It was the early 90s, and I was working as a sound engineer and
technician at an unusual, prestigious but small recording studio in
Canada. After enduring the crazy 70s and 80s, with all its rapid but
not always improved change, the flood of early transistor gear, the
horrible beginnings of CDs, the ridiculous claims of stereo-makers and
speaker companies, the dust had begun to settle among those with the
requisite decades of experience and mixed results.
Tube amps for musical instruments stubbornly remained supreme.
By now classic tube designs had been successfully resurrected into niche
hi-fi markets, and in the studio, legendary gear had begun to be bought
and sold as 'essential', from rare Neumann tube microphones, to the
hands-down winner, the Macintosh power amps, along with exotic pieces
like tube preamps and plate-reverbs.
The problem at first in the power-amp field was the bottomless craving
for more power, and somehow in the transistor shuffle, the lessons of
the early monster tube-amps had been lost and forgotten.
Now items like Macintoshs had become so pricey that people came to me
complaining, demanding someone make something as good. The glove was
thrown down. It had erked me and intrigued that the Macs had
proprietary transformers and 'secret' circuitry.
My scientific brain demanded that no 'magical thinking' was allowed. If
great amps could be made, they could be made with ordinary parts. They
shouldn't and didn't need exotic devices or gold-plated wires, or
special magic weights on top of my speakers.
Challenge Accepted:
I thank God that I had a very understanding employer at the time, who
was willing to allow some experiments, and front some investment cash
for parts etc., as part of their admirable ethic of supporting the arts,
even in the face of the rampant flakeness that flapped about everywhere
one looked.
Just prior to my plunge, I had had my interest in tube circuits
rekindled by another kind musician, who was also convinced about tubes,
but who was as poor as me, and who wanted help rebuilding / repairing
his ST70.
I quickly figured out that the ST70 was a piece of crap, that needed
redesigning from the bottom up. The EL34s were unsuitable for high
fidelity, the preamp section was trash, the power-supply circuit was
noisy, and even the power transformer was hopelessly underpowered for
the job of running a stereo class-A amp.
We rebuilt it and I became addicted to
Glass Audio and
Speaker-builder, and began to meet a whole pile of people whom I never suspected existed.
Among the various intriguing but always ultimately unsatisfying articles and attempts at producing tube Hi-Fi, I ran into
Kimmel's breathtaking work. Without ever knowing it, he led me to
the Fountain of Youth of tube circuits, the
Mu-Follower. This was the underappreciated gold lurking under everyone's noses.
I unashamedly plagarized and adopted the Mu-Follower for intense study
and analysis. My ears had already told me something extemely important
through the headphones: The Mu-Follower, properly used, could produce
a perfect copy
of the music signal. Suddenly a dream re-awoke in me that had been
gnawing away at me for years. A power amp that could perform better
than transistors.
But so much had to be overcome: Tube power amps were notorious for their 'color'.
The main selling point of tubes was their 'musical' coloring and 'warmth'.
But I was already dancing to the tune of a different drummer; I had
been immersed in studio for so many years, and was now obsessed with
sound purity.
I dove into every power-stage design documented, to grasp the
shortcomings and benefits of each. I collected everything published in
English back to the 1930s I could get my hands on, seeking for at least
one more key, some method to overcome the systemic flaws of output
sections. The closest thing to what I was looking for was the
Circlotron.
It was indeed a brilliant design, appealing to my mathematical
topological sense, and its elegant solution to 'crossover distortion' in
Class A/B was stunning. But, alas, it did not deliver the combination
of power and utter pin-drop clarity that I was seeking.
Again I looked at the Macintosh. It was using a type of feedback
circuit within the transformer to squeeze out the distortion. But I
wanted much more. I didn't want to have the need to do that at all.
All my experience with transformers told me they operated better if they
weren't pushed into distortion in the first place. I began to study
output transformers more closely. Exotic custom-made transformers were
too expensive. There had to be a way to make ordinary transformers do
what they were supposed to.
