Consoles 867the (hopefully) optimized input pair reigns. This is
mostly down to the impedances chosen around the
op-amp differential amplifier; some resistor values here
have been seen as high as 25 k:, guaranteeing a high
invariant noise floor for much of the gain range;
bizarrely, some implementations attempt gain around
this stage, too. The lesson earlier learned of keeping the
circuit impedances as low as circumstances permit is
salient here.
There are few reasons now for hand-knitting discrete
versions of this input arrangement, and good ones not
to, primarily suitable input devices: the much-favored
2SB737 is sadly becoming difficult to procure, and the
excellent integrated matched transistor sets such as the
THAT 320 are relatively expensive. Relative, that is, to
the cost of a purpose-designed IC! There are a few inte-
grated versions of this kind of arrangement offering
very acceptable performance in the convenience of a
little package; the Burr-Brown INA103 or 163 and the
THAT Corporation’s 1510 have multiparallel input tran-
sistor stages presenting OSIs about perfect for nominal
microphone impedance, the latter part taking to heart
the op-amp differential stage noise issue, with excellent
results. Texas Instruments / Burr-Brown’s PGA2500 is a
digitally gain-controlled version of this configuration,
which works very well indeed with fine gain steps and
inaudible gain-change (zipper) noise, solving a major
headache for digital console designers and others
desirous of remote control of mic gain. They have
proved themselves worthy with ribbon microphones, the
ultimate front-end-noise torture test.
With this configuration, although potentially offering
far higher and flatter input impedances than transformer
inputs, there are, as always, snags. Common-mode
signals directly gobble up head room in the first pair of
stages even if they are operating as followers; that this
common-mode stuff is substantially canceled in the
following differential amplifier is a bit of a stable-door
and bolted-horse routine. There is also the great danger
that common-mode signals (in addition to normal differ-
ential signals) can exceed the input swing capability of
the input devices. At best this will block the input stage,
at worst—if the common-mode signal is big enough and
at a low enough impedance (think ac line grounding
fault here)—serious destruction can result. A trans-
former input would blithely ignore such.
Radio frequencies simply adore base-emitter junc-
tions, and this configuration has them in abundance.
Successfully filtering microphone inputs sufficiently
without sacrificing noise performance or input device
high-frequency gain (increasing high-frequency distor-
tion and so on) is not a trivial task; it makes the
self-filtering properties of an input transformer seem
rather appealing.
Fig. 25-46C details the kind of helmet-and-armor
with which one has to attire electronic microphone front
ends to survive the fray:- The S L/C filters and additional coupling to ground
at high frequencies help against RF. The inductors
can either be single pieces, one in each leg, or prefer-
ably a pair of windings around a common (usually
toroidal) core. This has the twofold advantages that
the choking effect is concentrated on common-mode
signals—the most common (so to speak) interfer-
ence method—and that the inductances of the two
windings essentially cancel for differential signals, so
that there is much less effect of the RF protection
impinging on the desired audio. - The input dc-decoupling capacitors have to be pretty
huge in value to maintain low-frequency integrity
and at the same time have a high enough voltage
rating to handle typically 48 V off-load phantom-
power voltage. - The parametric varactor capacitance of the clamping
diodes has little to no effect on the audio but are vital
to protect the device front end from the very healthy
whack as the phantom voltage is turned on or off.
Even so, clamped or no, good common-mode rejec-
tion or no, phantom ramped or no, these transitions
are not exactly subtle. These diodes may help protect
the input from other nasties, too, but the aforemen-
tioned major ground fault would render all this toast
anyway. The good news: such are rare but in the real
world of touring sound not unthinkable.
25.10.6.9 Line Level InputsThe reader is referred to Chapter 11 for Bill Whitlock’s
excellent further coverage of real-world interfacing.
High-level balanced interconnections and systems
for the most part have been largely relegated to outside
world and intersystem interfacing; internal intercon-
nects are left unbalanced except within a very few
completely balanced console designs. The wisdom until
quite recently was that balancing implied transformers
and their performance limitations.
Transformers—good ones at least—are expensive,
large, and heavy. Not so good ones are still relatively
expensive, large, and heavy and represent a weak link in
a modern signal path, with their low frequency distor-
tion, hysteresis, and high/low frequency/phase response
effects. Transformers are best used only where their
impedance transformation capabilities, innate filtering,