the other. Each transistor requires a small
forward bias voltage base to the emitter
to make it start to conduct — roughly
0.7 volts. There is an optimum quiescent
current through the output stage that
minimizes distortion. No setting of this
current can eliminate the crossover
distortion completely, but the remaining
distortion is reduced by the overall global
This current is controlled by the bias
voltage applied to the two output
transistors. The bias control circuit
includes a temperature sensing transistor
that reduces the bias voltage as
temperature increases, attempting to
maintain the bias current at a constant
level. If this is not done, as the output transistors get hot,
the current increases and can reach a condition called
thermal runaway. When that happens, the current
increases until one or both of the output transistors is
destroyed. These output transistors need a sensing
transistor that is closely coupled mechanically. (Refer to
Figure 5 again.)
Note that the output transistors have 470 ohm
resistors in series with the base connection. These are
called stopping resistors. They prevent the output stage
from oscillating. I place these resistors right at the base
lead of the output transistors and use sleeving over them.
Delay connecting the output transistors until preliminary
testing is done.
■ FIGURE 4. Breadboard assembly.
With the output transistors NOT connected, connect
point A to point B (see schematic). This will be the output
point for testing. Before turning on the power, trace your
wiring against the schematic wire for wire. Measure
resistance from the power connections to ground to be
sure there is not a shortcircuit in the wiring.
Turn the volume control down all the way. For a
first test, don’t connect a load of any kind. Turn on the
power and measure DC voltage from output to
ground. It will either be at one of the power supply
voltages or nearly so indicating a problem in the
wiring; or, it will be within 10 or 20 millivolts of
ground indicating proper operation. If it is high, turn
off the power and trace the circuit again. By not
connecting the output transistors yet, you have
probably avoided destroying any of the low level
the power and carefully connect the output transistors. Be
sure to remove the connection of point A to point B.
Measure and set the 1K bias adjustment potentiometer to
maximum resistance. Measure because you might have
gotten the CW and CCW ends of the potentiometer
reversed. Maximum resistance insures a small bias voltage.
Now, turn the power on again and check that the output
voltage is on the order of millivolts.
Distortion is reconciled for minimum by adjusting the
bias voltage potentiometer while measuring the DC
voltage across the two 0.22 ohm resistors in series. Setting
the voltage to 26 millivolts, the theoretical best value
works fine. This adjustment gets my prototype very close
to the minimum distortion point. This is a no signal current
of about 59 mA (0.026/0.44 = 0.059 amps).
If you use a 20 turn potentiometer (recommended), it
will take several turns of the adjusting screw before you
see any voltage across the 0.44 ohms. Don’t turn too
rapidly! Wait a while for the transistors to heat up and
readjust the bias potentiometer for 26 millivolts again. The
heatsink may get warm but shouldn’t get hot with no
signal present. Connect the potentiometer as shown back
■ FIGURE 5. Temperature
sense transistor detail.
When the voltage checks out okay here, turn off
August 2013 33