SETA - Single Ended Tube Amplifier

This is my first vacuum tube amplifier, original design by Andrea Ciuffoli

Here some photos

top view
side view
bottom view
layout view (first test version)
Index
Introdution
Components
Simulation
The main schematic
Reference power supply
Cheap power supply
Sonic results
Layout views
Layout size
My address
Introduction
The aim of this project is to overcome the traditional single ended vacuum tubes amplifier
output stage limitation,adopting a cathode follower configuration.
Such a topology is pretty difficult to implement, due to a severe input voltage swing request
(over the power tubegrid). Such a large voltage can be obtained by the means of different
solutions:
Coupling Transformer, working also as voltage elevator. Requires a low plate impedance and
sufficient power for the driving stage reducing the range of usable tubes.
Inductive driver load, cannot work as an elevator, require a large output capacitor and a very
low series impedance (i.e.: very low Driver Plate Resistance), does not perform very well in the
lower range of the frequency spectrum.
Enhanced Triode mode, big names (Tim de Parravicini) adopted this kind of topology. It has a
good output swing, but the inverted usage of the grids generates a very high output
impedance. such a situation forces the usage of a power tetrode/penthode for driving the
output. Furthermore the stage linearity is pretty questionable.Shunt Regulated Push Pull,
require a supply voltage much higher (double) than the output stage. It means that a second
power supply is required.
Single Ended Hi voltage, same as above, plus the need of a sufficient power triode.In this
design the problem has been solved through the usage of:an inter stage coupling transformer
with a ratio of 1:1.125.a very linear driver, the 6SN7 indirectly heated triode.

Since a low damping factor is very difficult to achieve with a vacuum state amplifier, it is often
kept rather hidden,nevertheless its bad impact over any loudspeaker system, especially for
those with a complex impedance figure orsignificant phase rotation over the frequency
spectrum.
As output tube I am using a power pentode in triode connection instead of a direct heating
power triode because the cathode follower configuration add itself a local feedback that
increase the linearity of the output stage and is preserved the sonic characteristic of driver
stage.
This project use a reference design vacuum tube power supply rectifier with choke input.
The main vantage of choke input rectifier is to skip any current peak generation on rectifier
diodes that are present in normal capacitor input design.
The filaments are in AC and I have choice this type instead the DC because the normal DC
power supply design with large capacitors generate very high current peak about 50A and the
spectrum noise is amplified (the gain in this case 250V/6.3V = 40x so about 32dB) and go on
high voltage secondary given terrible effect on sound.

The main results should match these items:

Components
All the transformers used in this project are Lundahl products:
LL1627/140mA + PCB_B Output (price about 148$)
LL1660/10mA Interstage (price about 69$)
LL1651 Power supply transformer (price about 117$)
LL1638/8H 200mA Power supply inductances (price about 65$)
Lundahl is one of the world´s leading manufacturer of audio transformers used in many
professional studio recording and radio stations so He have a very big experince and all these
products use a custom audio C-core.
I have used the Lundahl transformers also in my new transformer coupled headphone
amplifier and in the DAC getting in both case incredible performances.
The total price of the transformers is 811$ or 616$ for the cheaper version without vacuum
state power supply and with only one inductance.
The output tubes to get very good sonic performances and max power in this project are the
KT90.
The KT90 is the higher power pentode designed for audio and it has been produced by Ei.
Ei tubes are made in the former Yugoslavia to the Philips license, using original Philips
machinery, drawings and test procedures.
These tubes can be find at good price by Korato (Beograd, Yugoslavia) about 18$.
Follows some KT90s Ei items:
photo
pins
datasheet page1
datasheet page2
I have tested also the KT88 Teslovak (considered with high sonic performances), the new KT88
Svetlana and the new 6550WE Sovtek.
The better alternative to the KT90 is the KT88 Svetlana
complete datasheet (in Adobe Acrobat format)
About the 6SN7 I have tested, togheter with my friend Francesco Pintavalle, many types and to
get the best sonic performances the 6SN7 Jan Philips is the first choice.
The tube diode rectifier GZ34 should be the Mullard version. Some DIY don't like the sound of
the GZ34 because consider this to much hard near solid state diodes but this is not true
because depend by the power supply design and with inductive power supply we will have the
best result.
About passive components types, I don't leave many choices: Allen Bradley resistors on anode,
Holco or Caddock resistances on cathode, ELNA Cerafine capacitors on cathode and power
supply.
For all my new projects, I am using the very good ELNA Cerafine capacitors that I love and
which can be find at good price by Welborne Labs.
The ELNA Cerafine capacitors contain super fine ceramic particles which, through chemical
reaction, improve the discharging speed between the anode and electrolyte with very low
distortion.
Any other electrolytic or polypropylene capacitor give very lower sonic performances
introducing some alteration. Only the Blackgate WKZ could give better sound, but I have not
tested it and I know of some problems during startup phase.
About wire forget the teflon ! use only stranded tinned copper wire with pvc isulator and for
loudspeakers I am testing the elevator cables (paralled wire with a good section).

