Noctua NF-S12A PWM: Example of unfortunate design

Cooling, Extra tests, Fans | | Ľubomír Samák

They never were as brilliant and amazing as marketing materials depicted (and still depict) them. Fans that are supposed to be top-tier for system positions are, at best, average. Well, judge for yourselves. We wouldn’t subscribe to the notion that these are currently aerodynamically attractive fans for PC cases. For the reason that they are not. And we can go straight to the specific reasons and arguments.

Wordpress Table Plugin

* When reading performance metrics, always account for a certain tolerance. For maximum RPM, a ±10% variance is typically noted, while minimum RPM can vary significantly from unit to unit—sometimes manufacturers even allow for ±50%. This must also be properly factored into airflow, static pressure, and noise level values. If a table cell lists only one value, it always refers to a situation at maximum RPM achieved at 12 V or 100% PWM duty cycle. The lower performance limits are not disclosed by the manufacturer in such cases. The price in the final column is always indicative only.
** Price is calculated per unit, but the fan is not sold individually. It is only available as part of a three-pack that also includes an ARGB hub.


Contents

   

Related articles

Cooling Topics at HWCooling Techtour ‘26 Belonged to DeepCool

Cooling, News | | Ľubomír Samák

Do you know how a heatpipe works? A DeepCool representative, in a talk at HWCooling Techtour ‘26, looked under the hood. In other words, he explained much of what often remains hidden. This included how they approach development and what they work on internally. That, too, was part of the program, which you can watch. The physical samples, which were also on display, come only at the very end of the whole journey. Read more “Cooling Topics at HWCooling Techtour ‘26 Belonged to DeepCool” »

Contents

   

Fan frame daisy-chaining trend. When is it (not) helpful

Analysis, Fans | | Ľubomír Samák

Easier installation? Yes, that’s why it’s done. Daisy-chaining fans using contacts in their frames is great, but it often also comes with pitfalls that reduce “cooling efficiency”. Manufacturers deal with this in different ways. One does it technically one way, another differently. In any case, they know that a frame that is too thick means less space for the blades. And that does not improve airflow per unit of noise. Read more “Fan frame daisy-chaining trend. When is it (not) helpful” »

Contents

   

Arctic BioniX P14 A-RGB: Elegant daisy chaining is just the beginning

Cooling, Extra tests, Fans | | Ľubomír Samák

High cooling efficiency is the foundation, but with the BioniX P14 A-RGB fans, Arctic has also made efforts on other fronts. In addition to the clever daisy-chaining method (with the aim of reducing cabling to a minimum), this includes, for example, illumination as intense as possible. The light is guided not only through the entire impeller, but also through the surface of the frame, from the side. There are light guides in these areas as well. Read more “Arctic BioniX P14 A-RGB: Elegant daisy chaining is just the beginning” »

Contents

   

Comments (22) Add comment

  2. when A12x25 came they overtaken both the pressure optimized and airflow optimized predecessors
    and the main issue I have with S12A is they become noisy before they start moving any air

    sure, if you can accept their full speed noise they are quite decent, but long time has passed since their release and you probably can buy 3 cheaper ones with similar characteristics at that noise level nowadays
    or pay some more for A12x25 or A12x25G2 and get superior performance at lower noise lever
    personally I’d simply either remove them entirely from the lineup or move them down to the redux line to be much cheaper, but again the motor and bearing in these is bad enough I see no place for them at all on the modern landscape

    EDIT: from empirical experience I wouldn’t say S12A use SSO2, and I don’t think they had that in specs when I bought one, my box says SSO only and certainly sounds different than A12x25 (might be a motor only,though the sound is pretty much mechanical), but at the same time on the back of the hub there’s SSO2 marking, interesting

    Reply

    1. — „… I wouldn’t say S12A use SSO2, and I don’t think they had that in specs when I bought one…“

      I would also have guessed—based on intuition—that the NF-S12A (PWM) fans were released later than SSO2. But it’s possible that these bearings were introduced into the fans over time as a “silent” revision? Hard to say… 🙂

      Reply

      1. yeah, as noted later my box says SSO, but the fan itself has SSO2 sticker, but something about it sounds much different than A12x25 does, if it’s the same bearing it has to be the motor
        whichever it is it can barely spin up because the “bearing-like” noise gets annoying at extremely low speeds and the impeller design at these speeds causes basically no air movement at all, A12x25 set to inaudible from the same distance is many leagues ahead

        Reply

        1. The difference may come from the motor, for example in how it switches between individual poles. But naturally, the sound you’re hearing can be influenced by various elements. 🙂

          Reply

  3. I wish it had been BlackNoise’s eLoop that had become the memetic high end, high perf, low noise fan line. I’ve heard recordings of them compared to their contemporary Noctuas and they’re far nicer on the ear, pleasant sort of Sci-Fi humming and they outperformed their Noctua opposite number too, IIRC. With ARGB.

