Power supply ripple remains one of the most stubborn obstacles in the quest for pure sound reproduction. Every audio amplifier draws from direct current rails that carry faint traces of alternating current leftover from rectification and filtering. These variations appear as small voltage wiggles at 100 or 120 hertz and their multiples. Even well designed capacitor banks fail to erase them completely. The ripple travels through the circuit and mixes with the music signal. Listeners notice the result as a subtle haze that veils low level details or adds a low level hum during quiet passages.
Designers have explored many fixes over decades. Passive filtering helps but falls short at low frequencies and under load. Active regulation adds complexity and heat. A simpler yet more effective path lies in controlling current flow itself. Transistor based constant current sources replace ordinary resistors that once set bias voltages and operating points. The change dramatically strengthens the circuit ability to ignore supply disturbances. Power supply ripple rejection improves by tens of decibels. Backgrounds grow quieter. Musical textures emerge with newfound clarity. The upgrade feels subtle at first yet proves profound upon extended listening. Circuits behave closer to ideal. They deliver the recording with less electronic overlay.
Many enthusiasts discover this transformation when comparing prototypes. One version uses resistors throughout. Another swaps critical positions for active current control. The difference reveals itself in effortless dynamics and deeper silence. The technique applies across solid state and even some tube designs though the focus here stays on transistor implementations common in modern amplifiers.
Power Supply Ripple Poses a Constant Threat to Audio Fidelity
Every mains powered audio system contends with imperfect direct current. Bridge rectifiers convert alternating current to pulsating direct current. Large electrolytic capacitors smooth the pulses yet residual ripple persists especially when the amplifier draws varying current. Class B output stages exacerbate the issue because their current draw follows the signal waveform. This creates signal related modulation on the rails. The disturbances couple into sensitive early stages where they get amplified along with the music.
Ripple acts like turbulence in an otherwise steady stream. It modulates transistor parameters such as collector current and base emitter voltage. In a resistor biased stage these changes pass directly to the output. The ear detects the artifact as loss of focus or artificial warmth that masks true timbre. High resolution systems expose the problem more readily. A quiet recording of solo acoustic guitar suddenly reveals faint buzzing beneath the strings. Orchestral crescendos lose their effortless rise because the supply wobble adds intermodulation.
Engineers quantify the intrusion through careful measurements. They inject artificial ripple onto the rails and observe the output. Without proper isolation even one volt of ripple can produce millivolts at the speaker terminals. At typical listening levels this becomes audible. The goal therefore centers on breaking the path from supply to signal. Constant current sources achieve this by presenting extremely high impedance to alternating variations while holding direct current steady. Resistors cannot match this selective behavior. Their fixed resistance passes both direct current and alternating disturbances equally.
Defining and Measuring Power Supply Rejection Ratio Accurately
Power supply rejection ratio known as PSRR measures how effectively a circuit attenuates variations on its supply lines. Engineers express the figure in decibels. The standard formula reads PSRR in decibels equals 20 times log base ten of delta V supply divided by delta V output. A value of 60 decibels for instance means supply changes appear at the output reduced by a factor of one thousand. Higher numbers indicate stronger isolation.
In practice testers apply a known alternating current signal to the positive or negative rail while the amplifier operates under normal conditions. They monitor the output for any correlated component. The ratio between injected amplitude and observed amplitude yields the rejection figure. Frequency matters greatly. Rejection often remains high at direct current but rolls off as frequency rises. Audio demands excellent performance from 50 hertz upward through the midrange where human hearing proves most sensitive.
Consider a typical differential input pair. With a simple resistor in the tail position PSRR might reach only 80 or 90 decibels at low frequencies. The resistor allows supply induced current changes to unbalance the pair. Replacing it with a constant current source raises the effective tail impedance from tens of kiloohms to several megaohms. The common mode gain drops accordingly. Overall PSRR climbs by 40 decibels or more. The formula for the improvement approximates 20 log base ten of Z current source divided by R resistor. When Z reaches 4.6 megaohms and R equals 22 kiloohms the theoretical gain nears 46 decibels. Real world results closely match this prediction.
Such numbers translate directly to listening. A 40 decibel improvement turns a noticeable hum into something 100 times quieter. It vanishes below the noise floor. Subtle cues in the recording such as hall ambience or finger noise on strings become audible without competition from the electronics.
Traditional Resistor Approaches Limit Rejection Capabilities
Resistors have served reliably for decades as current setters. They establish quiescent points in differential pairs voltage amplification stages and bias networks. Their simplicity appeals to builders. A single component replaces complex active circuitry. Yet this simplicity carries a hidden cost.
The impedance of a resistor equals its ohmic value at all frequencies. Supply ripple therefore creates proportional current ripple. In a tail resistor of 22 kiloohms a 100 millivolt supply wiggle produces about 4.5 microamperes of current variation. This modulation unbalances the input transistors. The error appears at the output multiplied by the stage gain. Higher value resistors reduce the relative variation but they also demand higher supply voltages to maintain the same operating current. Practical limits appear quickly.
Voltage setting resistors in bias chains suffer similarly. Ripple modulates the voltage drop across them. The reference point for later transistors wobbles. Base emitter voltages shift. Collector currents vary. Each stage adds its own contribution until the cumulative effect becomes audible. Builders who rely solely on resistors often compensate with extra decoupling capacitors. These help at higher frequencies but prove ineffective against slow mains fluctuations or load induced sags. The circuit remains vulnerable.
