How to diagnose noise in linear power supplies on the bench

Why noise in linear power supplies matters on the workbench

When designing or repairing high-precision electronic circuits, such as high-resolution analog-to-digital converters or audio amplifiers, the quality of the supplied power is the deciding factor between project success and failure. Linear power supplies are famous for delivering much cleaner direct current voltage than switching supplies because they use passive elements and transistors operating in the linear region to drop and stabilize voltage. However, even a good linear power supply suffers from imperfections inherent to the power grid rectification process, generating small periodic ripples known as ripple voltage. To ensure your circuit operates without erratic behavior, you must master correct bench measurement techniques using an oscilloscope, an instrument that displays electrical signal variations over time.

In practice, ripple is nothing more than the leftover alternating current from the wall outlet that manages to pass through the filtering circuit after being turned into direct current by rectifier diodes. If this leftover is too large, it manifests as hum in audio circuits, bit slips in sensor readings, or instability in sensitive microcontrollers. In power supply voltage drop? How to measure ripple and transients, we explore general power behavior, but here the focus is strictly on high and low-frequency oscillation. Understanding the origin of this noise allows you to separate an actual design defect, such as an exhausted capacitor, from a procedural setup error when touching the probe tip to the circuit board.

Setting up the oscilloscope and proper channel coupling

The first mistake engineers and hobbyists make when measuring power supply noise is leaving the oscilloscope channel in DC coupling, which passes both direct current voltage and variations. Since your power supply might be outputting twelve volts while the ripple is only twenty millivolts, the giant direct current voltage pushes the signal to the top of the screen, forcing you to use a vertical scale so wide that the noise details simply disappear. The solution is to switch the oscilloscope channel coupling to AC mode, which acts as a blocking capacitive filter for direct current, allowing you to view only the fluctuating oscillation around the average value.

Besides coupling, it is vital to adjust the channel bandwidth or limit the bandwidth to twenty megahertz in the instrument's menu. Linear power supplies do not generate extreme radio frequency noise like switching supplies do, so keeping the bandwidth open at three hundred megahertz or more only serves to capture FM radio interference, amateur radio stations, or the office computer network signal. By limiting bandwidth, you eliminate spurious noise that is not part of your power supply's real behavior, resulting in a clean and faithful trace on the oscilloscope screen.

The danger of long ground leads during voltage measurements

Another critical point on the workbench is how you connect the oscilloscope probe to the circuit under test. Almost every oscilloscope comes with a standard probe featuring a long, flexible alligator ground clip that looks harmless but acts exactly like an inductive loop in the room's electromagnetic field. When you clip this ground lead to the power supply ground and touch the probe tip to the output, this long loop picks up 60 Hertz power grid noise and fluorescent light interference, creating a false signal on the screen that does not actually exist in the circuit.

To eliminate this error, replace the alligator clip with a spring ground adapter or use the barrel technique, where you remove the plastic cap of the probe and touch the metal ground ring directly to the board's ground plane while the center tip touches the test point. This rigid and extremely short connection reduces the loop area to almost zero, ensuring that the noise displayed on the screen is purely the ripple generated by the power supply rather than airborne interference. If you notice the noise changing size just by moving the oscilloscope wire, you already know the problem lies in the connection method.

Applying real resistive load for on-demand testing

A linear power supply might look perfect when it has no load connected to its terminals, displaying an almost imperceptible ripple of just a few millivolts. However, the true acid test happens when the power supply starts delivering current to power your real circuit, demanding effort from the regulating transistors and filter capacitors. Therefore, never check power supply behavior in an open circuit alone; use an adjustable bench load or appropriate power resistors to simulate the real current consumption your system will demand during normal operation.

When applying a load, observe how the ripple increases proportionally to the required current, as electrolytic capacitors need to discharge and charge more rapidly between power grid cycles. If the oscillation spikes above acceptable limits for sensitive analog circuits, it usually indicates that the filtering capacitance has lost value over the years due to drying of the internal electrolyte. If you need to simulate sudden consumption variations to check transient response, it is worth checking detailed methods in how to test power drops with UPS and bench load units, adapting the concept for power supply testing on the electronics workbench.

Identifying commutation spikes and diode noise

Even though they are linear, power supplies using standard silicon rectifiers can exhibit high-frequency noise spikes known as diode reverse recovery noise. When the diode transitions from conduction to cutoff, a small momentary current inversion occurs, generating a quick, sharp voltage spike at the output. On the oscilloscope, these spikes appear as thin thorns superimposed on the smooth sinusoidal wave of the main ripple, and they carry enough energy to lock up fast digital circuits or introduce harsh buzzing in sensitive audio circuits.

To clearly view these spikes, you will need to increase the oscilloscope time base speed to the microsecond or nanosecond per division range, allowing you to see the detailed waveform of each transition. If these spikes are excessive, the practical bench solution usually involves soldering small one-hundred nanofarad ceramic decoupling capacitors in parallel with the rectifier diodes or right at the power supply output, creating a low-impedance path to divert these high frequencies to ground before they reach the final circuit. To explore these architectural concepts further in production, see our related technical analysis on How to design a redundant battery bank for UPS.

Final considerations on stability and filtering

Diagnosing noise in linear power supplies requires methodical discipline and rigorous attention to physical connection details on the workbench. The oscilloscope is a powerful tool, but it will faithfully reflect any procedural error, turning environmental interference captures into false equipment failure diagnoses. Ensuring proper coupling, limiting bandwidth, using short ground connections, and applying real load are mandatory steps for anyone seeking reliability in sensitive, long-lasting electronic designs.

With continuous practice of these techniques, you quickly learn to differentiate normal thermal and magnetic ripple from critical problems such as exhausted capacitors or diodes with anomalous behavior. This diagnostic clarity saves precious bench hours, prevents unnecessary replacement of expensive components, and elevates the technical standard of any electronic development or maintenance laboratory.