How to test ATX power supplies with a dummy load

Why an ATX power supply refuses to turn on alone

When removing a computer power supply from its chassis for isolated testing, the first common surprise is seeing that it shows absolutely no sign of life when plugged into the wall. In practice, this happens because ATX power supplies rely on a startup command sent by the motherboard, specifically the green wire known as Power On. To simulate this behavior on the bench, we need to jumper the green wire with any black ground wire using a small piece of wire or a switch, allowing the internal circuit to understand that it must begin supplying power to the components.

However, simply turning on the power supply does not guarantee that it is operating perfectly or under real-world usage conditions. Modern switched-mode power supplies utilize short-circuit and undervoltage protection circuits, meaning that cheap or faulty units might turn on without a load, but immediately shut down or deliver completely unstable voltages as soon as we try to draw current from them. This is precisely where the need arises to use a dummy load to simulate the actual power consumption of a processor and a graphics card during testing.

To better understand how components behave under electrical stress, it is worth checking our detailed guide on how to test motherboard voltages with a multimeter when a computer fails to display video. This preliminary knowledge helps correlate the faults you find in the power supply with the symptoms exhibited by the rest of the computer hardware on the repair bench.

What is a dummy load and how to build one

A dummy load is nothing more than a set of robust resistors capable of turning electrical energy into heat, simulating the consumption the power supply would experience if it were feeding a complete machine. Without this load on the 5-volt and 12-volt rails, the supply's cross-regulation fails and voltages spike to dangerous levels or drop below acceptable thresholds. To assemble this setup on the bench, we typically use high-power wirewound resistors, known as cement resistors, mounted on an aluminum heat sink to prevent them from melting under the intense heat generated during tests.

Calculating resistance and power values depends on the rail you want to test, keeping in mind Ohm's law guides us on the relationship between voltage, current, and resistance. For instance, to demand significant current from the 12-volt rail, we need resistors with lower resistance and high power dissipation capacity, otherwise the component will burn out within seconds. Additionally, bench fans or old computer fans can be attached to help exhaust the heat accumulated in the resistors during prolonged maximum load tests.

Measuring output voltages with the multimeter

With the power supply turned on via the jumper and under the load of the resistors, the next logical step is to check whether the main voltages are within the acceptable tolerance stipulated by the ATX specification. The standard tolerates a maximum variation of five percent upwards or downwards on the 3.3-volt, 5-volt, and 12-volt rails. To perform this measurement, we set the digital multimeter to the direct current voltage scale and touch the probe tips to the corresponding pins of the main 24-pin connector and auxiliary connectors.

If you notice that any of these voltages are fluctuating wildly or completely out of the permitted range, the problem usually points to faults in the supervisor integrated circuit or filtering issues in the secondary section. To deepen your diagnosis on semiconductor components that commonly fail in these stages, it is worth reviewing the article on how to test Schottky diodes on graphic cards, whose semiconductor principles apply perfectly to the high-current rectification stages of ATX power supplies.

How to analyze ripple and electrical noise with the oscilloscope

Although the multimeter is excellent for measuring the average DC voltage value, it is completely blind to ripple, which is that residual high-frequency undervoltage left over after energy rectification and filtering. In practice, excessive ripple can cause random system crashes, corrupted files on hard drives, and chronic processor instability, even if the multimeter shows that the 12 volts are perfectly correct. To visualize this noise, we need to connect the oscilloscope to the power supply output using AC coupling, which blocks the direct current component and magnifies only the small unwanted variations.

When configuring the oscilloscope, it is essential to use the probe with the shortest possible ground loop, connecting the alligator clip very close to the probe tip to avoid capturing environmental electromagnetic noise from the workshop that would mask the true reading. An acceptable ripple on the 12-volt rail should remain below 120 millivolts peak-to-peak, while on the 5-volt and 3.3-volt rails the tolerable limit is 50 millivolts. If these values are much higher, the culprit is almost always worn-out electrolytic filtering capacitors, a topic you can master better by checking our guide on how to test electrolytic capacitors with a capacitance meter on the bench.

Safety precautions and final considerations

Working with open power supplies on the bench requires extra attention, because even after unplugging the unit from the wall, the large capacitors in the primary circuit retain dangerous voltages exceeding 300 volts for several minutes. Never touch the hot heat sinks of the load resistors and always wear safety glasses in case any component suffers excessive thermal stress. Testing with a dummy load and an oscilloscope turns power supply maintenance from blind guessing into a scientific and precise diagnosis, ensuring no expensive hardware is connected to an unstable power source.