Thanks to 6c45-pe's
performances, the first stage drives without problems the
211's grid. "Soft" A2 Class it's possible. The 6c45-pe
is a low power triode (8 watts max. plate diss.) with high-m,
very high gm and low rp. It's cheap and
readily available. Further, the equivalent noise resistance is 50ohm!
The power supplies are very
simple and effective: no vacuum rectifier in this context.
Alls caps in the H.V. power supplies are
in polypropilene film.
Photo
1 - A Lab Prototype
During the last two years this design concept has evolved and
here I'll present the full development including the simulation decks.
Fig. 1 The Block Diagram
Basically this amplifier is constitued by six sub-units, The
driver circuit (that included the 6c45 circuitry), the output stage circuit (with
211 circuitry) and four power supply unities respectively. Local GNDs are joined
in a common ground bus solidal with an attentively evaluated point
on the copper plane. The copper plane represents also the mechanical support for
irons and power supplies capacitors. The signal paths has been keep very short.
The Driver and the Output Stage, very close beetween their, join a very small
area close to the chassis front-end.
Photo1
Photo 2
Photo 3
Photo 4
Fig. 2 The Amplifier
Many ways exhist to realize a two stage SE amplifier. Base on my
personal taste (derived both sonic and technical consideration), I have matured
the following staircase of execellence:
a) 2 stage SE with Interstage Transformer;
b) 2 stage SE with LC Coupling;
c) 2 stage SE Loftin-White based;
d) 2 Stage SE with RC Coupling.
The solution with the Interstage Iron it's very difficult since
the high cost requested to posses this kind of iron often is not supported by an
adeguate electric response. The design of a good interstage transfomer is true
White Magic Art!! So I have preferred the option b) relying in the convincement
that tube require magnetics as load for the best sonic response. The simplicity
of this amplifier can be admired in Fig. 1. Take in mind that a two stage SE
amplifier having a 211 tube in the output is not an easy task! In fact as
pointed in other section of this web you can obtain the cumulative requirement
of an high amplifiyng factor in union with good current capabilites with the
special category of tube having an high s.
s = gm * m
There are so little triodes having the specification required.
The adopted choice is based on the use of EC8020, WE437A, CV5112/3A167M.
Unfortunately this triode are hard to find and uncheap. The introduction on the
western market of the wonderful Russian triode 6C45-PE
in the late '90 opened new way. I found this first set of 45's in the 1996
thanks to an Hungarian channell and subsequently by the Russian dealer Navigator.
Unfortunately in this latter years due to an increasingly worlwide request the
price of the 6C45 boosted without (fortunately) approaching that of previously
mentioned triodes.
A key concept often negleted in similar design is that the major
efforts are in the power supply side. I mean thechnical and economical efforts!
Fig. 3 The Power Supply (Click on it to enlarge)
In fact any SE amp cannot produce a good sound without an
attentively evalued and designed power supply. As the amplifying section you can
consider a staircase of execellence but there the choice is more simple because
you has facet with:
a) Choke input filter power supply;
b) Capacitor input filter power supply.
Vacuum Diodes with no form of electronic regulation often
produce better results. Taking a look on Fig.3, you observe a DC
regulation for the 211' cathode (with the only IC regulator of the overall
amplifier). The monitor is a simple circuitry able to temporize power supply
events (more details later).
I have evaluated the performance of this block with a
PSpice simulation, Fig.4.
Fig. 4 The schematic used in the power supply simulation
You can observe the absence of any kind of parasitics element in
the schematic of Fig.4. This condition was purposely introduced in order to run
worst-case simulation set.
Fig.5
Steady-state voltage are a bit higher of real values due to the adsence of
parasitcs resistor inside the magnetics. Although the grid voltage is
intentionally kept to the nominal value from the start you can observe "strange
movement arount the first temporal events subsequent to the power on (in this
simulation I have supposed no intervention of the monitor circuitry). A zoom
reveals more attentively what happening, Fig. 6.
Fig. 6
In the time interval 150..250ms the output is subject to a
strong stress due to the cumulative result of positive bumps on the plate
voltage and (mainly) on the grid voltage; the resulting dynamic dissipation is
depicted in the Fig. 7.
Fig. 7 The dynamic power dissipation on 6C45 and 211 plates
Real world results are however mitigated by parasitic resistor
of magnetics .
Fig. 8 Diode pulse currents at start-up
Fig. 9 Repetitive diodes peak current
The contribution of the C-L-C filter is evident observing the
residual hum on the output pin. Afte 5s this component is negligible, Fig. 10.
Fig. 10 Residual hum at the output pin after 4s from start-up
A great deal of attention was reserved also to the simulation of
the amplifier block. For the performances in terms of square-wave response I
have used the schematic in Fig. 11.
Fig. 11 The simulated amlifier block for the square-wave
response evaluation
This schematic is based in a worst case evaluation of main
parasitc elements involved in the amplification process. RLC block take in count
parasitcs of the connections on supply rails, ground bus and devices used for
the amplifier building.Clearly the values are depending strongly by the quality
of wiring.
Fig.12 The 100Hz Sqaure Wave Output
Fig.13 The 1kHz Square Wave Output
Fig.14 The 10kHz Square Wave Output
Solid State designer can be a bit amazed about this poor
electric response but Vacuum Tube base ones say that results can be considered
good particularly if we think that this results refer to the full output power.
However as know the main limiting factor for a SE triode amplifier is the output
transformer.
For the evaluation of other parameter I have particularized
again the schematic of Fig.11 by the introduction of a simple passive
model for the loudspeaker that replace the output resistor, Fig.15 and Fig.16.
Fig.15 Impedance response for the loudspeaker model
Fig. 16 The simulated amlifier block including the loadspeaker
model
The first consequence deriving by loudspeaker model introduction
is a worsening in terms of linear distortion. Respect to resistor load
case, Fig. 17
Fig. 18 Frequency Response with resistor load
Fig. 19 Frequency Response with loadspeaker
Fig. 20 Frequency Response with loadspeaker detail
The Zoom of Fig. 20 shows in depth this harmful mechanism
however kept within 1dB.
Fig. 21 The Output Power Supply Rejection Ratio (Driver Stage
injection)
Fig. 22 The Output Power Supply Rejection Ratio (Output Stage
injection)
Depending by Loadspeaker characteristic we have different
performances in terms of output power and THD. In genreal a rise in the load
impedance produces a lower output power with a lower THD and mutatis mutandis to
low loads correspond more power and higher THD. This behaviour is perfectly
mapped in this amplifier as you can observe, Figg. 23-28.
Fig. 23 The 100Hz Output Sinusoidal wave in union with the
Output Power
Fig. 24 The 100Hz THD
HARMONIC FREQUENCY FOURIER NORMALIZED PHASE NORMALIZED
NO (HZ) COMPONENT COMPONENT (DEG) PHASE (DEG)
After this extensive set of simulation results take a look to
building phases......... sterting from the detail of regulated power supply for
the 211 heater. Remember this is the only solid state element inside this
amplifier!
Photo 5 The 211 Cathode Power Supply Module (Top View)
Photo 6 The 211 Cathode Power Supply Module (Side View)
Photo 7 The 211 Cathode Power Supply Module inside the amp