    Reply

    1. eh, for me eLoop is really whiny, seems the motor dislikes low speeds a lot as this noise gets quieter as the speed rises, but before it disappears the aerodynamic noise takes over so at no speed these are usable for me

      Reply

      1. Several fans tend to “whine” at the very bottom of their speed range. It’s usually because the voltage is already too low for the motor to operate comfortably. With PWM, the situation often improves (even at otherwise comparable speeds), since the fan continues to receive 12 V, just with very long gaps between individual pulses. 🙂

        Reply

        1. Yeah, it’s a fairly common feature to set a floor in the PWM or whatever silicon so the things won’t trigger outside of their tolerable noise ranges these days, IIRC. Both the Pro arctics and the InWin Neptune series boast this.

          Honestly, to be top-end in the current market, they’d just need to find a way to improve the light spreading of the polycarbide and add a trigger floor to the PWM silicon on the eLoop IMO. Once they had that, and maybe try and make a top-end noise and performance AIO and heatsink… maybe even get exclusive OEM rights from the first client 80 Plus Ruby PSU, I think they’d be able to make a comeback as a client fan line, not much amendment to their old design needed at all.

          Reply

        2. that should be the case generally, unfortunately B12P is already PWM and it still whines a lot
          also direct PWM drive works as you described, but latest Noctua fans have some “interpreter” that alters voltage to eliminate heavy “kicks” of torque that may sound clicky, don’t them?

          Reply

          1. Hmm, in my opinion PWM control works with 12 V pulses that are delivered at lower duty cycles at lower speeds. Noctua fans have a controller that modifies these pulses in some way, but in my view it doesn’t actually reduce the voltage to a lower level. The voltage is then reduced only when using adapters (L)NA. 🙂

            Reply

            1. well, the graphs they presented suggest signal taking a different shape to a square wave, with the leading edge being smoother
              I’m not sure how can that be interpreted, lowering voltage would make more sense imo, otherwise they’d have to do kind of PWM-in-PWM and hope that doesn’t behave clicky? I’d assume full torque applied even for a 1% of an usual PWM cycle would still end up causing a rather sudden application of force

              that potentially could also be a current limiter instead of voltage, as in the end it gives the same result of lower power
              the most important part being: it’s validated to work this way, and with LNA in place too, so the motor doesn’t complain, while other fans may have never been internally tested with lower voltages

              but then again, with B12P I’m pretty sure even when plugged in directly it whines at low PWM, which would be a different topic
              maybe an artifact of slow pole changing?

              Reply

              1. Yes, the PWM waveform presented by Noctua—with a sloped rise instead of a sharp edge—could indeed be quieter. However, I assume the difference will be so small that it will fall below the resolution of human hearing. It seems like one of those details that looks impressive, but from which no major impact should be expected. That said, it’s a beautiful detail that clearly shows how delicately Noctua approaches fan development.

                — „… maybe an artifact of slow pole changing?“

                I really don’t know, but it’s possible. 🙂

                Reply

                1. for what it’s worth with A12x25 at slow speeds I can hear slight ticking/clicking when I put my ear close enough while on G2 I haven’t experienced that
                  not that it matters from regular distance or for most users that run them twice as fast and aero noise would mask it even if it was still present, but that’s a difference I noticed

                  Reply

                  1. Do you also hear that clicking noise on multiple NF-A12x25 G1 units? If so, it’s possible that this is handled differently at the design level in the NF-A12x25 G2—but I can’t say for sure. I assume Noctua would have highlighted this if that were the case… 🙂

                    Reply

        1. I’m given to understanding the best performing of that linage is the 140mm designs, where they have a fair claim to being the best general purpose ARGB fans on the market.

          Reply

          1. In the 140mm format they have an odd number of blades, which is a general standard in the field of axial PC fans, but the 120mm fan has an even number of blades, which is quite unique. 😉

            Reply

        2. Ha! As soon as it becomes possible, we’ll test the 120 mm NB-eLoop fans. With Noiseblocker, it’s difficult for us to establish communication (we haven’t been able to find a suitable contact channel), and distribution is also weaker. Some NB samples were purchased for us by mino_85 a long time ago. Years have passed since then, and we still haven’t managed to get to those fans. So this factor plays a role too—those unfortunate time constraints that simply have a hard limit. You know I’d love to test everything right away, but unfortunately I don’t know how to use a magic wand yet, haha.

          Reply

          1. you should ask Elsa or Potter or Saxana… 😀 🪄🎩
            I don’t have a magic wand… 🙁
            By the way, … didn’t it come with a manual? 😀

            Reply

      2. An even number of blades is indeed quite rare. It shouldn’t lead to particularly attractive results in terms of acoustic characteristics. 🙂

        Reply

Leave a Reply

Your email address will not be published. Required fields are marked *