Simulation
The projects below have been simulated with spice release 3f4 running over a UNIX in order to
optimize their performances. The spice3f4 tube models have been computed by a complex
function, with a custom program working on the real plot for the output characteristics (Ip vs
Vp @Vg) for any of the tubes he is interested in.
see here the match (in red tube model)
The distortion spectrum below has been plot by Spice3f4 using a my custom command
"plotfour" add to the original Berkeley's programs.
The Spice3f4 sources are available for donwload at ftp://ic.berkeley.edu/pub/Spice3/ and can
be compilated on many Unix operation system (on PC use Linux).
A nice porting on Windows could be find at S. C. Wong's Home Page.
Here some simulation plot results at the max power:
first stage
seconda stage
grid output stage
cathode output stage
output (see below)

The main schematic

We have a first simple voltage stage with one section of the 6SN7 coupled by a capacitor, the
uinque interstage capacitor of this amplifier, to next stage.The voltage gain of this first stage is
about 17x (22dB) and little chaneg of the anodic resistance should be done to optimize the
sound and gain from 27K to 39K.
The second stage is loaded by .......
The cathode resistance of output stage should be 150ohm (about 46watt 140mA) or 180ohm
(about 36watt 120mA) using the KT90 and 220ohm using the KT88 (about 36watt 120mA).

Reference power supply

The Hybrid Graetz Bridge concept after a Fulvio Chiappetta paper, published by the Italian
Magazine 'Costruire HiFi' a reference for the DIYers.
He tested and comparated this topology versus a pure tube vacuum state power supply design
with a 300B output tube amplifier and there was no difference both while listening and
measuring the output noise spectrum.
I built this power supply and the sonic result is very high even if it uses semiconductors the
sound depends on the rectifier tube.
Infact the vacuum tube rectifier is in series with the solid state diode and the switch
characterictic is of the lower speed device so the tube.
The 0.27 ohm resistors could be adjusted to have correct 5V on filaments of tubes and a value
range from 0.18 ohm to 0.33 ohm should be good to test and match any rectifier.
To reduce a little the cost of this project without compromise the result is possible to use only
one tube rectifier per both channel but the two secondary 11-13 and 12-14of the transformer
must be paralleled to balance flux.
Another little reduction can be obtain using a unique LL1638/8H 250mA inductance instead the
two LL1638/10H 200mA that follow the rectifier bridge

Cheap power supply

This must be considered as the last solution when cost and size are very important or as a
starting point to test this amplifier.Here the same power supply transformer of the reference
version is used but in this configuration the high voltage secondaries are in parallel to get 250V
instead 500V.To reduce more the cost also one bridge and one inductance LL1638/8H 25mA
could be used per both channels, instead two LL1638/10H 200mA, but the separation will be
compromized.

Sonic results
I am currently finishing up the final adjustments for the driver stage bias/load line. The
amplifier has been tested with different output tubes and supply configurations, by two
listeners (Andrea Ciuffoli and Francesco Pintavalle)The overall sonic result is excellent : werll
balanced and very natural.Lower end does not resemble a tube amp so much, it is clear and
tight, deep and defined. Effects of the good damping factor.
It is of interest the sonic influence of the supply tubes, where the usage of a 5R4 RCA gives
impressive results in particular for the firmly focused soundstage.
The KT90 since to be the better choice for output power but the KT88 Teslovak and the new
product KT88 Svetlana used with the 5R4 RCA vacuum diodes (instead the GZ34 Mullard) given
a more hot sound near to direct heating tubes like 300B.
Layout views
click on the images to get real scale

Layout size
Attention: some 100uF+100uF 500V ELNA Cerafine capacitors are higher so you should use
an higher chassie!
click on the images to get real scale
3 mm aluminium with external size 490 x 220 mm

3 mm aluminium with external size 490 x 120 mm