An amplifier is an electronic circuit that uses a small input signal to control a larger power source, producing an output signal that is a magnified replica of the input. In a real circuit, an amplifier changes the power level of the signal, actively drawing energy from a DC power supply to boost voltage, current, or both to drive a load. Beginners commonly confuse active amplification with transformer action; a transformer steps up voltage but steps down current proportionally to conserve power, whereas an amplifier injects entirely new power into the signal path from its supply rails.
Amplifier Topologies: Choosing Your Output Stage
When figuring out how to make an amplifier, your first major design decision is selecting the output stage topology. This choice dictates your circuit's efficiency, thermal management requirements, and audio fidelity. The topology determines what fraction of the input waveform cycle the active devices (transistors or MOSFETs) are actually conducting current.
| Topology | Conduction Angle | Max Theoretical Efficiency | Typical Real-World Efficiency | Primary Application | Thermal Profile |
|---|---|---|---|---|---|
| Class A | 360° (Full cycle) | 25% (Resistive) / 50% (Inductive) | 15% - 20% | Audiophile audio, low-noise RF | Extreme heat; requires massive heatsinks |
| Class B | 180° (Half cycle) | 78.5% | 60% - 70% | Historically RF; rarely audio (crossover distortion) | Moderate; devices run cool at idle |
| Class AB | 180° to 360° | 78.5% | 50% - 65% | Standard audio receivers, PA systems | Manageable; requires standard extruded heatsinks |
| Class C | < 180° | > 80% | 70% - 85% | RF transmitters, oscillators | Low; high efficiency but high harmonic distortion |
| Class D | Switching (PWM) | 100% | 85% - 95% | Portable Bluetooth speakers, subwoofers | Very low; devices act as switches, not linear resistors |
The Math: Calculating Gain and Output Power
Let's walk through a worked numeric example for a classic Class AB push-pull audio amplifier driving an 8Ω speaker. We will use a dual-rail power supply of ±15V DC.
1. Maximum Output Power
In a perfect world, the output transistors can swing the voltage all the way to the supply rails. In reality, bipolar junction transistors (BJTs) have a saturation voltage ($V_{CE(sat)}$) of about 1.5V. Therefore, our maximum peak output voltage ($V_p$) is:
$V_p = 15V - 1.5V = 13.5V$
To find the continuous power delivered to the 8Ω load, we must convert the peak voltage to RMS (Root Mean Square):
$V_{rms} = \frac{V_p}{\sqrt{2}} = \frac{13.5}{1.414} \approx 9.55V$
Now, apply the power formula:
$P_{out} = \frac{(V_{rms})^2}{R} = \frac{(9.55)^2}{8} \approx 11.4W$
2. Voltage Gain Requirements
Suppose your audio source (a smartphone DAC or a preamp stage) outputs a maximum of $0.5V_{rms}$. To drive the amplifier to its full 9.55V output, you need a specific voltage gain ($A_v$).
$A_v = \frac{V_{out}}{V_{in}} = \frac{9.55V}{0.5V} = 19.1$
In audio engineering, we usually express this in decibels (dB):
$Gain_{dB} = 20 \times \log_{10}(19.1) \approx 25.6 dB$
When designing the input stage (often an op-amp like the NE5532 or a differential transistor pair), you set the feedback resistors to achieve this exact gain. For a non-inverting op-amp configuration, $A_v = 1 + (R_f / R_i)$. To get a gain of 19.1, if $R_i = 1k\Omega$, then $R_f$ must be $18.1k\Omega$ (use a standard 18kΩ resistor for a gain of 19, which is close enough).
Where You Meet Amplifiers in Practice
Understanding how to make an amplifier extends beyond audio. The core principles of gain, bandwidth, and efficiency apply across multiple engineering domains.
Audio and Consumer Electronics
In home theater receivers, you will find Class AB topologies for the main left/right channels to ensure low Total Harmonic Distortion (THD), while the subwoofer channel almost exclusively uses Class D to deliver 500W+ without melting the chassis. A common failure mode in DIY audio builds is crossover distortion, which occurs in Class B/AB amps when the signal crosses the 0V zero-point. This is fixed by adding a $V_{BE}$ multiplier bias circuit to keep the output transistors slightly turned on at idle (typically 20mA to 50mA of quiescent current).
Instrumentation and Sensor Signal Conditioning
When reading a 10mV signal from a thermocouple or a load cell, you aren't driving a speaker; you are driving an ADC (Analog-to-Digital Converter). Here, power efficiency is irrelevant, and precision is everything. You will use instrumentation amplifiers like the INA114. The critical metric here is Common-Mode Rejection Ratio (CMRR) and input offset voltage, not output wattage. According to Texas Instruments' amplifier design guides, failing to properly decouple the power supply pins of precision op-amps with 100nF ceramic capacitors placed within 2mm of the IC will inject switching noise directly into your amplified signal.
RF Transmitters
In radio frequency applications, such as building a 433MHz transmitter, linearity is often sacrificed for raw efficiency. Class C amplifiers are standard here. Because the transistor conducts for less than half the cycle, the output looks like a series of pulses. A high-Q LC tank circuit (inductor and capacitor) is placed at the output to filter the harmonics and reconstruct the pure sine wave at the fundamental frequency.
FAQ: Common Design Pitfalls
Why is my amplifier output clipping before it reaches the supply voltage?
Clipping occurs when the output signal hits the physical limits of your power supply or the saturation limits of your transistors. If you are running a single 12V supply and expecting 12V peak-to-peak output, you will clip. Furthermore, if your biasing network is misaligned, the quiescent operating point (Q-point) will not be centered on the load line. For a 12V single-supply amp, the output should sit at exactly 6V DC with no audio input. If it sits at 8V, your positive swing will clip prematurely. Check your voltage divider bias resistors and ensure your coupling capacitors are blocking DC from shifting the input stage.
Why does my amplifier oscillate at high frequencies when I connect a long cable?
This is caused by parasitic capacitance and inductance in the feedback loop or output wiring, creating an unintended phase shift that turns negative feedback into positive feedback at RF frequencies. To fix this, add a small 'Zobel network' (typically a 10Ω resistor in series with a 100nF capacitor) across the output terminals to ground. This provides a low-impedance path for high frequencies, stabilizing the amp. Additionally, ensure your feedback resistor values are as low as practically possible (e.g., under 10kΩ) to minimize the impact of stray board capacitance, a principle well-documented in foundational semiconductor literature.
Do I need a heatsink for my output transistors?
Almost always, yes. Let's look at the math for a TIP31 transistor in a TO-220 package without a heatsink. Its junction-to-ambient thermal resistance ($R_{\theta JA}$) is roughly 62.5°C/W. If the transistor dissipates just 4 Watts at idle (common in Class A or heavily biased Class AB), the junction temperature will rise by $4W \times 62.5°C/W = 250°C$ above room temperature. Since the maximum junction temperature is 150°C, the silicon will be destroyed in seconds. You must attach a heatsink with a thermal resistance low enough to keep the junction under 125°C under maximum continuous load.
