A 2 Watt audio amplifier made from discrete components.
This was one of the earliest circuits that I ever designed and built, in Spring 1982. At that time I had only an analogue meter and a calculator to work with. Although not perfect, this amplifier does have a wide frequency response, low harmonic distortion about 3%, and is capable of driving an 8 ohm speaker to output levels of around 5 watts with slightly higher distortion. Any power supply in the range 12 to 18 Volts DC may be used.
The amplifier operates in Class AB mode; the single 470R preset resistor, PR1 controls the quiescent current flowing through the BD139/140 complimentary output transistors. Adjustment here, is a trade-off between low distortion and low quiescent current. Typically, under quiescent conditions, current is about 15 mA rising to 150 mA with a 50 mV input signal. The frequency response is shown below and is flat from 20Hz to 100kHz:
The circuit is DC biased so that the emitters of the BD139 and BD140 are at approximately half supply voltage, to allow for a maximum output voltage swing. R9 and R10 provide a degree of temperature stabilization which works as follows. If the output transistors are warm, the emitter currents will increase. This causes a greater voltage drop across R9 and R10 reducing the available bias current. All four transistors are direct coupled which ensures:-
(i) A good low frequency response
(ii) Temperature and bias change stability.
DC Voltages of Prototype
The following voltage checks were made on my prototype. All voltage are made with respect to (wrt) 0 Volt and shown in the table below.
|DC Voltages Measured wrt 0V
The BC109C and 2N3906 operate in common emitter. This alone will provide a very high open loop gain. The output BD139/140 pair operate in emitter follower, allowing the amplifier to drive low impedance speakers.
The signal to noise ration is shown below:
Signal to Noise Ratio:
This amplifier has a S/n ratio of 115dB at 1kHz. Overall gain is provided by the ratio of the 22k and 1k resistor. A heat sink on the BD139/140 pair is recommended but not essential, though the transistors will run "hot" to the touch.
A quick measure of the distortion of this amplifier was performed. Operating on a 15V DC power supply with an input sinusoidal waveform of 100mV peak to peak at 1KHz produced the following results in Tina1.
The number of samples was set to 4096 and Fourier coefficients up to the 16th harmonic were calculated. The sum of the
all harmonics up to 16KHz amounted to just under 2.9% total harmonic distortion, the results are plotted below.
The second and third harmonic are the biggest contribution to overall distortion. Choosing a different amplifier design, a different viasing scheme or more evenly match components can reduce distortion accordingly. At the time this
amplifier was made, I only had an analogue multimeter, so all things considered, it was not too bad an effort.
Picture of my Prototype
Finally an image of the original which has stood the test of time. The BD139,140 power transistors can be seen on the left hand side, the preset near top centre, the BC109C centre right and 2N3906 is buried under a miniature screened audio cable, centre bottom.
Original March 1999
1Tina - Toolkit for Interactive Network Analysis
Last Revision 26-6-06