darTZeel Audio NHB-108 B version - Manual (Page 24)

darTZeel NHB-108 model one
User Manual, version B 1.0 Page 24 of 35
So for now, and maybe for still a little while, we
will probably not use them directly in audio appli-
cations.
T5.2.2. FETs
FET stands for "
F ield E ffect T ransistor", of which
there are two main categories. For small signals,
they are called as is, FETs.
For power applications, their most common name is
MOSFET, which stands for "
M etal O xide S ilicon
F ield E ffect T ransistor". They behave similarly to
FETs, but their internal structure may vary.
FETs are being used more and more, for several
reasons, including the most obvious one:
It is commonly accepted that they perform very
closely to vacuum tubes (or electron tubes or
valves). Tube lovers generally like MOSFET ampli-
fiers due to their similar behaviour.
Here, we are forced to say that we do not share this
point of view. Not entirely, at any rate…
It is true that the sound of MOSFETs tends to be
soft and warm, a bit like those found in tubes. But
their electrical behaviour is not similar, simply
because they are not made of similar materials.
The only true similarity is that MOSFETs – and
FETs of course – are voltage driven, like tubes.
Apart from that, they are in two different worlds,
vacuum for tube, silicon for FETs.
Conduction in any type of transistor always acts in
a solid, here the metallic silicon, hence their name
of "solid-state" devices.
In a vacuum tube, electrons move in… a vacuum.
Coming back to FETs, what is their sound? For
most designs, we would say that it is only a matter
of taste. We are convinced that one can design a
"standard", good sounding machine in the 3 tech-
nologies mentioned above, and even with the fourth
- tubes - with virtually the same sound.
Obviously, if you like music, you definitely need a
darTZeel NHB-108 model one
.
Yet, and contrary to common belief, MOSFETs are
less linear than bipolar transistors available today.
MOSFETs often need more sophisticated added
circuitry to make the best of their possibilities, and
that means a longer and more complex signal path,
and hence a reduced resolution.
MOSFETs are also slower, in absolute terms, than
bipolars, and are much harder to drive in pulse
mode.
Reproducing very fast transients requires a lot of
instantaneous current drive, not easy for the power
supply to properly feed, so temporal integrity could
suffer.
T5.2.3. Bipolars
These have been in existence for more than fifty
years. The industry knows them very well. Today's
bipolars are better than ever. Ultra linear and ex-
tremely fast, they are perfectly suited for very high-
end audio applications.
Their excellent linearity makes it possible to mini-
mize the number of peripheral components, thus
obtaining a straightforward signal path, with many
less superfluous correction circuits.
In the NHB-108 model one, the audio signal trav-
els through only 6 transistors, from input to output,
maintaining low THD and IMD levels, and without
using any global NFB, the output stage even operat-
ing in a totally open loop! The slowest transistors
used have a bandwidth of more than 30 MHz, much
higher than for a MOSFET. This extreme intrinsic
speed allows a total phase respect across the whole
audio range, without any static or dynamic defor-
mation. In brief, no Temporal Distortion.
In terms of music, these breakthrough advantages
bring you closer to the musicians. Now you can
share the emotion of the violinist, the fire of the
conductor or the sweetness of the flutist.
Words are however not enough to describe what a
single bipolar transistor pair is capable of. The best
is still to invite your friends home, making them
understand, share and believe your ears.
T5.3. The circuit of darTZeel
T5.3.1. Criteria of choice
Simplicity.
The audio circuit of the NHB-108 model one uses
only 14 transistors in all, including current sources.
Only three different bipolar transistor devices are
used, all of the same brand.
Purity.
The version B of the darTZeel NHB-108 model
one
takes unprecedented care of the precious musi-
cal signal:
There is absolutely no connector, switch, fuse or
relay (save for the XLR inputs) in the signal path.
Even better, we did not use any of these devices
except in the AC mains input.
By using one single pair of output devices, and by
not limiting the output current, we are able to obtain
unbeatable dynamic reproduction.