Push-Pull circuits were the only ones that met the power requirements
needed by typical audio environments, because SE to get sensible power
required exotic high-power expensive tubes. This wasn't any better than
expensive transformers! Perhaps worse! You only buy a transformer
once.
I studied the transformer distortion more closely, in particular the
I.M. distortion; where did it come from? yes the non-linearity of the
tubes added harmonic components, and with complex signals these could
become progressively more non-musical, less related to the signal. But I
could audibly hear far more than mere tube distortion in the speaker.
The headphones made it obvious.
But OTL circuits were mostly just rumours in the 90s, and required so
many tubes that nothing was saved, and much maintenance was incurred.
This was no real solution at all.
PSRR: Real and Imaginary
I began to investigate claims about Power-Supplies and their
contribution to distortion. They were the single most common problem
for most tube circuits, adding hum and buzz, and modulating the music
signal.
Even when Power-Supply noise seemed to be under control, tube amps
sounded non-musical and mediocre. The problem wasn't solved, only
masked. The more current a circuit drew, the less filtering and the
louder the ripple. It came sneaking back in with a vengeance whenever
the music demanded power. Power supply sagging was another addition to
the problem, with its own non-linearity.
Like many a desperate builder in those days, I threw more capacitance
into PS circuits, almost recklessly, trying to eliminate noise. But the
law of diminishing returns prevented any cost-effective solution in
that direction, and indeed, such methods caused their own problems, with
damaging surge-currents and part-failures both diode and cap. But I
did finally learn how to build reliable power supplies, with delay
circuits and surge-limiters.
The Mu Follower Again:
My ears were drawn back again to
the Mu-Follower.
Why was it so pure? Why was it noiseless?
(1)
It was a Constant Current circuit. Sure. That explained
part of its magic: The Power supply was never asked to do more than it
always did. Ripple did not increase with varying music signals. But
that wasn't the whole story:
(2)
It was a Very High Resistance circuit. Most of any noise
voltage appearing in the Power Supply was divided across the high
resistance portion of the series-circuit. This second fact was what
gave the circuit an almost miraculous PSRR. Most of the noise was
dissipated as heat in a non-amplifying part of the circuit off from the
signal path.
Because of this, the Mu-Follower is the ideal preamp circuit. Critiques
aside, it reproduces and amplifies the cleanest signal possible in a
tube circuit. Its only drawback, the requirement of a relatively high
impedance load, is no limitation for a preamp: You merely add a
high-quality cathode-follower as the next stage.
Output Transformer Blues
I turned back to the Output transformer. I studied its behavior more closely.
A balanced D.C. current in its bifilar primary winding was supposed to do two things:
(1)
cancel itself out, preventing any magnetization of the iron core, which would use up precious saturation headroom.
(2)
reject 99% of powersupply hum and noise, because this also would be balanced in the primary.
(1)
cancel itself out?
I thought about it. This is actually nonsense.
In a Push-Pull circuit, the tube on one side conducts more,
while the tube on the other side conducts less.
Its current, not voltage that magnetizes iron;
the currents must swing significantly from side to side in the primary.
This means that these currents are sucking back iron-domains, and if
the current wave-form isn't a duplicate of the original signal, power is
being used in the transformer not belonging to the signal. Any
distortion products are also being distorted also in phase due to
hysteresis. To put it bluntly,
transformers amplify distortion, adding yet more.
It became obvious to me that one must prevent any distortion from
entering the transformer in the first place, at all costs.
(2)
reject 99% of powersupply hum?
Now lets imagine a 120 Hz ripple (with harmonics).
While there is no music signal,
and the current is balanced on each side of the primary,
we know each tube is acting like a static resistor (and diode in series).
The same percentage (amplitude) of the ripple voltage appears across each half-winding in the primary.
So far so good.
