Updated 2012.
Content :-
Picture of 5050 stereo amp, general description and specification.
Schematics for 1 amp channel, power supply and protection. Full explanations.

5050 integrated amp
          on bench.
Integrated 5050.
This 5050 was custom was custom built in 2000 for a client in Victoria.

The chassis has a 55mm high brass channel brazed at each corner.
There are two top plates, aluminium for tubes and the steel bottom of box for
PT and two C-cored 5Kg OPTs.
Bottom plate is perforated steel to allow ventilation, and box over transformers
is painted aluminium.
Not shown is the perforated steel cover screw fixed over the tubes.

There is an inbuilt preamp, source select switch for 5 inputs, and Alps Black dual
gang volume control. There are two 50Watt Ultralinear output stages.

Power maximum at onset of clipping, resistance load, 1kHz
4r0, 53Watts AB1 with first 10Watts class A.
8r0, 36Watts AB1 with first 25Watts class A.
Weight = about 30 Kg.
4 Output tubes = any brand of KT88, 6550, KT90, KT120. Sovtek KT88 in picture.
4 input tubes = 4 x 6CG7 in picture. ( Later versions of this amp used 6 x 6CG7 )
Bandwidth at 40Watts into 6 ohms, 14 Hz to 68 kHz.
Distortion at 40Watts, less than 0.2%,
Distortion at 3Watts, less than 0.05%,
Output impedance less than 0.3 Ohms.
17 Db of global NFB.
Stability is unconditional, and the amp can be used with any type of speaker load.
Fixed bias is used with full active protection and bias balance led indicators are
mounted for easy viewing on the front panel. When both are extinguished, bias
current balance in each channel  is set correctly. No external preamp is required.
There is an octal socket at rear to supply B+ and heater power to a tubed phono

Schematic of  50 watt UL AB1 power amp....

The pictured amp from 2000 has two triodes of a single 6CG7 for V3&4.

In 2006 I thought it better to use two paralleled 6CG7 for V3 and V4, shown here.
The change from original circuit allowed other slight improvements shown above.
Bias resistors R17, R18 have been reduced from 220k to 68k to give better dc
stability since less Vdc is generated across the 68k as tube age.  The 68k
are low value cap coupled loads for V3+4. But the parallel triodes cope
very well with RL shown. They also improve the dynamics of the sound and
give reduced THD.
How it works :-
The input signal is selected by the rotary wafer input switch and is attenuated by
about 4 dB by the C1, R2, R3, and balance control network.
V1 is 1/2 a 6CG7 working as a simple SET preamp stage, gain = about  x10,
and drives 50k log Alps Black volume control pot.
Because no long cables are used between the preamp and power amp the bandwidth
remains wide despite the simplicity.

V2 is the first triode of the power amp and is a SET stage acting as a differential
amplifier to accept the signal input at its grid and the feedback signal at its cathode,
so the difference between the input and feedback signals form the Vg-k between
grid and cathode. V2 anode Va is fed to the network formed by C3, C6, R12, and R13.
This network reduces gain and phase shift at below 20Hz to allow excellent LF
stability. C7+R11 are a Zobel network to reduce gain and ultimate phase shift above
10kHz. It is a "gain shelving" network. The LF and HF networks ensure that with GNFB
the function is always unconditionally stable. The networks are also known as critical
damping components and R&C are chosen to best suit the L&C characteristics
of the OPT.
Additional stability networks were used, R32+C11 at output, and C12 across R33.

Warning! Others building this kind of amp won't have OPTs equal to the OPT I made.
They may choose slightly different tube types. Therefore this schematic is a guide
only for general method. Using wrong R+C in stability networks can cause
oscillations at LF or HF or both.

V2 Va is applied through the networks to the one active grid input of V3.
V3 and V4 form a differential amp with commoned cathodes and CCS to a negative
B- rail. V4 has a grounded grid, and both grids are biased at 0V.
This differential amp is also called a "long tailed pair", or LTP. V3+V4 have exactly
equal ohm loads for anodes. The cathode CCS, ie, transistor constant current sink
for an unchanging Idc flow, and there is no signal current change between cathode
and -87Vdc negative rail derived from supply for fixed grid bias.

The CCS uses a MJE340 which has Vdc between base at -24V and -87V rail kept
constant. There is a constant Vdc of 62Vdc across R16 5.17k (approx) giving Idc
= 12mA.  The collector input resistance is many megohms. The bjt has has no
active signal controlling effect, and no sonic signature. I could have used a
pentode tube for this CCS but there is no need since the tube would have no other
function than providing a high ac impedance at the common cathodes.

