The "Nemo" line level preamp and "Bunnerong" power supply for line and phono stages :-

The "Nemo" line level preamp is on the left and it might need a better name.
On the right is the "Bunnerong" power supply.
( There is an old coal fired power station in NSW at Bunnerong, and a friend
said that the preamp power supply looks like that power station.
There is something about all those large value electrolytics, with potted power
transformers at the rear.)

The Nemo is in its 3rd  regeneration; it should be close to heaven because it has
been born again 3 times. It started with a simple schematic with common cathode
gain triode of 1/2 a 6CG7, resistance loaded, with direct coupled cathode follower
output. Conrad Johnson used such a simple good sounding circuit way back in
about 1955, and very little improvements have ever been nade since then.

Initially, in 2000, a DACT attenuator was placed before the gain tube, Siemans
NOS 6CG7.
The Siemans µ is about 21, so gain became 18x, or 25dB, and a little too high.
The owner who initially bought the amp sold it to its present owner in 2004.
After some use, the new owner regarded the high gain as a nuisance, and
felt that digital recordings tended to sound too "toppy", harsh, on his Vienna
Acoustic Motzart speakers.
So I then tried a µ-follower instead of the 1955 common cathode gain stage and
cathode follower. The DACT remained ahead of the µ-follower, and a shunt NFB
loop added to reduce the high gain. In the NFB loop, there was a switchable set
of 3 x R&C series networks to give a lowered and shelved HF response for
treble above 2kHz. These two new features sounded well. NOS Siemans 6CG7
still tended to sound better than Russian 6H30 which we tried.

The present form in 2011 is unchanged from 2005, and is described in the
schematic and notes below.
It sounds quite excellent with Sieman's NOS 6CG7. Gain has been reduced
to only about 12dB with the shunt FB which also allows the shelving networks
to be included in the FB loop. Overall, there are less HF losses in the signal
path and balance is more accurate.

Controls in the picture from left to right :- source select, HF shelf levels, gain,
mains on/off, phono amp on/off.

The line stage for dec 2005....

The line stage amp only has one gain triode and one output cathode follower,
so one pair of 6CG7 is all that's needed for both channels.
V1 anode has its anode dc supplied through the MJE350 constant current source
which was selected to keep the ac load on the tube high, and THD low.
The actual collector resistance is over many meg-ohms, and the transistor
acts passively, and does NOT inflict any sonic signature. The triode's
best sonic abilities are much aided by the presence of the constant current
source.
The input signal from the signal source selector feeds the Record Out for those
wanting direct connection for recording purposes. The input signal also is applied
to the R1 resistance which is the first R of the shunt NFB network formed by
R1 and R5. It may be assumed that whatever source is selected, it will have a
low source resistance of say 600 ohms. Where source resistance is higher, the
amount of NFB applied becomes higher, but few sources today have Rout
above 10k0.
The gain of the 6CG7 with the R5 270k load and the following DACT 100k
switch is about 73k, so the triode gain = 18x approximately.
This means if 0.1Vrms+ is applied to the grid, then Va = 1.8Vrms-, and the
phase of Va is opposite to grid signal.

So there is 1.9Vrms appearing across R5 270k, so current = 0.00703mA.
It isn't much, only 7.03uA. Grid input resistance is many meg-ohms.
But there is 0.1Vrms across R2 470k, so current = 0.213uA. The R2 is
used because it biases the grid to 0V if the source signal is a "floating"
signal that has no biasing resistance, or the source is from the output
side of a capacitance with no grounding resistance. An alternative is to use
a 100k R between source switch pole and 0V. But doing it my way keeps
the input resistance at just above 47k.

Now the current in the R1 47 must be the total of the currents in R5 and R2,
ie, 7.23uA, so the voltage required across the 47k = 0.3398Vrms,
or say 0.34Vrms. The signal input from the source must become 0.34 + 0.1Vrms
= 0.44Vrms+. This means that with NFB you need 0.44Vrms+ input to produce
output = 1.8Vrms-, and signal voltage gain with NFB = 4.09x, or +12dB.

Hands up who didn't have a clue what the last two paragraphs were about?
Gee, I see a sea of hands. But that does not matter, the point is that the gain
is only 12dB, or 4x, and that's enough, because a CD player produces up to
1.4Vrms so the V1 anode signal maximum will reach 5.6Vrms.
The THD at this level without NFB would be approximately 0.2%, and with
the NFB it is reduced to about 0.05%. Now most of the music level will be
1/20 of the maximum level when THD will be also 1/20 of the maximum
level, so expect to measure less than 0.01% for 90% of whatever you like
to listen to. Meanwhile, the signal to noise ratio is just fabulous with the
source input raised in level before being applied to the DACT volume switch.
Placing the volume attenuator BEFORE the gain triode means that the source
signal is reduced hugely, then applied to the gain stage, and the noise of the tube
gain stage is at a much higher relative level to the signal, and SNR is much
worse than having the volume attenuator after the gain stage, as I have it.

