e015008.PDF
(
748 KB
)
Pobierz
<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01//EN" "http://www.w3.org/TR/html4/strict.dtd">
AU
DIO
PPP HiFi
Valve Power Amplifier
using KT88 valves
Design by G. Haas (Experience Electronics)
Thanks to their pleasant sound, valve-type power amplifiers continue to
enjoy uninterrupted popularity. With such an amplifier, you can eliminate
the impression of coldness, sterility and artificiality that many people expe-
rience with CDs.
The design for a power amplifier that is pre-
sented in this article is based on the PPP
principle. PPP stands for ‘Parallel Push-Pull’.
‘Push-pull’ means that the output stage is
composed of two active elements acting in
phase opposition. One of the valves handles
the positive half-cycles, while the other one
handles the negative half-cycles. In
a Parallel Push-Pull configuration, the
valves in the output stage are con-
nected in parallel with respect to the
ac signal. The disadvantage is that
the power efficiency per valve pair is
less than with classical Class AB
push-pull operation. Otherwise the
PPP principle has only advantages.
This output stage configuration was
invented in the early 1950s, and it
was the configuration of choice in
studios. Reduced distortion, good
sound and a wide frequency
8
Elektor Electronics
5/2001
AU
DIO
response were more important in this appli-
cation area than high power efficiency. With
the triumphal march of semiconductor tech-
nology, PPP was unable to retain any signifi-
cant territory, due to its relatively low power
efficiency and relatively high price. However,
with the aid of modern resources it is possi-
ble to construct excellent HiFi PPP power
amplifiers at an acceptable price.
u2
n
u1
n
1
n
N1
u1
N2
u2
AB versus PPP
R
g2
R
g2
In order to enable you to better understand
the PPP technique, we first have to delve into
a bit of theory. The schematic diagram of
Fig-
ure 1
shows the basic circuit of a classical
Class AB push-pull circuit. Each of the con-
trol grids of the power valves is driven by a
half-wave signal from a paraphase circuit.
This is shown symbolically in the figure; in
practice the two valves would naturally be
driven by positive half-wave signals, since
otherwise the circuit would not work. Here
the left-hand valve is responsible for the pos-
itive half-cycle and the right-hand valve for
the negative half-cycle. The output trans-
former combines the two half-cycles to form
the complete sine wave. The figure shows the
polarities of the voltages that appear in oppo-
site phases across two halves of the primary
transformer winding (N1 and N2). They gen-
erate a voltage on the secondary side of the
transformer that is reduced by the trans-
former ratio
n
. From the figure, you can also
see the response of the power stage to hum
components in the supply voltage. If a hum
voltage is superimposed on the supply volt-
age U
B
, it will be fed into the two windings
N1 and N2 in equal measure. If the circuit on
the primary side of the transformer is com-
pletely symmetric, which means that the
windings, output valves and quiescent cur-
rents are identical, the hum voltages cancel
each other out due to the antiphase feed into
the transformer. Unfortunately, perfect sym-
metry cannot be achieved in practice, so U
B
must be adequately filtered.
Figure 2
shows the ac equivalent circuit of
the basic circuit of Figure 1. The power sup-
ply U
B
represents a short circuit for ac sig-
nals. The ac internal resistance is given by
the series connection of the internal resis-
tances of the two valves and the two wind-
ings N1 and N2 with their equivalent resis-
tances Ra. Both output valves are operated
with a non-zero quiescent current, in order to
avoid cut-off distortion when the sine wave
signal passes through zero.
The situation with a PPP output stage is dif-
ferent, as can be seen in
Figure 3
. You will
notice that there are two supply voltage
sources and that the valves are connected to
U
B
000118 - 11
Figure 1. Basic circuit of a conventional Class AB push-pull output stage.
R
a
R
a
R
i
R
i
U
B
000118 - 12
R
I1
= 2·R
i
+ 2·R
a
= 2·(R
i
+R
a
)
Figure 2. The ac equivalent circuit of Figure 1.
U
B
1
n
u1
u2
U
B
000118 - 13
R
I2
= R
i
/2 + R
a
/2 = 1/2 · (R
i
+ R
a
)
R
I2
= 1/4 · R
I1
(!)
Figure 3. In a PPP output stage, the signal current flows through the entire primary
winding of the transformer.
5/2001
Elektor Electronics
9
AU
DIO
it is connected to the mains with no
secondary load. The PPP power
stage can be seen as a predecessor
of the standard transistor push-pull
output stage. As we have seen, a
whole series of indisputable benefits
are associated with this design, to
wit low internal resistance and tol-
erance of short-circuit and open-cir-
cuit loads, as well as an inherently
good frequency response.
nection, the input resistance drops
to 8.5 k
, but the output stage can
be fully driven with only 0.75 V. Suit-
able wire bridges are provided on
the circuit board. If the amplifier is
wired with a Cinch socket, the input
sensitivity remains at 1.5 V.