V3+V4 Grid bias is held at 0Vdc via R13 at V3, V4 grid is connected directly to 0V.
Incoming signals to V3 cause Ia current change which is opposite phase in V4,
because the sum of the positive and negative moving Iac must always remain
constant and equal to the constant current flow into the collector of the MJE340.
RLa for V3 and V4 anodes is 33k // 78k = 23k approx. If max Va = 42Vrms, then
max peak Ia swing = +/- 2.6mA, less than 1/2 the idle current of 6mA, and higher
Va swing is possible without V3+V4 overloading.
Because Iac and anode RLa of each of in V3+V4 are equal, the two Va are equal
amplitude and balance is within +/- 1%. There is not much need to use perfectly
matched triodes for the two triodes in this LTP with a common cathode CCS.
Vac balance depends solely on the equality of the resistance loads on each 1/2
of the LTP, and modern metal film resistors allow excellent balance and stability.
The LTP here has input to V3 grid with V4 grid at 0Vac. There is Vac at the common
cathodes = 1/2 x V3 Vg, so that Vg-k for both triodes is nearly the same. 

The balanced LTP Va are applied to the V5+V6 grids. These are biased through
R17, R18 , 68k, which are fed a negative voltage from the adjustable network R19,
R20, R22 plus a balancing potentiometer to adjust the balance of dc current in
each output tube.

The biasing network has a fixed -87Vdc bias supply. This -Vdc supply is well filtered
but not regulated. If mains Vac rises, then so will B+, and we want the same %
variation for the -87Vdc, so that if B+ increase increases Iadc, then grids become
more negatively biased which helps prevent idle Pda for tubes rising too much.

Bias adjustment for the owner is to equalize the Idc of each output tube.
VR1 is a 10k linear wire wound pot to balance Ikdc. The 3W rated 24mm dia pots
bolted into top plate with slotted shafts pointing up which allow easy turning with
screw driver pointing down through holes in well perforated steel tube cover which
is not shown in the picture.

With VR1 set in the center position, you can expect V5+V6 grid bias voltage
= -60Vdc. But the Eg-k needed by each output tube to give equal Idc is always
slightly unequal. The rotation of VR1 in either direction allows easy balancing
to get Idc of each tube within +/- 5%. The idle Idc of each tube is monitored by the
two LEDs at the front of the amplifier which normally stay extinguished.
If one LED starts glowing, it means Idc in adjacent tube has risen a little above the
Idc in other tube. If the wanted Idc in each tube is 44mAdc, you may find 48mA in
one tube and 40mA in the other so a net difference = 8mA, which can cause
unwanted extra THD and a tube which is hotter than the other. A quick turn of VR1
while watching LEDS can reduce the Idc difference to less than +/- 2mA.
It is possible that 2 output tubes will be found to idle with 65mA and 55mA if Eg
is the same -60Vdc in each. When balanced, expect to get 60mAdc in each,
and that means Pda is above the wanted maximum of about 23Watts.
Pda at idle = Ia dc x Ea dc = 0.06A x 500V = 30Watts.  I have rarely ever found
any healthy KT88 give Idc more than +/- 10% of a center value.

No special tools, voltmeters or technical expertise is needed for adjusting
the bias of this amp. As tube age, the Eg-k for a wanted Idc varies more widely,
so a larger turn of VR1 is needed, and it is possible to find VR1 turns to its max
position one way, and a red LED is still alight. This tells an owner an aging tube is
unable to be biased to function within a safe working zone, and the tube should be
So the LEDs indicate a fault with output tube, or if he has a shorted speaker cable
or has some other fault causing imbalance of Idc.
Thus an owner is made aware of a problem well before any serious damage
occurs. But he should know he cannot ignore a red glowing LED. And if the Idc of
any one output tube becomes about twice the Idle Idc for longer than 4 seconds,
the amp automatically turns itself off. 

The output stage is "Ultralinear". The OPT has 50% screen taps.
This set up has been used by many brands since about 1950 to linearize the
transfer function of power beam tetrodes and pentodes. With UL, THD spectra
becomes more like a triode with far less odd number H. The effective dynamic
anode resistance, Ra, is reduced from a typical 32k for a KT88 without UL taps
to about 5k. At less than 10Watts, anode load is approximately same ohm value
to damping factor = 1. The GNFB much reduces effective Rout and increases
DF to at least about 8.

The power supply for 5050.