Now in addition to R5 = 270k, there is also R6, R7, R8 and C3, C4, C5.
At LF below say 1kHz, these networks have very little effect, but above 1kHz
they begin to reduce the total resistance between the grid and anode output.
The netoworks may be switched to give a shelved HF response above 2 kHz
so that nasty harsh recordings can be tamed a little if required.

The output signal from the pole of the 100k log DACT volume switch feeds
the grid of V2 cathode follower output buffer which has fixed bias.
Such cathode followers have utterly transparent sonics, and prevent HF losses
caused by long cables or high shunt capacitance of some horrible solid state
power amplifier. The THD generated in V2 is reduced from about 0.4%
without NFB at 5.6Vrms output to about 0.03% by the "follower connection"
which itself is application of about 16dB of local series voltage negative feedback.
At 1/20 of the max level, THD will be theoretically 0.0015%, so nobody need
worry that their triode amp is the cause of the distortion that they may claim
to be able to hear.

There is some shunt regulation of the B+ voltage with a zener string.
When the phono amp is separately switched on or off, there are no strange
LF noises or cone wobbles in speakers due to B+ rail variations.

I do not believe that Auricaps or many other brands of coupling capacitors
sound any better after having recently trialed the above line stage where
Wimas were in one channel and Auricaps were in the other.
A customer friend and I used the same mono sound source through each
channel in turn with me trying to trick my friend when I asked him to say
which channel was better. After about 6 changes of with 2 different recordings,
my friend could not pick any change or state any preference which was better
than chance, ie, he liked the Wimas just as much as the Auricaps.
I certainly could not hear any difference let alone a "better" sound with Auricaps.
However, my friend proceeded to have me replace all the Wimas in his
preamps and power amps with Auricaps.
I will always consider that my customers are always right, and work as
directed, but I don't myself think I am missing out on better sound because
I have not used Auricaps in my own system. 

Between 2006 when I wrote about the Wimas Vs Auricaps, and now, 2011,
the Auricaps remain in most of my friend's amps, except for one where he
asked me to instal some Russian made caps, probably polyproylene, but army
green color suggesting they were spare parts for Russian Army electronics.
My friend has another friend who talks to other friends and the consensus
of the "friend network" stretching across the world and including mobs of
typing bullshit artistes is that Russian Brand X caps sound the best, therefore
must be installed, "or I won't listen to any music at your place when I visit".
I've known audiophiles to change capacitors every 6 months until nearly
every possible available brand of capacitor has been tried, each lot better than
the last lot. But not one A-B test was ever conducted. Oh no, the irrational
audio nutter can't stand logic or common sense, and he cannot let himself ever
be cornered into a position when he must rely solely on his powers of hearing
to discern what sounds best.

I do not think many of the myths about special parts have much validity.
The circuit design and careful tube choice are far more important to the
sonic signature than brands of the same type of cap, or different types of hook
up wire or solder or RCA sockets or speaker cabling.
The main reason to use better quality parts is reliablity and tolerance quality.
For example a cheap Taiwan made dual gain potentiometer costing $4.00 may
have 15% difference in L and R levels at the -20dB gain setting but otherwise
work perfectly - for 3 years, maybe, if used a lot. Then it begins to make noises.
I won't use less than an Alps Black pot which is much more expensive but they
last 40 years. Well, we might assume that, if the present production of Alps
pots is as good as it obviously was back in 1980. There was good matching on
their dual tracks at all levels, and the carbon tracks didn't wear through easily
by countless rotations by the metal slider running around the circular carbon track.
But now its possible the Alps "Black" 27mm square bodied sealed potentiometers
are being made in mainland China or some other place and its difficult to say
present production quality is the same as its been in the past. The last lot I bought
from RS Components Australia measured well, and the price has come down by
nearly 50%, which suggests RS is now sourcing Alps parts from a supplier who
is far cheaper. An Alps Black pot would be an easy thing to copy.

The above Nemo has a DACT switched attenuator which is still working well
after many years, and the level match is much better than an Alps pot which
usually within +/- 0.5dB for L&R channels, and good enough.
I doubt there is any difference in the sound of either.