At the input to the amplifier, the 2.2-
µF bipolar electrolytic capacitor C1
ensures low-impedance coupling
and at the same time blocks any dc
component of the signal. Resistors
R3 and R4 block coupled-in RF inter-
ference. Valve V1a provides the main
amplification, with RF oscillations
being suppressed by C2. Valve V1b
is wired as an impedance converter,
and it serves as a low-impedance
source for the signal applied to para-
phase circuit valve V2. The grid of
valve V2a is driven directly without
a coupling capacitor, since the dc
potential has been brought to the
proper level by suitable selection of
component values. The control grid
of V2b is connected to the same dc
potential as the grid of V2a via R14.
Capacitor C10 provides a short-cir-
cuit path to ground for any ac com-
ponents that may be present at the
grid of valve V2b, which is driven by
the ac signal voltage via its cathode.
Valve V2 must provide the full ampli-
tude of the output voltage, since the
output valves V3 and V4 only pro-
vide the signal with the current
needed to achieve the desired
power. The latter two valves act as
cathode followers, in the same way
as transistors are used as emitter fol-
lowers in the output stage of a tran-
sistor amplifier.
This also gives an answer to the fre-
quently hotly debated question,
‘How do different types of valves
sound?’ Since the output valves of a
PPP design do not contribute any-
U
B
R
L
U
B
A power amplifier in
three building blocks
Figure 5
shows the complete
schematic diagram of a monophonic
power amplifier. In consists of three
blocks, and is consequently divided
into three circuit boards. This allows
the amplifier to be built such that the
valves show to best advantage, as
can be seen from the photograph.
The transformers and the large elec-
trolytic capacitors, in contrast to
many other amplifier designs, are
‘embedded’ in the equipment.
Block 1 is the amplifier circuit, block
2 is the power supply and block 3 is
the switch-on current limiter. At the
input to the amplifier, provision is
made for an E-1220 audio trans-
former. If you want to have a floating
symmetric input, you can use the
transformer together with an XLR
socket; otherwise, a Cinch socket is
adequate. Thanks to the transformer,
it is possible to use floating sym-
metric cabling, as is common prac-
tice in studios. This avoids earth
loops and quality losses due to the
cabling. Furthermore, the trans-
former can be connected either 1:1 or
1:2. In the 1:1 configuration, the
input resistance is around 34 kΩ ,
with an input sensitivity of 1.5 V for
full output power. With the 1:2 con-
000118 - 14
Figure 4. The ac equivalent circuit of Figure 3.
the transformer in a completely different man-
ner. The ac signal current flows through the
complete primary winding of the transformer,
which is similar to the situation with a mains
transformer. Here the valves are connected in
parallel for ac signals, which reduces the
internal resistance of the PPP circuit by a fac-
tor of four. This has various benefits. First of
all, the transformer may have a smaller trans-
former ratio. This reduces the effects of stray
inductance and winding capacitance, which
yields an improved frequency response.
Besides this, a genuine ac current flows
through the complete primary winding,
instead of one half-cycle per half winding as in
the classic Class AB push-pull circuit. This
avoids the much-feared ‘flyback’ effect that
occurs with Class AB push-pull output
stages, in which a half-wave voltage across
one half of the primary winding generates a
voltage that is twice as large on the plate of
the opposite valve, due to the 2:1 ratio of the
autotransformer formed by the primary wind-
ing. If the transformer core is magnetised and
the stored energy is not fully consumed on
the secondary side, the voltage on the pri-
mary side rises sharply, and to make matters
worse it is again multiplied by the same 2:1
factor. The end result is a voltage breakdown
at the weakest point — valve, socket or trans-
former. This may be a desirable effect in an
automobile ignition coil, but in an amplifier it
is devastating. In the event of a short circuit,
a Class AB push-pull stage will become hot,
but it will not be destroyed. Such a valve out-
put stage is thus by nature short-circuit proof,
but not open-circuit proof.
