LEVEL PREAMP 2003.
Fig 1. Amp on bench,
Fig 2. Rear amp on bench,
Fig 3. Schematic of 1 channel of amp,
Full explanations are included.
This preamp was custom built in about 2003 for a customer who
wanted to try something with tubes to see
if it made his music any better. He was delighted.
Pardon the morning break mug of coffee and apple, but they
indicate the size of the amp and its ultra simple
remote control inside a block of re-cycled mahogany.
Remote control is only for volume up or down only. Sensible motor
speed is used to control an Alps pot, 27mm
square body. Manual volume control is possible. The aluminium case
work is hand made, machine engraved,
powder blasted to an even matt finish then anodized. A blue LED
indicates on/off, and 5 sources are selectable,
and there is a pair of record outlets.
The infrared "sender" electronics and battery is within the small
wood block. The IR receiver in the amp drives a
motor to turn the pot.The inside of chassis volume is packed with
PSU and point to point wiring.
The rear panel has mains input, fuse and each channel has 7 RCA
5 inputs, 1 output, and 1 record output.
Here is one channel of the line level amp. No balance control was
included because the customer agreed there
was no need for one. Gain is set for less than 2 times because the
only source the owner has was CD, or other
fairly high input sources. There are Constant Current Sources
anode or cathode loads to convey Idc to 12AU7
so that Ia = about 4.0mAdc for idle current for each 1/2 of the
The CCS use means the dc carrying load element to the tube acts as
a resistance with R > 2Mohms, so that there
is no ac loading by a resistor from B+ to anode or from B- to
cathode which would have values of about 30k, and
with other AC coupled loads such as the Volume pot of 50k after V1
anode, plus NFB resistance of 120k, the total
load = 35k.
The result of using the CCS dc supply transistors as passive high
ac impedance dc elements much reduces the
distortion by approximately -10dB, ie, the CCS use reduces THD by
a factor = 1/3, a welcome amount which probably
translates to a much larger subjectively perceived betterment in
The THD is somewhat determined by the level of signal before the
volume pot. Let us assume the CD maximum level
is 1.4Vrms, then maximum V1 anode signal would be approximately
16Vrms, and THD = 2% if there was zero shunt
NFB, ie, R3 and R8 were omitted. The use of the volume control at
the 12 o'clock position would give Vo = 0 1.6V, and
the output cathode follower would slightly reduce this to 1.4Vrms,
and thus create far to much volume level in most
power amps which need only 1Vrms to cause clipping.
Therefore the use of a shunt NFB network using R3 and R8 is needed
to reduce overall gain at maximum volume setting
= 1.7 times. With maximum CD input = 1.4Vrms, V1 output = 1.7 x
1.4 Vrms = 2.4Vrms, and THD < 0.1%, because
the NFB is reducing the THD by factor = 1/4, and much less signal
exists at the V1 anode.
The average level from a CD player may be only 0.2Vrms, and the
preamp THD is approximately proportional to output
voltage, so it will be found that THD under normal operating
conditions = < 0.03%, and quite negligible.
The maximum average signal levels ever wanted by most people is
0.25Vrms and the low gain of the preamp provides this
easily. This suits people with normal 50W power amps which require
a maximum 1Vrms of input for the maximum 50W
The sound from this type of circuit has the natural warmth of the
music, excellent detail and dynamics, a sense of
smoothness and spaciousness and precise imaging. This type of
preamp is the very essence of what captures the soul of
the music and warms hearts of anyone listening. It will seem
to smooth out the harshness that may otherwise accompany
the efforts of some of the more raucous singers like Celine Dion;
it will make an old Italian violin or cello have you in tears,
and piano will never sound so good, massed strings will sizzle
with warmth without the searing harshness of solid state.
Brass will have sparkle without edge.
During the years I've made such preamps, people have found that
replacing their solid state preamp with a tube pre-amp
such as the one above has done wonders for the sound even if they
have a solid state power amp. One fellow had a listen
to this type of preamp and asked if I'd accept a well known brand
of solid state preamp as a trade in for a preamp like the
2003 model and I could only politely refuse him because how would
I ever sell an SS preamp after anyone with good ears
had a listen to a well done tube preamp?
