News New Alphacool fans with noise isolation via metal frame

[StefanR5R]

Earlier this month, Alphacool announced 120 mm fans (a 2000 RPM and a 3000 RPM model) with a decoupled metal frame construction:
https://www.alphacool.com/en/news/new-apex-stealth-metal-fan-2

In short, it is a quite normal plastic fan in a round plastic frame, which is then suspended via rubber rings in a square, heavy metal frame. As far as I can tell, the main difference of these from your run-of-the-mill fans is that the latter aren't easy to mount mechanically decoupled and at the same time airtight for good pressure and flow rate.

A first in-depth review is out now:

in German: https://www.igorslab.de/alphacool-a...intergrundinformationen-und-vielen-messungen/
translated: https://www.igorslab.de/en/alphacool-apex-stealth-metal-power-fan-in-an-exclusive-review/​

According to these measurements by igor'sLAB, both aerodynamic and acoustic performance are quite a bit ahead of the current top of computer fans.

Construction cost is high, and so is the price: € 30 a piece. Availability is supposed to be sometime later this year.

A specific drawback of these fans, mostly only relevant if they are to be used as CPU tower cooler fan: The weight would put more stress on the cooler mounts. Apart from that, the high weight of the frame of these fans is one important reason for their supposedly excellent acoustics. Other contributing features are good precision in manufacturing (according to igor'sLAB), a 6 pole motor, and fluid dynamic bearing.
________________________________________________________

I am currently in the process of building a watercooled computer which is meant to be running 24/7 during the colder months of the year and with quite some heat output. Fans with such a good performance would be good to have for such an almost-always-on computer. However, before I upgrade fans, I first need to thoroughly decouple the pump. As long as I haven't done this, fan noise of that computer isn't really relevant. The fans which I currently have, Arctic P12, are cheap but are doing their job while being drowned out by the pump. As it happens, P12 PWM PST can be had for € 30 too — but as a pack of 5, not as a single piece.

 

[crashtech]

I wish they were ball bearing for longer life, given the expense. Afaik ball bearings increase noise at least slightly though, so that might have been a reason not to supply them.

 

[StefanR5R]

Apparently reviews from other sites are appearing now. So far I only came across der8auer's video review (English version/ original German version). Conclusions by der8auer:

1. It is a relatively quiet fan at 1000 RPM (at which many 120 mm fans are reasonably quiet) and at 2000 RPM (somewhat more so than other fans).
   Note, der8auer tested on an open bench = without case which would necessarily resonate to some degree, and he tested in a normal room where you can't quantify noise very well, putting it diplomatically. Still, I wouldn't disregard these findings of his.
   
   ​
2. The thermal performance in CPU cooling with this fan, tested with a 250 W CPU and a liquid cooling with dual fan 240 mm radiator, as well as with a single fan 120 mm tower style air cooler, was average at 1000 RPM and below average at 2000 RPM.

IgorsLab's earlier test within a mini air tunnel, in contrast, found far above-average aerodynamic performance of this fan, and apparently nobody else can reproduce this. It seems some comments were made by others that IgorsLab's use of a vane wheel anemometer was not combined with adequate straightening of the fan's downstream. Meanwhile, Igor Wallossek has updated his article with a brief preface about these discrepancies (English version, original German version). Among else he says that he is working to identify the cause of the deviations between IgorsLab's and others' tests.

What I am thinking so far:

• I for one wouldn't expect dramatic improvement of aerodynamic performance from a new variation of a 120 mm × 25 mm radial fan.

• Nor do I expect revolutions with respect to air noise. Low air flow → low air noise. More air flow → more air noise.

• Good build quality and good quality assurance processes at the manufacturer are the recipe for low motor noise and low bearing noise. The latter also calls for FDB, not ball bearings.
  Good and constant manufacturing quality costs money.

• Now, what happens when the end user straps a fan to a computer case, an air cooler, or a liquid cooling radiator? Depending on the case etc., some resonances will inevitably happen. To counteract them, you need to suspend the fan in a way which decouples the fan as good as possible, and also dampens the vibrations of the fan frame a bit. But you also want a somewhat pressure tight seal between fan frame and case, or fan frame and radiator. Which makes it harder to suspend the frame in a decoupled fashion.
  
  This is where Alphacool's idea comes in: Implement the decoupling within the fan frame, and also add a bit of weight to the case-side/ radiator-side part of this construction. That way, either the user doesn't need to provide his own decoupling solution anymore, or the built-in decoupling works as an add-on to the user's decoupling efforts.
  How well Alphacool executed this idea — notably, how soft the rubber O-ring within their frame construction is — that's not yet known to me.

 

[StefanR5R]

• Nor do I expect revolutions with respect to air noise. Low air flow → low air noise. More air flow → more air noise.

PS, of course this puts it very simple, too simple certainly. A significant part of the air noise comes from obstructions which are very close to the fan blades. The designer of the fan can influence the number and shape of blades, and the spokes of the fan frame, and that's it. All the rest which influences air noise, i.e. near and far obstructions, depends on the application, i.e. can't be optimized via fan design within the constraints of a given fan form factor, e.g. 120×25.

 

[crashtech]

Looking again at these fans, I wonder if they will make replacement motors available. They appear to be serviceable from the illustrations.

 

[MrTeal]

Cybernetics also has testing up on the new fan, using a much more comprehensive test than debaurer's.

Cybenetics Labs – PSU Efficiency & Noise Level Certifications - Fans

www.cybenetics.com

The Alphacool fan appears to be a good fan, but not a crazy revolution.

STS has a review at all with comparison audio.

The noise profile of the Apex Stealth at least in that video is very nice, especially compared to the Phanteks. Performance at 25dBa (~1700RPM) might not have as good a static pressure and and airflow as the T30, but it sure isn't as whiny. It'd be nice to normalize at a CFM through a radiator and then evaluate noise though, since at the same P-Q the T30 would probably be spinning slower.