A PPP output stage, by contrast, is safe under
both extreme operating conditions. Since the
transformer is connected as previously
described, the output stage does suffer from
‘flyback’. It can be compared to a mains
transformer, which also suffers no ill effects if
10
Elektor Electronics
5/2001
AU
DIO
C6
3p3
R10
220k
A1
V1 = ECC83
V2 = ECC81
R5
2k7
R11
47k
G1
R4
R15
R27
C4
C5
V3
D1
56V
3
47µ
450V
47µ
450V
C8
220n
4
D2
1
00V
R23
10k
5
V2a
8
630V
KT88
R19
D3
1
00V
R25
R17
22k
V1b
*
see text
D4
1
00V
C7
220n
M1
*
voir texte
P1
D1
C11
C13
25k
630V
R21
C2
1N4007
47µ
100V
47µ
100V
Tr2
8
Ω
V1a
50V
Tr1
E1220
10p
R7
R8
4
Ω
C1
R13
a
R3
2k2
*
M
2
Ω
2µ2
0
R22
R1
R2
R6
R9
50V
R14
b
C3
1N4007
C12
C14
AP-234
P2
47µ
100V
47µ
100V
100p
D6
25k
M
R26
M2
R18
22k
R20
C10
8
D1
1N4007
C9
R24
10k
KT88
C1
1000µ
V2b
5
R1
100
Ω
22µ
100V
4
220n
630V
40V
R16
R28
D2
Re1
C3
C2
3
V4
R12
47k
G2
1N4007
1000µ
40V
1000µ
40V
A2
R29
V1
V2
V3
V4
R2
100
Ω
2
2
R3
4
9
5
4
9
5
6V3
7
7
100
Ω
Tr3
R4
100
Ω
R5
R30
100
Ω
R6
100
Ω
R7
100
Ω
Si1
R2
1k
G1
R8
100
Ω
C1
200mA T
100n
1000V
R3
C2
C3
R1
GL1
S1
470µ
450V
100µ
450V
Si3
M1
B500C1500
2A T
A1
50V
230V
A2
50V
M2
Si4
B500C1500
2A T
R4
R6
C5
C6
GL2
470µ
450V
100µ
450V
C4
Si2
R5
1k
G2
100n
1000V
200mA T
000118 - 15
Figure 5. A circuit according to the building-block principle: the equipment can be configured as a monobloc or stereo amplifier, with
different types of output valves.
5/2001
Elektor Electronics
11
AU
DIO
COMPONENTS LIST
R13 = 2k
7, MO, 2W
C15 = 1nF, MKH, lead pitch 7.5mm (see
text)
R14 = 33k
Ω
R15,R16 = 10k
, MO, 2W
Power Output Stage
with KT 88 or 6550A
Semiconductors:
D1 = zener diode 56V 1.3W
D2-D4 = zener diode 110V 1.3 W
D5,D6 = 1N4007
R17,R18 = 10k
Ω
R19,R20 = 220k
Resistors:
R21,R22 = 33k
Ω
R23,R24 = 3k
3
R25,R26 = 10
Ω
, MO, 2W
R27,R28 = 270
(unless otherwise stated, use metal film
types, 0.7W, 1% tolerance;
MO = Metal oxide, 5% tolerance)
Miscellaneous:
V1 (Rö1) = ECC83
V2 (Rö2) = ECC81
V4 (Rö4),V5 (Rö5) = KT88 or 6550 A
Tr1 (Ü1) = E-1220 (if necessary)
Tr2 (Ü2) = AP-234
2 off ‘Noval’ (9-way) valve socket, PCB
mount
2 off ‘Octal’ valve socket, PCB mount
Tr1 = Mains transformer NTR-P\7 (Mono)
or NTR-P\5 (Stereo)
, MO, 2W
R29,R30 = 47
Ω
P1,P2 = preset 25k
R1,R2 = 68k
Capacitors:
C1 = 2
R3 = 2k
Ω
2
R4 = 150k
F2 50V bipolar
C2 = 10pFceramic
C3 = 100pF ceramic
C4,C5 = 47
µ
F 450V axial
C6 = 10-33pF (fit only if parasitic oscilla-
tion is noticed)
C7,C8,C9 = 0
, MO, 2W
R5 = 2k
Ω
7, MO 2W
R6 = 2k
2
R7 = 1M
Ω
R8 = 2k
7
R9 = 22k
Ω
, MO, 2W
R10 = 390k
F22 630V, MKS 4
C10 = 10
µ
F 100V, lead pitch 5mm
C11-C14 = 47
Ω
R11,R12 = 47k
Ω
, MO, 2W
F 100V, lead pitch 5mm
12
Elektor Electronics
5/2001
Plik z chomika:
gasma
Inne pliki z tego folderu:
bge.jpg
(26 KB)
detail1.htm
(5 KB)
detail2.htm
(4 KB)
detail3.htm
(4 KB)
detail4.htm
(6 KB)
Inne foldery tego chomika:
1974
1975
1976
1977
1978
Zgłoś jeśli
naruszono regulamin