After many years the 2003 preamp above is still working perfectly
and the owner has tried a few different 12AU7 from the
suppliers around the world since different brands can have a
slightly different timbral quality, or voice. The measurements
would all be nearly equal. Why some tubes sound different has
never been fully explained.
One may find many fine wines from different wineries to be all
very drinkable, and enjoyable, and nobody worries whether
they have the same chemical makeup. So it is with many tubes from
different makers. I supplied the best 4 selected 12AU7
twin triodes from my collection of about 40 12AU7 at the time so
my customer had a pair of spares.
I rigorously test tubes for low noise and low microphony and try
where possible to carefully match them for brand and gain.
Usually with 12AU7 the gain is within a dB between samples and
there is no need for a balance control.
The 12AU7 is an old fashioned choice of tube because these days
the premier makers like to use 6922 / 6DJ8 / ECC88
or the Russian 6H30. I usually prefer 12AU7, or 6CG7, or 6SN7 and
I don't think more modern type numbers such as
6922 can sing as well as the best of the NOS 12AU7 where they test
well for noise and microphony.
Some more recently made Russian tubes and selected NOS do not have
low noise and low microphony.
Usually I find the tubes that pass the quality test for low noise
invariably sound glorious, probably because the tube is as
perfect a sample of manufacturing as possible, with tight fitting
electrodes and a hard vacuum and a well made cathode.
These aspects of a tube cannot be judged by the appearance or just
by testing with a tube tester. The electronic testing
I perform reveals the real condition of a tube beyond the
capability of a "tube tester" which only tells anyone if the
tube works or not, ie, if it has the rated correct emission.
The other fabulous triodes such as the 6SN7 and 6CG7 have slightly
more gain than the 12AU7 but they will work
almost identically as the 12AU7. But the 12AU7 needs a 9 pin mini
socket, and uses 12.6V at 0.15A across pins 4 to 5
for heating but 6CG7 needs 6.3V at 0.6V. 6SN7 needs an octal base,
so such tubes are not interchangeable.
12AT7 or 12AX7 may be used instead of 12AU7, with alterations to
cathode R and CCS current setting R.
All have been tried, and the 12AU7 or 6CG7 are probably best.
Usually, if 6CG7 is used, then 6922 or 6DJ8
may be used without any circuit alteration.
I don't make a lot of preamps, and I doubt I will attempt to make
a similar small sized preamp such as the one pictured
which delighted the owner who didn't actually know exactly what he
would be getting when he ordered it.
But when he did get it the appearance matched his expectations and
it blended in well with the rest of his gear.
I now prefer to make a line stage preamp slightly larger.
In October 2011, I have begun to make a line level preamp using
sub-miniature triodes which are single unit triodes
only 35mm long, dia = 8mm, so size is quite a lot smaller than a
normal 9 pin mini tube. These tubes do not use a
plug in socket and have leads which solder directly to a circuit
board, thus allowing considerable saving in circuit size.
Such tubes may be clamped to a chassis with copper or alumimium U
straps to aid cooling. The sub-mini tubes were
originally produced to suit wearable hearing aids used with
battery supplies but later were used wherever compact size
was required especially in military gear for aircraft where
reliability was imperative. Their production has long ceased,
but stocks remain. Their production ceased because of the change
to solid state in about 1960 for 99% of most
applications. I'm planning to use a similar circuit for the next
preamp as in the 2003 model, but instead of using MJE340
for CCS, it is just as easy and effective to use a sub mini
triode. So a total of 8 individual tubes will be used and they
will fit easily to a chassis about 350mm long x 150 deep, x 55mm
high with a PSU box encasing PT and diodes at one
end and heatsink and triodes at the other, and with plenty of room
for the remote volume control using a stepper motor
to drive a DACT switched attenuator.
Back to Index Page