The voltages and opposing forces cancel:
no PS ripple is transmitted to the secondary,
and no current is induced in the iron core either.
Now suppose the music signal is a 30 Hz bass-note.
For the first two cycles, the music signal unbalances the two tubes:
one conducts more, and the other less current.
The voltage also rises at one plate and falls at the other.
now there is momentarily a "D.C." (slow changing A.C.) across the whole primary.
In slow motion if you will, one tube acts as a large resistor,
and the other acts as a small one.
More current flows down one half-winding, magnetizing the core.
(so far so good, because this is the music signal).
But now the PS Ripple also appears more on one leg than the other,
because of these slowing changing voltage-dividers.
The PS Ripple, modulated by the music signal is pushed into the secondary as an AC voltage.
This happens four times (in alternate directions) every 30 Hz.
But most critically important, it is happening most of the time!
During this low music signal, there is darn little PSRR!
The same rule applies again:
We must never let the PS ripple appear across any part of the primary!
One obvious rule, if there weren't enough reasons already,
is
not to include a transformer in a multistage feedback loop!
What circuit achieves both goals?
The Mu Follower.
It became obvious to me that not only would the Mu-Follower work in the Output stage, it was essential!
I built several amplifiers, on a commission basis during the 90s,
as a result of my findings.
I also discovered that even the tubes chosen were not really critical,
except to meet power and current-handling objectives.
I used tubes from WWII, vintage, NOS, and modern tubes,
all with great results.
I received a very complimentary review by a writer for
Stereo Times, as a personal endorsement.
Quote:
“Let me tell you about an amazing experience I had not long ago.
I was sitting in a local recording
studio, listening to my friends lay down the tracks for their upcoming
CD. The music consisted of a vocalist, stand up bassist, some
percussion, and two guitarists; one acoustic and one electric. As I
listened to the live mic feeds, and subsequently the master tape
playback of this music, I was struck once again by the notion that home
playback just doesn’t sound like this. Even through the small, modest
speakers being used as monitors there was a richness of tone, attack,
transparency and effortless high frequency detail that was stunning.
There was a sense of live musicians coming through those monitors that
escapes even the most expensive home systems I’ve heard. In contrast,
most CD’s sound "canned," like all the lovely aural information has been
truncated, stuffed into a playback medium that makes a two dimensional
caricature out of the original sonic event. So, I sighed and enjoyed the
experience, realizing I could never bring that sound home with me.
Here’s the kicker. I wandered into a
room down the hall from the recording studio where (Nazaroo), a
technician working for the studio, had set up his newly designed mono
block tube amplifiers. The amps were driving a tiny little pair of NHT
"1.5s." I’ve heard those NHTs several times before; nice speakers,
pretty flat response, low coloration design, but nothing that’s ever lit
my fire. For kicks, from a nearby shelf full of CDs I grabbed an
Cassandra Wilson CD, loaded it into the non-descript CD player hooked up
to the system and pressed play. My jaw dropped. I’ve been trying to
scrape me jaw off the floor ever since. What I heard was an incredibly
pure sound. So pure that the sound coming from this commercial CD was
startlingly close to the experience of listening to my friend’s master
tapes in the recording studio. There was none of the electronic hash
overlaying the music that I was used to hearing through consumer sound
systems. Instead there was a harmonic purity, clarity and lushness that I
was accustomed to hearing only in the recording studio. I immediately
recognized that I was hearing a level of sound quality that I’ve rarely
encountered before, even among the heart-stoppingly expensive systems
that populate high end audio. And all this coming from amplifiers
putting out around 20 watts into a relatively cheap, small pair of
speakers. Clearly (Nazaroo)’s amps were doing something special.
So I had (Nazaroo) build me a pair of
mono block tube amplifiers. These put out approximately 28 watts a side
- a little more powerful than the ones I listened to at the studio.