Comparisons reveal the gap clearly. A resistor based prototype might show 70 decibels PSRR at 120 hertz. The same topology with active current control reaches 110 decibels or better. The difference manifests as blacker backgrounds and more stable imaging. Instruments stay anchored in space even during complex passages. Resistors simply cannot isolate the signal path from the supply with the same effectiveness.
Transistor Constant Current Sources Offer Robust Alternatives
A basic transistor constant current source uses a bipolar junction device with a stable base voltage and emitter resistor. The current through the collector approximates open parenthesis V reference minus V base emitter close parenthesis divided by R emitter. V base emitter stays near 0.65 volts for silicon transistors. Designers derive the reference from zener diodes light emitting diodes or another transistor network. The output current remains nearly independent of collector voltage over a wide range.
For greater precision engineers employ current mirrors. Two matched transistors share a base connection. One sets the reference current. The second copies it. Output impedance rises dramatically because the mirror action opposes changes in load voltage. Adding emitter degeneration resistors further stabilizes the value against temperature and beta variations.
More sophisticated topologies like the Wilson mirror stack three transistors. They cancel base current errors and multiply output resistance. Typical values exceed one megaohm. Cascoding adds another transistor in series. It holds collector emitter voltage constant on the mirror pair. The result approaches ideal current source behavior. Ripple on the supply now produces almost no change in delivered current. The stage above operates in splendid isolation.
These circuits require only a handful of parts yet deliver performance impossible with passive components. Builders select transistors with high Early voltage to maximize output resistance. Careful layout minimizes stray capacitance that could bypass the high impedance at high frequencies. The effort pays dividends in measurable and audible ways.
Specific Mechanisms Driving the Substantial PSRR Gains
The core advantage stems from impedance disparity. A resistor offers fixed resistance. A constant current source offers dynamic resistance that can reach millions of ohms. Supply ripple encounters this barrier and generates tiny current. In a differential pair tail the small current change splits equally between the two legs. It appears as common mode signal and cancels at the differential output. The rejection multiplies.
In a resistor tail the same ripple creates larger unbalanced swings. The cancellation remains incomplete. Measurements confirm the effect. Replacing a 22 kiloohm tail resistor with a current source of 4.6 megaohm impedance yields roughly 46 decibels better PSRR. Additional stages compound the benefit. A voltage amplification stage loaded with a current source instead of a collector resistor sees similar isolation. Its output becomes far less sensitive to positive rail wobbles.
Another mechanism involves reduced modulation of internal voltages. In a resistor biased current source ripple affects the base emitter junction directly. Current varies. With an active reference such as a current regulating diode or secondary source the base drive stays constant. V base emitter modulation drops sharply. One common upgrade replaces the single bias resistor with such a device and records 30 decibels better local PSRR. The entire amplifier inherits the improvement.
These gains hold across frequencies of interest. Low frequency performance benefits most because passive filters struggle there. The circuit maintains stability even when mains voltage sags or surges. Designers achieve consistent sound regardless of wall outlet quality.
Widespread Use in Modern Audio Amplifier Topologies
Modern solid state amplifiers place constant current sources in several strategic locations. The input differential pair almost always employs one in the tail. This single change transforms common mode and power supply rejection. Active loads in the collectors raise differential gain while preserving balance. The voltage amplification stage commonly uses a current source as collector load. It provides high gain and excellent isolation from the positive rail.
Some designs extend the concept to output stage biasing though thermal tracking requires care. Preamplifiers benefit equally. Phono stages with their high gain prove especially sensitive to supply noise. Current sources keep the delicate signal clean from the first transistor onward.
A representative topology starts with a long tailed pair input. The tail current source sets 2 milliamperes total. Each leg runs at 1 milliampere. The collectors drive a current mirror load. The single ended output feeds a voltage amplification transistor loaded by another current source. This arrangement delivers open loop gain exceeding 100 decibels with PSRR better than 100 decibels from 20 hertz to 1 kilohertz. Closing the global feedback loop suppresses residual artifacts even further.
Builders report consistent results across power levels from milliwatts to hundreds of watts. The technique scales gracefully. Smaller signal circuits use low power transistors. Higher current stages employ larger devices or paralleled units. The principle remains identical. Replace the voltage setting resistor with an active source and watch rejection soar.
- Stable operating points that resist supply and temperature shifts
- Reduced susceptibility to load induced rail modulation
- Higher effective gain in early stages without added noise
- Cleaner separation of musical elements in complex mixes
- More consistent performance across different power line conditions
These advantages accumulate into a noticeably more engaging presentation.
The evolution toward constant current sources reflects a deeper understanding of circuit behavior. Engineers once accepted resistor limitations as inevitable. Today active control stands as standard practice in serious designs. Listeners reap the rewards every time they play a favorite recording. The music flows more freely. Details once hidden emerge naturally. The amplifier steps aside and lets the performance shine. This quiet revolution continues to raise the bar for audio fidelity one stage at a time. The results speak for themselves through years of satisfying listening.
Our Telegram Channel
Our YouTube Channel