More Discrete BJT, Low Power AF Amp Experiments
In March 2009 an audio power amplifier with relatively low distortion and modest power gain was designed, built and tested. All of the amplifier development was performed on my work bench. The goal was to build a speaker-level, audio stage for use in future radio receivers. Popcorn-grade parts were utilized. I am happy with the outcome; a rather strange looking, but simple low power amplifier.
This web page shows the evolution of a simple audio power amp and hopefully further experimentation with this amp will be added over time.
Low Power Popcorn Audio Power Amp
Above in Figure 1 is the ultimate output of over 15 hours of experimentation. A clean sine wave output is realized from "just on" to the full 375 mW output power. Beyond 375 mW nearly symmetrical clipping of the top and bottom of the sine wave occurs. Perfectly symmetrical clipping (and absolutely maximal clean output) can be obtained when R2 is replaced with a 100K or so potentiometer and the output purely balanced using an audio signal generator and oscilloscope. This likely would be overkill, as it works fine as built in Figure 1. As currently shown, the Q1 voltage divider will establish a good stage bias with a variety of DC operating voltages. The output up to the maximum power is a pure sine wave. In the 3 versions I built, the output power varied from 360 to 375 mW. The final version was tweaked and is presented as Figure 1.
Q1 is not set up as a voltage amp, rather it is a pre-driver. Q1 sets the bias voltage of Q2 and also the midpoint voltage of the closed negative feedback loop. There is something like .013 mA emitter current flowing through it at quiescent. A 5 to 10K potentiometer worked the best with this circuit for volume control. Note the positive feedback from the top of the speaker voice coil back to the emitter of Q3. This increases Q2 voltage gain by by bootstrapping up the AC impedance that the Q2 collector sees and likely enhances transistor switching. The parallel transistor Q4/Q5 power follower topology was developed on this page and works very well.
In certain receivers, additional decoupling may be required if the amplifier oscillates (howls). A 10 ohm resistor can be added between the DC positive supply and this circuit. You may also need to increase the first 100 uF filter capacitor in value. Q3 is a DC level shifter and its bias may also be further tweaked by replacing the 2K7 and 1K resistors with a trimmer pot from 1 to 10K ohms.
Shown above is the third breadboard of the Figure 1 amplifier. I built 3 to ensure the design was reproducible. The small orange capacitor is a 2.2uF tantalum type. Any type or value from 2.2 to 22 uF will work fine.
Please refer to Figure 2. This version of the Figure 1 amplifier has just a single Darlington pair for each of the power followers. Note the maximal undistorted sine wave power difference. The quiescent current is very low. If you build this amp, crossover distortion may occur in your design. To remedy this, increase the Q3 bias current. You could replace the 2K2 ohm resistor with a 2K5 to 2K7 ohm resistor, use a trimmer potentiometer or tweak the 2K2 and 1K values to have the lowest quiescent current possible while eliminating crossover distortion. Other than the Q4/Q5 power followers and Q3 bias, all other aspects of this amplifier are identical to Figure 1.
Above in Figure 3 was my prototype of this amplifier topology. This amplifier does not give a perfect sine wave. Nevertheless, it helped me to learn about biasing Q1 and using both positive and negative feedback to get a wide swinging AC signal into the speaker. Q2 works best when driving a very high impedance such as the Darlington pairs used in Figures 1 and 2. This amp later evolved into Figure 2 and then ultimately Figure 1.
Shown above is the Figure 1 breadboard from another angle. This is the highest output 12 volt amplifier I have built to date. Up to the maximum power rating, the sine wave is pure. This amp is relatively quiet considering it uses low cost, junk box transistors.
Please refer to Figure 4. Shown is a simple voltage amp that may be used to drive Figure 1 or 2. From a line level device such as a tape player, no pre-amplifier is required for Figure 1 or 2. In a receiver, some voltage amplification is normally required and the amount of voltage gain required will depend on the receiver RF gain. Transistors are 2N3904 or 2N3906 in Figures 4 and 5.
Shown above in Figure is another experimental, high gain voltage amplifier. This is the basic voltage amp from Figures 1, 2 and 3. Resistor RF is 3K3 ohms which increases the gain by lowering the negative feedback. This amplifier will give a pure sine wave only when driving a high impedance load. The 0.1 uF shunt bypass capacitor can be a lower value, but must be used for stability. This is a great experimenter's stage.
Notice: Some comments may be seen by hovering over the following photographs with your mouse cursor