They sound like a fairly ‘powerful’ 28 watts - enough to drive most
mini-monitors easily. In my smallish room they have also acquitted
themselves quite well in driving my Von Schweikert VR-4 Gen II speakers -
a large, full range floor standing speaker with a 90 dB sensitivity and
6 ohm impedance.
Hooking up (Nazaroo)’s amplifiers to
almost any speaker makes the speaker "disappear" as a sound source. The
amplifiers produce a signal that is so clean and transparent that the
low-level ambiance of the recording is revealed. It’s like the very air
between your speakers is altered, turned into the recorded space of the
CD, and inhabited by palpable sonic images of musicians playing their
instruments. Listen to a recording of someone playing acoustic guitar.
The high strings of the guitar have a life-like sparkle, without any of
the electronic edge or etching to the finger picking that is added by
cruder amplifiers. This unnatural electronic etching normally obscures
the naturally beautiful overtones of the acoustic guitar. (Nazaroo)’s
amps remove these electronic impurities, allowing you to hear the
differences between every string, and the resonating wood body of the
guitar. Hell, it sounds like you can hear right inside the body of the
guitar. In classical music recordings, woodwind instruments in the back
of the orchestra that are often obscured or homogenized by other audio
equipment are rendered clearly, with each instrument being easily
identified, and their melodic line effortlessly followed by the ear. You
can hear the texture of a reed buzzing in an oboe, and the vibrating
column of air emitted by the body of the instrument.
And it goes on and on. Each recording
I’ve played through (Nazaroo)’s amplifiers has sounded unique, with
it’s own recorded ambiance, and pallet of mic techniques being
disclosed. The sonic signature of individual stand-up basses, classical
guitars, pianos, saxophones etc. are unveiled. (Nazaroo)’s amplifiers
simply offer a cleaner, clearer sonic path to the original recording
than any amplifier I have encountered. Even my very expensive Conrad
Johnson Premier 12 monoblock amplifiers are revealed to somewhat color
and obscure the sound in relation to (Nazaroo)’s design. The same is
true for other well-regarded amplifiers that I have had at my house. I
have not yet heard a more transparent audio signal. Plus, when paired to
appropriate loudspeakers, (Nazaroo)’s amplifiers deliver among the
richest, tightest bass I have heard from any tube design.
So, is there anything more I could
ask from these little marvels? Well, yes. In (Nazaroo)’s amplifiers I
have encountered something marvelous, but not perfect (I’ve yet to
encounter *any* perfect audio component). I find that, in relation to my
CJ amplifiers and some other good tube designs, (Nazaroo)’s amplifiers
can sound a little less fulsome in the upper midrange. Female voices,
while startlingly clear and natural, can loose a little of the physical
density or presence that I get with my CJs. Same goes for instruments
that inhabit or play in the upper registers, such as alto sax or
trumpet. On my CJ amps these instruments move a little more air in my
room. Also, one must be careful to pair (Nazaroo)’s amplifiers to the
right speaker. I’ve had these amplifiers exhibit excellent control over
the bass region of some very large speakers. But if the speaker
sensitivity drops too low (or the woofer size gets to big), watch out,
the bass can get a bit out of control. Hey, we’re only dealing with 28
watts here, so we must be sensible.
I should add that the above caveat
regarding a thinness in the midrange seems to be a minor point. While I
do hear it, many of my audiophile/musician guests who listen to music
through these amps do not pick up on any thinness. In pretty much every
case where I have A/B’d (Nazaroo)’s amps against a competitors my
guest’s have preferred (Nazaroo)’s, feeling they offer a substantially
more realistic sound. “
So, there you go; right now,
(Nazaroo)s’s amplifiers offer a level of sonic purity that would be hard
to match, even by some of the ridiculously expensive amplifiers offered
by other high end amplifier designers. If he continues to perfect his
design, I feel he will have a product that, at a fair price point,
should make other high end manufacturers hang their head in shame. “
- Reviewer's name withheld
