Some users don’t put performance on top, but silence. This article is dedicated to these users because even the PSU can be a notable noise source, especially high power density or low-efficiency ones. Since I have access to Cybenetic’s internal database, hosting thousands of results, it isn’t hard to extract the required information. I will base my findings on 115V and 230V input, which typically is the worst-case scenario since efficiency is lower at lower voltage input. If I have many entries in a category, I will limit them by listing only the PSUs with below 30/35/40 dBA average noise output depending on the Wattage.
UPDATED 13 February 2023
I will only include in today’s quietest PSU picks article only products available on the market. Performance doesn’t count, nor does efficiency, although silent PSUs are usually highly efficient. Moreover, I will have an extra category for the passively-cooled PSUs, which are not so many and have limited capacity.
If you want the best PSU picks according to performance or ATX v3.0 compatibility, you should look at the articles below:
- Best ATX v3.0 & PCIe 5.0 Ready PSUs Picks 2023 – Hardware Busters
- Best PSUs Picks 2023 – Hardware Busters
I will put some affiliates link for the products that I will recommend. You don’t pay more by using these affiliate links, but you help me keep this site alive and kicking!
Output Noise Measurements Test Set-Up
In the schemes below, you will find an outline of my hemi-anechoic chamber with the exact placement of the MIC and the DUT. I strictly follow the corresponding ISO Standards:
- ISO7779: 2018 – Acoustics – Measurement of airborne noise emitted by information technology and telecommunications equipment.
- ISO 11201:2010 (E) – Acoustics – Noise emitted by machinery and equipment – Determination of emission sound pressure levels at a workstation and at other specified positions in an essentially free field over a reflecting plane with negligible environmental corrections.
Background Noise Calculation
The background noise in our chamber can range from 5.5 dBA to 8 dBA, depending on the external conditions. Therefore, to have the best possible conditions, I prefer to take noise measurements at night, when the ambient noise is lower than during the day.
In no case the background noise can exceed the sound of interest, and in practice, the output level of the DUT has to be at least 3 dBA higher than the background noise for the measurement to be accurate. Still, a correction has to be applied to get the correct result. The background noise correction is defined as K1, and the amount of the measured source level is reduced to obtain the background noise-corrected source level.
The procedure for measuring a DUT’s noise output is the following:
• Measure the total noise level (LDUT + LBG) with the DUT in operation, paying extra attention to its lowest noise output mode.
• Measure the background noise level (LBG) with the DUT switched off.
• Calculate the difference between the readings mentioned above (LDUT – LBG). If it is less than 3 dBA, the background noise is too high to measure accurately. A correction must be applied if it is within a 3-10 dBA range. There is no need for correction if the difference is greater than 10 dBA, but I still use it for up to 20 dBA differences.
The formula for calculating the noise source without the influence of the background noise is the following:
In a spreadsheet, the formula above can be given as 10*log(10^(LDUT/10)-10^(LBG/10))
The chart below can be used to make background noise corrections to the source signal.
In this category, most PSUs have low efficiency; hence energy losses are increased, making it a challenge to have relaxed fan speed profiles. Nonetheless, there are low average noise output units, like the Corsair CX450M, the Asus TUF 450B, and the XPG Pylon 450.
be quiet! has the quietest 550W PSU entry, followed by Corsair and Super Flower.
be quiet! leads the silence race in the 650W category, followed by Cooler Master and Corsair.
Asus has the quietest 750W PSU, with the new Corsair RM750x Shift following. The German brand, be quiet! takes over third place.
Asus also dominates in the 850W category with SilverStone, and be quiet! following, with products having below 20 dBA average noise output, which is excellent for this capacity.
The 1000W category includes dead silent PSUs, with two below 15 dBA average noise output, which is a fantastic result for a high-capacity unit. Asus has two units in the top three, and FSP’s Titanium PSU takes over second place.
In the high-power category, be quiet! achieves first place, with Cougar taking over the following two places.
The dominance of be quiet! continues in the 1251-1500W category with the Dark Power Pro 12. Corsair is following, with XPG coming third.
The Asus Rog Thor 1600W achieves an impressive result, which is far away from the competition. Seasonic comes second, with Corsair’s AX1600i taking third place in the 115V noise chart. I don’t have 230V data for the Rog Thor 1600W, but I am sure it would achieve first place in the corresponding chart.
As crazy as it sounds, because it is the highest capacity SFX unit, the Cooler Master V1100 SFX Platinum achieves the lowest noise output, followed by the budget Raijintek Ermis 550B and the SilverStone Extreme 850R Platinum.
Asus has the upper hand in the SFX-L category with its Loki models. Seasonic’s unit comes third. Corsair should take care of the fan speed profile of the SF850L to drop it lower. For the SF1000L, there is room for improvement, too.
There are several passive PSUs available in the market today. Still, if you ask for my expert opinion, it doesn’t worth spending so much to get one of them, unless you need the absolutely lowest noise output possible.