a flannel pillow slip, more to protect the table than anything else! In the post-move picture above, it is shown on an IKEA butchers block on a cabinet filled with sand bags and bolted to the wall.

Sub_Integrator.jpg (126447 bytes)Line Stage 002.jpg (24008 bytes)My current system architecture uses a single sub-woofer.  I built a 2A3 line stage that has separate main channels and sub-woofer outputs.  The main channels have 50Hz high pass filters:  The idea is to remove the large swing sub LF signals from the main amps and speakers, thereby simplifying the job of the amplifiers and reducing the speaker cone movements.  This implementation really improved clarity and imaging.  The left and right sub outputs are integrated using a transformer signal summation (transformers shown in picture on right) approach rather than simply using the L & R inputs on the sub plate amp which, presumably, mix the signals in some parallel fashion.  I discovered that this approach causes some form of destructive interference between the channels resulting in a "cracking" sound, audible when the sub was played alone.  I tried mixing the channels using an external parallel method but the problem remained.  The summation method (described in more detail on the 2A3 line stage page) eliminated it.

Pioneer DV-563A multi-format player and  Fisher FM-200B tuner.




Line_Conditioner.jpg (90500 bytes)

One fine day, I shall get around to developing a linear power regenerator.  Until then, I have built this simple line conditioner using a re-cycled UPS (APC Back-UPS 600), based on Mike Vans Evers design from Sound Practices, Vol 1:#4  The design uses the two heavy 8.5V secondaries as a common mode line choke with 3uF/2000V capacitors across the line before and after the choke.  I had been lucky and picked up some low $ hospital grade receptacles to replace the cheap units that came with the UPS.  Does it improve my system?  I think so.