Noise measurement for PSU and case fan acoustics
A quiet PC is shaped by more than the specification printed on a fan box or power supply label. Motor tone, bearing noise, electrical buzz, airflow turbulence, resonance, and the interaction between components can determine whether a system sounds unobtrusive or distracting at a desk. That is why Hardware Hounds treats acoustic testing as a repeatable measurement process rather than a quick impression from a listening session.
Our approach combines calibrated instruments, controlled operating conditions, and careful listening. Sound-pressure readings provide comparable data, while subjective notes help identify qualities a single decibel figure cannot describe. The result is a clearer picture of how a PSU or case fan behaves at idle, during gaming, and under sustained load.
Building a controlled acoustic environment
We perform measurements in a quiet room with the test system isolated from desks, walls, and other large reflective surfaces. The room’s background level is recorded before every session. If ambient noise rises enough to influence the result, testing is paused rather than allowing an inaccurate reading into the chart.
The computer is positioned consistently, with the microphone placed at a fixed distance and height from the primary noise source. We keep cables, mounting hardware, and panels in their normal positions because removing a side panel or suspending a power supply can produce results that do not represent real ownership. The same orientation is used across products whenever practical.
Environmental conditions are logged as well. Temperature affects fan control and PSU cooling behavior, while humidity and nearby equipment can influence the background floor. These variables may seem minor, but controlled testing makes small differences between products easier to trust.
Calibrating microphones and sound meters
Acoustic testing starts with calibration. We use a calibrated sound-level meter or measurement microphone with a known reference source, checking the instrument before a test run and verifying it afterward. This confirms that a reading has not drifted because of handling, battery state, or equipment changes.
Measurements are generally taken using A-weighting, written as dBA, because this weighting approximates the sensitivity of human hearing at moderate volume levels. For deeper electrical hum or low-frequency fan resonance, we also examine frequency content when the equipment allows it. Two products can show similar dBA values while producing very different listening experiences.
The microphone is kept away from corners and hard reflective surfaces, and its position is marked so that repeated tests use the same geometry. We avoid touching the stand or cable during a reading. The meter records a stable interval rather than a brief peak, with unusual spikes documented separately instead of being treated as the product’s normal acoustic output.
Testing case fans at useful operating points
Case fans are tested at several controlled speeds rather than only at maximum RPM. We record behavior at a low-speed idle setting, a common mid-range value, and full speed. When a fan supports PWM control, the controller signal and measured rotational speed are recorded together because a nominal 50% setting does not produce the same RPM across different models.
Each fan is tested in the same case position, with the same filter, grille, and mounting arrangement. This matters because restrictive mesh can increase turbulence, while a rear grille or drive cage may create a distinctive rushing sound. We test both the fan’s acoustic output and its effect on system temperatures, since a quieter fan that moves too little air may not be a practical improvement.
We listen for tonal behavior during speed changes, including clicking, ticking, bearing chatter, commutation noise, and sudden resonance. A smooth broadband airflow sound is usually less intrusive than a narrow, high-pitched tone at the same measured level. Notes from these observations accompany the numerical results so readers can understand the character of the noise.
Measuring PSU acoustics under load
Power supplies are evaluated at multiple output levels that reflect common usage. Low-load readings show whether the fan remains stopped or starts at an unexpectedly high speed, while gaming-level and sustained-load tests reveal the cooling curve during realistic operation. Higher loads help expose fan ramping, bearing noise, and coil whine that may not appear on the desktop.
The PSU is tested inside the same system or on a controlled load platform, depending on the review procedure. Input voltage, output load, fan mode, and internal temperature are recorded. Semi-passive models receive extra attention because their fan-start threshold can create a noticeable transition between silence and active cooling.
