Fan through a frame?
Due to the atypical spacing of the mounting holes for its format, the new Noctua NF-A14x25r G2 PWM fans cannot be installed in 140mm positions as standard. You need a conversion frame for that, and one such frame was designed, made and sent to us by an HWCooling reader. So, we were able to simulate and test one of the most anticipated fans of recent years (NF-A14x25 G2 PWM) already now.
The new generation of 140 mm fans has already come out, but only “partially”. This in the form of the Noctua NF-A14x25r. While these have the same aerodynamic design as the more versatile NF-A14x25 G2 PWM, they are mainly tailored for use on tower cooler heatsinks that have their clips customized. Like the NH-D15 G2 cooler, for example. For most other applications, the mounting hole spacing (105 × 105 mm) corresponding to standard 120 mm fans is unsuitable.
Some cases with “dual” positions for both 120 and 140 mm formats will accommodate these fans, but if the typical hole spacing for 140 mm fans (i.e. 125 × 125 mm) is required, you’re out of luck. That is, until you reach for a frame to which you could attach the fan using 105mm holes and then install the frame itself using 125mm holes.
A reader nicknamed “the patient” sent us such a frame. They also designed the frame themselves, and the source files for the 3D model are public. You will be able to print the frame yourself, if it makes sense to you. It fits not only the Noctua NF-A14x25r G2 PWM fan, but also other 140mm fans that use the 105×105mm hole spacing.
Basically, it is a 5 mm thick sheet with outer square dimensions of approximately 137.5 mm per side. The inner circle has a diameter of 132 mm. Depending on the material used and the 3D printing itself, it is possible that these dimensions will vary slightly. In the case of the Fiberflex 40D piece we received, we can say that it fits well and holds its position well with the fan as well. This despite the fact that the elasticity of the material is high, but at least it does not crack during screwing, which would reduce the mechanical durability of the whole solution. On one side of the frame, there are also well-modeled holes for recessing the screw heads.
Admittedly, a 4mm wider opening wouldn’t hurt (the frame walls wouldn’t interfere with the Stepped Inlet Design steps), but this one still avoids the axis of the blades, which is the main thing from a performance point of view.
The test results will differ somewhat with the results of the “September” square-frame fans for a number of reasons. We mentioned one in the previous paragraph, and then there’s something else. Although the frame, when installed on radiators, seals well in the most critical places, around the fan mounting holes, there will still be some leaks through which pressure will be lost. These are due simply to the frame being perched on anti-vibration pads, which are the highest points of the fan tunnel. This means that there is a narrow gap on the sides, which is not the case with the NF-A14x25 G2 PWM models. For the reason that these fans will have a full-circuit gasket pre-installed. This compensates for any height differences with a view to avoiding a drop in static pressure and thus airflow through an obstacle. For example, through a radiator. But the latter will be only one of the three obstacles which we will test with the frame on the NF-A14x25r G2 PWM. The others are a plastic filter and a hexagonal grille.
Different results than with the fans with originally a square frame (NF-A14x25 G2 PWM) are also obtained with this combination with the DIY frame due to the different impeller to obstacle clearance, which in this case is larger by the thickness (5 mm) of the frame. Theoretically to a positive direction (greater separation of the leading edges from the obstacle typically on a plastic filter and grille reduces unwanted microturbulence and thus reduces noise with simultaneously higher airflow), but practically it should be noted that the frame used by us does not represent any “optimization” for the NF-A14x25e G2 PWM, and on the contrary is characterized by features that can degrade the performance-acoustical characteristics. And that’s whether it’s a collision with the SID or other imperfections, such as unwanted grooving of the tunnel’s inner wall, which is a result of the natural manufacturing process of 3D printing by layering.
The tests are following the standard methodology, in the standard wind tunnel, but with a non-standard number of only one piece. We don’t have more NF-A14x25r G2 PWM fans, so keep that in mind. We’ll find out how one piece might differ from another later, comparing 15 samples following the same template as this test of the inexpensive Endorfy Stratus 120 PWM fans. But don’t write that to Noctua yet, lest they start making a selection of “better pieces”. 🙂
If you’re interested in details of the new Noctua NF-A14x25 G2 PWM fans, you’ll find them in this article.
- Contents
- Fan through a frame?
- Results: Airflow through a plastic filter
- Results: Airflow through a hexagonal grille
- Results: Airflow through a thinner radiator
- What we have arrived at
Hmm, a supposedly top fan performing so poorly on plastic filter with a spacer… A possible explanation is that the adaptor seems a bit leaky on both sides, and the fact that the gap (between the fan and the adapter) is spaced away from the obstacle might have introduced a significant path of least resistance. This issue is less apparent on obstacles with less impedance (radiator and hex grill).
To validate whether the adaptor is working as intended, I would suggest testing the adapter on one or two other fans with anti-vibration pads (perhaps Noctua A14), at one or two noise levels, to see how much impact the adaptor will makes. Of course, if time allows.
There are more of those little leaks. They are also on the rounded sides (between the fan and the frame) and then also between the frame and the obstacle. Fiberflex 40D is a very elastic material and the shape of the frame is also shaped by the force with which it is attached to the fan or to the obstacle. A more rigid material with higher hardness would not be as pliable, would not flex as much, and the results might be better. It is hard to say by how much. We will be able to speculate on that after we know the results of the NF-A14x25.
Among 140mm fans, the airflow of the Noctua NF-A14x25r G2 PWM with the DIY frame through the filter is below average, but still always higher than the NF-A12x25. Note that the Toughfan 14 Pro is not much better and on the grille the Thermaltake fan is already lagging behind. Well, we’ll see, we’ll get to a proper conclusion in September. 🙂
The model already supports fan sizes ranging from 200 mm to 92 mm. Just keep in mind that it may perform slightly better without anti-vibration pads. Please take the code with a grain of salt, as I’m not a professional programmer. I would gladly accept your suggestions. Of course, you are welcome to modify and improve the code as you see fit.
Thanks for the follow-up.
A combination of a stronger material that would be less subject to deformation due to the screws (to keep the flexing of the frame as small as possible) and a hair larger hole (to accommodate the Noctua SID) would probably lead to better results. The question is by how much, it would probably only be to a negligible degree. I suppose the biggest improvement would be better sealing of the frame from both sides, as each gap reduces static pressure, but I wouldn’t see even that as dramatic in this case. The key to functioning on a radiator in my opinion is that there is a good amount of sealing at the corners. With these we already have an idea of what effect they can have on the airflow from the tests of different corners on the BeQuiet! Silent Wings Pro 4.
I was concerned about the strength of the narrowest point by the central opening. The cross-sectional thickness in this area is 3.5 mm, which I believe is a reasonable minimum (though the model allows shrinking this dimension to 3 mm). I didn’t want to exceed the standard fan size format, but increasing the main dimensions could provide greater strength. As for the material, it might be better to use something with a 60D hardness, which could also be a better option from a printing perspective.
Here’s a direct link to the archive containing the model for English-speaking readers.