Why Tube Gain Stages?
The stupid answer is the name on the unit is "Manley Labs" and that
is what we do. Unlike a current trend, we do not use tubes for THD,
clipping character, cool marketing buzz-words, or plagiarism. We
began building tube gear because we preferred the sound when it
was un-fashionable and re-introduced these glowing gain bottles to
both the hi-fi and studio communities when there was virtually no
fresh tube designs available. We also stress that, its not just the
tubes, but the way they are used. The sound of a piece of gear is due
to many design details and many of the components - always has.
In the Massive Passive, the tube gain stages are new designs
developed for this unit. We try to use minimalist techniques where
ever possible and use the appropriate technology for the purpose.
Simple vacuum tube circuits can excel for medium gain voltage
amplifiers and high headroom output stages. A simple tube stage
offers better linearity than an equally simple transistor circuit.
Transistors have a logarithmic transfer curve and are essentially
current devices. Transistor circuits are typically built with huge
excess gain which is used for more feedback in order to tame the
linearity (THD) but this feedback seems to cause audible transient
problems and is directly responsible for the harsh clipping character.
Op-amps, which can be less noisy and lower THD, are complex
circuits which force music through many transistors and may also
bring crossover distortion artifacts and headroom issues into play.
The only alternative would have been for the Massivo to use FET /
MOSFET high voltage discrete circuits. Someday, we may introduce
a version like this but don't phone us up every month asking if we are
working on it. We'll let you know.
We use an exceptional op-amp/discrete circuit for the input buffer
in order to drive the 150 ohm (worst case) EQ circuit. Not an
appropriate place for tubes. The first design used a transformer (3:1)
for impedance conversion but it had a 10 dB voltage drop and thus
10 dB more noise. The new input circuit isolates input loading and
allows the tube circuits to be better optimised. We use two similar
all-tube gain stages per channel for interstage and output line drivers
which together cancels some distortion. The output is capable of
driving up to +37 dBu! This stage also uses a separate winding on
the output transformer (also custom designed) for a little negative
feedback to allow lower output impedances and minimal transformer
coloration. In other words, because we expect some engineers
wanting to boost 20 dB at 100 Hz occasionally, the circuits had to
be capable of cleanly delivering it (regardless if the next piece in the
chain can deal with it). We used tubes for more headroom (300 volt
power supply) rather than more clipping. Generally too, tube circuits
clip a little smoother than mega-negative feedback IC circuits.
Some may question "tube reliability" but most major studios have
many 30 or 40 year old tube compressors and EQs running every day
and some with the original tubes. Not many 15 or 20 year old
transistor units are still working or wanted. Tubes will eventually
burn out (so do transistors), however, usually you can easily get the
type of tube used 30 years ago and you won't need a soldering iron,
schematic or technician. Your parents probably used to "fix" the old
TV. The bottom line is, good gear tends to be more reliable, and if
a problem develops, is both easy to fix, and carries great factory
support. We understand that it sometimes involves your professional
livelihood and this is indeed often serious and you depend on it. If
this is the case, consider getting a few spare tubes which covers
90%+ of emergencies with immediate fixes. Our service department
has a great reputation with phone support and fast turn-arounds too.
Phase Shift?
Deadly topic. This is probably the most misunderstood term floating
about in the mixing community. Lots of people blame or name phase
shift for just about any audio problem that doesn't sound like typical
distortion. We ask that you try to approach this subject with an open
mind and forget what you may have heard about phase for now.This
is not to be confused with "time alignment" as used in speakers, or
the "phase" buttons on the console and multi-mic problems.
First - all analog EQs have phase shift and that the amount is directly
related to the "shape" of the EQ curve. Most digital EQs too. In fact,
one could have 3 analog EQs, 3 digital EQs, and an "acoustic
equivilant", and a passive EQ, each with the same EQ shape, and
ALL will have the same phase shift characteristics. This is a law, a
fact and not really a problem. The two exceptions are: digital EQs
with additional algorithms designed to "restore" the phase, and a
rare family of digital EQs called FIR filters based on FFT techniques.
Second - Opinions abound that an EQ's phase shift should fall within
certain simple parameters particularly by engineers who have
designed unpopular EQs. The Massive Passive has more phase shift
than most in the filters and shelfs and leans towards less in the bells.
Does this correspond to an inferior EQ? Judge for yourself.
Third - Many people use the word "phase shift" to describe a nasty
quality that some old EQs have and also blame inductors for this. Its
not phase shift. Some inductor based EQs use inductors that are too
small, tend to saturate way too easily, and create an unpleasant
distortion. The Massivo (of course) uses massive inductors (compared
to the typical type) which were chosen through listening tests. In fact
we use several different sizes in different parts of the circuit based
on experiments as to which size combined the right electrical
characteristics and "sounded best". The other very audible quality
people confuse with phase shift is "ringing". Ringing is just a few
steps under oscillating and is mostly related to narrow Qs. It is more
accurately described as a time based problem than phase shift and
is far easier to hear than phase shift. For our purposes, in this circuit,
these inductors have no more phase shift or ringing than a capacitor.
Fourth - A given EQ "shape" should have a given phase shift, group
delay and impulse response. There also exist easy circuits that
produce phase shift without a significant change in frequency
response. These are generally called "all-pass networks" and are
usually difficult to hear by themselves. You may have experienced
a worse case scenario if you have ever listened to a "phase-shifter"
with the "blend" set to 100% (so that none of the source was mixed
in) and the modulation to zero. Sounded un-effected, didn't it, and
that may have been over 1000 degrees of phase shift. Group delay
and impulse response describe the signal in time rather than frequency
and are just different ways of describing phase shift. Some research
shows these effects are audible and some not. The Massive Passive
tends to show that group delay in the mids is more audible than
towards the edges of the spectrum and there may be interesting
exceptions to generalities and conventional wisdoms. The audible
differences between EQs seems to have more to do with Q, distortions,
headroom and topology than with phase shift.
Fifth - Phase Shift is not as important as functionality. For example,
we chose very steep slopes for some of the filters because we
strongly believe the "job" of a filter is to remove garbage while
minimally affecting the desired signal. A gentler slope would have
brought less phase shift but would not have removed as much crap.
12