Electrical noise is handled separately from airflow noise. Coil whine may come from inductors or transformers and can vary with graphics-card frame rates, power states, or specific loads. We document whether a sound is present at idle, during menu rendering, under GPU load, or during a combined CPU and GPU workload. A dBA reading alone cannot communicate the annoyance factor of a sharp, changing buzz.
| Test condition | What it reveals | Typical measurement focus |
|---|---|---|
| System idle | Fan-stop behavior and background noise | PSU silence, low-speed ticking, residual room level |
| Light desktop load | Low-power control response | Fan start threshold and tonal changes |
| Gaming load | Realistic sustained behavior | Fan RPM, airflow noise, coil whine |
| Combined CPU and GPU load | Worst-case thermal demand | Maximum acoustic output and ramp stability |
| Fixed fan speed | Direct product comparison | Noise generated at equivalent RPM |
| Full-speed operation | Upper performance limit | Peak dBA, turbulence, motor and bearing sounds |
Separating useful data from misleading numbers
Sound levels are logarithmic, so a small numerical difference is not always meaningless, while a large-looking difference may be less obvious in a noisy environment. We avoid presenting a single reading as an absolute verdict. Instead, results are shown alongside the test distance, ambient level, operating condition, and whether the reading represents an average or a stable plateau.
The practical noise floor also matters. If a system measures close to the room’s background level, the result should be described as below or near the measurement limit rather than treated as a precise zero. This is especially important for fan-stop PSUs and low-speed case fans, where room ventilation or distant traffic can dominate the microphone.
Repeat measurements help identify outliers. We run each condition more than once, allow temperatures and fan speeds to stabilize, and investigate any result that changes unexpectedly. If a product oscillates between two fan speeds, that behavior is retained as part of the finding instead of being averaged away.
Connecting acoustics with thermal performance
Noise data has value when it is connected to cooling performance. A case fan may be very quiet at 800 RPM, but its airflow can be insufficient for a high-power graphics card. Conversely, a louder fan at 1,500 RPM may provide a meaningful temperature reduction when the system is under sustained load. Our comparisons therefore consider acoustic output, temperature, airflow setting, and power consumption together.
For PSUs, the quietest result is not automatically the best result. A fan that starts late but ramps aggressively can produce a sudden burst of noise, while a slowly increasing fan curve may remain less distracting even if its peak reading is similar. We also report whether the PSU maintains stable behavior over a long session rather than judging it from a short benchmark run.
This combined view helps distinguish genuinely efficient products from components that achieve a low reading only under limited conditions. It also gives builders useful context for selecting a fan curve, choosing a case layout, or deciding whether a higher-capacity PSU is worthwhile for lower fan activity.
How readers should interpret our results
Our charts are designed for comparisons made under the same test conditions, not as guarantees of identical results in every room. A different case, desk position, ambient temperature, or graphics card can change the final sound. Fan-control software and motherboard headers may also use different PWM or voltage behavior, affecting the operating point.
When we describe a product as quiet, we consider both its measured output and its acoustic character. A low dBA result with intermittent clicking may be more noticeable than a slightly higher broadband airflow reading. Listening notes therefore remain part of the review, particularly for coil whine, tonal motor sounds, and resonance that instruments may understate.
Readers who notice an unusual behavior in their own hardware can share details through the Hardware Hounds contact page, especially when the sound changes with a particular load or configuration. Specific information about the case, fan speed, PSU model, and workload makes it easier to compare experiences meaningfully.
Recommendations for quieter builds
The most reliable way to reduce system noise is to match component capacity and cooling demand before adjusting fan curves. These practical steps generally produce better results than chasing the lowest idle reading:
- Choose case fans with a smooth motor and bearing design, then compare their noise at the same RPM rather than relying on maximum airflow claims.
- Leave clearance around fan intakes and exhausts so grilles, filters, and cables do not create avoidable turbulence.
- Select a PSU with enough capacity to remain efficient and within its preferred fan-control range during normal workloads.
- Use a gradual fan curve to prevent repeated speed changes caused by short temperature spikes.
- Check for coil whine separately from fan noise, since changing frame rates, power limits, or component combinations may affect it.
A quiet system should remain quiet during the tasks it is built to perform. By measuring PSUs and case fans at controlled speeds, realistic loads, and consistent distances, Hardware Hounds turns acoustic impressions into useful evidence for purchasing and upgrade decisions. Explore our hardware testing coverage to compare sound, thermals, and value before your next build.