ELLA AMPLIFIER (Old Version)
Circuit : Shunt Regulated Push-Pull (SRPP) configuration using both sections of a 6922 Medium Mu Triode. This is directly coupled to a Medium Mu 6SN7 Twin Triode, with both sections configured as a relative to the 'long tailed pair' phase splitter. The output is fed, via coupling capacitors, to a pair of EL-34 Pentode Power tubes, operating in 'Fixed Bias' mode (Class AB). The OPT is connected in 'Ultralinear' mode, with a Zobel Network connected across the Speaker output.
Negative feedback, from the Speaker terminal to the input tube, is incorporated.
The design is such that EL-34's, KT-88's and 6550's can all be used
I first built the P to P (point to point) stock version back in December 05 (slightly enhanced with Kiwame resistors and bronze obbligato capacitors). As I was new to soldering, the first version was a tad messy. The unit didn't work on switch on, but thanks to numerous communications to and fro with Mr Brian Cherry (diyhifsupply) - voila, ELLA up and running. Slight hum from the speakers but nothing major. Sounded pretty damn good, much better than my original MF Synthesis ss amp. Of course, during the build I became smitten with the bug. You know the one, why does the circuit work as it does, what can I do to improve things, 'TUBE ROLLING', what's all the fuss about re SETs, etc etc. As so often happens when the bug strikes, wallet opened and I purchased an LD+/Mk2 (300B Single Ended Triode monoblocks) and Basie Preamp. Units worked first time , better than the ELLA with such delicacy and accuracy of sound - Tremendous ! BUT, my main speakers were 88 db, just loud enough but no more. Eventually blew the 300B driver valve - oops. I was going to build my own full range speakers incorporating Brian's own driver (Neo-Fone), with higher efficiency, but this would have to wait, as I decided to embark on another leg of the ELLA journey.
By this time, my soldering prowess had improved markedly, and I was somewhat ashamed with the spaghetti under the original ELLA's hood, certainly in comparison with the SET gear. I read the various forum notes and recommendations for possible upgrades, read Morgan Jones 'Valve Amplifiers' cover to cover several times, and decided to upgrade the amp. New holes in the chassis pretty much demanded a full rebuild, and anyway that awful wiring had to go. Take it all apart, start from scratch (but take time to do it right) and incorporate Triode/UL mode switching, NFB On/Off, and perhaps most importantly better components and wiring looms. An expensive decision was also warranted. In all texts that I had read to date, the main problem with valve amplifiers, in almost all their various incarnations, (excepting the quirky OTL (output transformerless) designs), was the Output Transformer. Irrespective of circuit design, without a top quality OPT, all was seemingly lost (or sound quality at the very least greatly compromised). As it happened, Mr Cherry was selling an exotic one, full of Amorphous material. I had to have the best, so my wallet opened yet again !!
Other bits and pieces in the New ELLA:-
3 dp/dt switches, two for Triode/UL switching and the other for the NFB On/Off
High Resolution Bias Pots (much better than the stock ones, could easily set each valve to within 1mV of each other (as measured across the 10ohm Cathode resistors).
Teflon valve bases (tightest possible connection for the valves).
Silver wire, and silver solder for all signal wires, OFC copper for all ground connections and bias power supply.
I kept the Kiwames and Obbligatos as is.
Took three months of steady effort to complete the ELLA MK2. Worked first time.
So to the results....
Kit used to assess the build included Sony sacd/cd player, coaxial digital out to Musical Fidelity DACV3, MF 10V3 tube buffer, both with stand alone power supply, Rogers LS-7 BBC Monitor Loudspeakers, Tannoy ST-50 Supertweeters, REL Quake Sub-base unit. Cable - Atlas Compass and Equator interconnects, Mark Levinson Speaker cable.
Absolute silence from the amp, no hum, nothing, even with ears up to the loudspeakers, full volume.
Using Valve Art matched KT-88's, and after letting the amp play for some 50 hours, power was seemingly less with the Amorphous than with the stock OPT, in UL mode. Switching from UL to Triode mode on the fly reduced gain by only a very small amount, indicating that the UL taps on the Amorphous OPT were set for minimum distortion, rather than maximum power. UL mode was clearer in the highs, and tighter in the lows when compared with Triode mode. This indicated a true high end quality to the Amorphous OPT's.
Massive change in sound switching NFB off. Much more gain, opening up the sound stage somewhat, but at the expense of tighter bass – bass seemed to wallow around (could this be due to a lower damping factor ?). I like tight bass, so back to Negative feedback.
Hey, and the ELLA was now in SET territory, but with more power - Yippeeeee...
UPDATE - 6th June 2008
Ordered Active Bias System / New Teflon Coupling Caps from Diyhifisupply. Watch this Space !!