Silicon rectifiers are used to give the best efficiency, and regulation and trouble free
operation. C-L-C filtering of the +500Vdc anode supply is used to keep hum noise to
extremely low levels. 12V regulators 7812 are used with an 1N4004 in the adjust leg
to get close to 12.6V for dc heater supplies. The power supply uses a large
conservatively rated power transformer with a low temperature rise, and excellent self

The active protection schematic and bias balance indicator amps :-

The above schematic is a digitally tidied up copy of the messy hand-drawn
sketch which I had in 2001.
There is an effective bias balance indicator, 3 second delayed turn on to limit
inrush current at turn on, and active protection in case of excessive Ik in one
or more output tubes. This schematic should be read in conjunction with
5050 amp and PSU schematics. The operation is as follows :-

The mains switch for the amp has two poles, one for the 240V Mains Active
line in, and the other to operate as a commutator switch in series with
+16Vdc auxiliary PSU rail and SCR C106D anode.

At turn on, the mains switch connects Active to large PT primary and to small
7VA PT which powers the protection and relay 1 and 2. The mains Neutral line
connects directly to primary of small PT.
The 7VA PT is only controlled by the manual on-off mains switch.

The mains Neutral line to large PT primary is connected through contacts of
relay 1 and relay 2.
Relay 1 is open at turn on so that mains input current flows in 500r x 20W.
The resistance limits the peak currents so the rise in B+ is slowed and tubes
are heated less quickly. When B+ reaches about 2/3 its full peak value at
about 3 seconds after turn on relay 1 closes to shunt 500r, and B+ will then
rise to its peak value. The "inrush current" surges at turn on and at 3 seconds
later are about 1/2 the amps which would otherwise occur without the delay
circuit. Therefore a lower and more usefully sensitive mains fuse can be used.

If too much Idc flows in one or more output tubes, relay 2 opens the Neutral
line to large PT thus turning off all power for the amp. The 7VA PT will remain
on, and red LEDS will light up to tell an owner the amp has a problem and
has turned off internally. Without this active protection, nobody could rely on
a mains fuse to blow is one output tube goes into "thermal runaway mode"
where it overheats, glows orange, and conducts maybe 600mA for a minute
before glass melts and it fuses internally. Overheating of OPT primary or
HT windings is avoided.

After turn on, Idc in each output tube will reach the wanted idle level between
about 44mA and 55mA after 20 seconds. This Idc generates between +0.44
and 0.55Vdc at tops of 10 ohm current sensing R23+R24k in both channels.
The four Ek Vdc connect to protection board at K1, K2, K3, K4. Any Vac or
noise voltages at each 10r0 is filtered low by 1k5 and 100uF.
Each of the four Ek voltages is applied to a base of a BC337npn bjt. There are
4 x BC337 in two differential amps, one amp for each channel. The pairs of
collectors in each BC337 diff amp feed bases of a direct coupled diff amp
with BC327 pnp bjts. The BC327 control a pair of red LEDs. These LED can
only be turned on if there is a significant difference on the pair of Ek voltages from
each pair of OP tubes in each channel. The circuit is arranged to turn on an LED if
one tube of a pair conducts about 4mAdc more than the other. The LED are arranged
so they indicate which tube has more Ik than the other, and such imbalance is noticed
by an owner from across the room. He may then operate the small balance pot for
each channel and rotate one way or other to alter the grid bias voltage so that
BOTH LEDs remain unlit. My experience is that bias balance adjustments are
rarely ever needed. The LEDs will turn on and flicker during warm up, and then
again at turn off when momentary Idc imbalances occur. If there is a shorted speaker
cable, the amp will have unbalanced Idc, and LEDs will flicker, warning an owner
that something is wrong.

If excessive volume is used with a speaker with Z < 3 ohms, or if teenagers try to
remove roof tiles or break window glass with audio levels, or if for any other reason
one or more output tubes conducts excessive Idc for longer than a few seconds,
the signal at the SCR gate may reach 0.65Vdc and the SCR will turn on, and
remain turned on. The SCR turns on relay 2 and contacts open and the large
mains transformer will be turned off, and remain turned off indefinitely until someone
resets the amp by turning off the mains switch, waiting 2 seconds, then turning back
on. If the problem causing high Idc remains, the amp will again turn itself off, and
an owner will know he needs technical help.

My protection circuits have saved owners from paying high repair bills. 

People are mad if they build a wonderful 5050 amp without a board
with a circuit like this to manage the behavior of the amp, which needs
to learn it shouldn't upset anyone

Back to power amp directory

Back to Index Page.