Tube choice for all the above preamps does have an effect on the sound.
This is true especially with the Rocket which has no global FB for RIAA eq.
Any tubed phono amp without loop NFB such as the Rocket is more prone
to trouble from noisy tubes and microphony.
Unless tubes are properly tested for noise and microphony, not just tested
in a tube tester, then the sonic outcome may be sub-optimal. I find the least
noisy and least microphonic tubes give the best performance in terms of detail,
dynamics, conveyed natural warmth, musicality, tonal balance, sound stage,
imaging and air.
Tubes that are old and tired, and possibly not so NOS as you may think they
were when you bought them will perhaps sound less than optimal. It does no
harm to have a few different tubes to try, but one should never change tubes
each week because the pin grippers in tube sockets may loose their spring
pressure and become loose, resulting in noise, intermitancy, and a repair bill.
This is particularly so with some Chinese made brands of tube sockets
which do have nice ceramic mouldings, but have appalling metal alloy used for
the pin grippers. Some tube sockets I have had to replace in 1990s Chinese made
amps have been just terrible. The Quad-II-Forty amps I have worked on have
abysmally bad sockets. But I have a more recently made Shanling amp here
which produced many clouds of smoke after 3 repair attempts my someone else.
I have yet to establish why small fires occurred on the circuit board, but I can
say the tube sockets are excellent.
NOS Mullards are supposed to have a "polite sound" with less perceivable bass
and treble while NOS Seimans are supposed to give the best all round dynamics.
I find the NOS tubes made here in Australia 40+ years ago are as good as
anything made in Germany.

Under the NEMO :-

As everyone can see there isn't much room under the chassis with Auricaps !
The messy tape with writing on it is just so I know where things are when
I have to service the amp and normally the bottom cover plate completely
covers over such mess.
Everything is hand wired with three dimensional circuitry. The wires from
input sockets to source switch are very fine teflon insulated solid copper
silver plated wires which are fragile but supposed to be good sounding.
Solder has silver content, but just what a small % of silver in solder does
remains un-quantified.

The ""Bunnerong"" power supply :-

This separate remote power supply is about 250mm long x 220m wide x 170mm high.
It used to have a tube rectifier and tube regulator for a line stage only when I sold
it to fellow who ordered it with those features in 2000.
But that fellow sold it to another fellow ( for $2,000 ) in 2004 and he contracted
me to do the Rocket phono amp stage. To enable the PSU to power both Nemo
line amp and Rocket phono amp, I suggested we increase the power output
capability and upgrade the supply by removing tube rectifier and regulator and
use only solid state active devices. The use of the silicon diodes etc allowed the
PSU to function without getting so hot. I found that heat with the PSU tubes was
enough to nearly cook the internally mounted electrolytic capacitors. The plastic
sheathing  of a couple had begun to shrink and tear from prolonged excessive heat.
On hot days when room temperature might be 32C, the underchassis temperature
might reach 70C. 
With tubes gone there was room to place TWO power transformers, one for B+
and a negative -16V rail for the phono amp, and the other for the heating power
needed for a total of 8 signal tubes. There was room for some high value electrolytic
capacitors. The power supply has functioned flawlessly for years, and always runs cool.

I most certainly do not believe tubed power supplies improve the sound.

T1 is a NOS spare part for 1950s Navy electronics, and well potted and sealed.
T2 is a generic 100VA tranny bought from Jaycar electronics parts in Oz and it
runs cool and quietly enough after I potted it.
The large value power supply caps used for the dc heater supply are what I have
acquired from miscellaneous purchases sales here and there but they are very
well over rated for their job.
C2, C3, C4 and R1, R2 form a CRCRC filter for the B+. Further RC filtering is
done within the chassis for the line stage and phono stage which has a solid state
follower regulator, so there is NO NOISE from the rails or heaters in this amplifier
set up.
The dc heater supply does have a shunt regulator to steady the heater voltage
applied to the line stage. This was prudent because the switching on and off of
the phono amp meant that the heater voltage varied too much to the line stage
unless it was regulated. This simple regulator allows the removal of the line stage
umbilical cable and the regulator transistor, 2N3055 will pass all the heater power,
about 12.6V x 0.6 amps, and without getting too hot. Zener diodes alone are not
as effective as the use of the zener and trimmer diodes shown that are working
with the power transistor.
My shunt regulator is very rugged, effective and simple.
 
The heater supply to the phono stage has been made adjustable, see R14 to R18
because there may be a time when someone wants to change a tube type to
something with more or less heater current.

To Rocket Phono Amp

Back to Index Page