Question Fan Spec Comparisons Useless?

[Mantrid-Drone]

With my latest build I'm currently looking at fans, deciding what to buy. There's an aesthetic element involved but I'm most interested in performance.

I'm actually making a comparison chart of all potential 120mm and 140mm fans.

Except for Akasa few fan suppliers seem to provide noise level dB(A) figures for the range of speeds the fan supports. The figures all appear to be single figure dB(A) and CFM ratings at max RPM.

In my primary PC I have Akasa Viper 120mm CPU cooler and case fans capable of 1900 RPM, used along with a couple of 140mm (1600 RPM) case ones. I have the BIOS set to use 75% PWM on all of them and even under heavy load in the height of summer, it rarely reports four figure speeds. Even when it does they're kicking in temporarily to get the temperature down and go nowhere near their 1900 RPM max speed.

So the question: is it fair to judge a 1900 RPM fan like the Akasa Viper 120mm whose published figure is 28.9dB(A) at max speed when it is being used more typically at half that speed?

At the slowest speed supported, 600 RPM the figure quoted is 6.9dB(A).

I realise the relationship between fan speed and noise level is not linear, the figures for the Viper show that clearly but let's assume it is.

Averaging the Akasa Viper figures out using the lowest and highest numbers at 1000 RPM around 44 CFM/13dB(A) as the more typical airflow/noise level.

If you look at the Noctua NF-F12 with its stated 1500 RPM max/55 CFM/22.4dB(A) and do a quick, again linear, comparison with the Akasa Viper at 1500 RPM the Akasa appears significantly better at 66 CFM/17.68dB(A).

The better spec Noctua NF-A12 figures using the same criteria are 45CFM/16.95dB(A) ie. near enough the same noise level but 30% less airflow.

At 1000 RPM the NF-A12 beats the Akasa Viper on noise at 11dB(A) so it is quieter but again the, roughly calculated, airflow at that speed is 30 CFM, 33% less.

Of course this all relies on manufacturers' figures being trustworthy. If you accept they are: surely just comparing fan dB(A) and CFM at max speed is not like for like and therefore useless to potential customers. Wouldn't both max and min CFM/dB(A) figures be more helpful? Why not a mid-range figure like 1000 RPM (for 120mm fans) too?

 

[Paperdoc]

While I can accept that cheap sleeve bearings are made of plain metals, the most common version of those I encountered for decades was what I called "Oilite" bearings. These were brass sleeves around a steel shaft, but the brass was cast in a particular way to include microscopic porous voids, then soaked in hot oil for a time before installation. That had the effect of providing an invisible oil reservoir in the bearing sleeve. Still, those hearings did require re-oiling over their lifetime. In smaller items they usually were so securely fastened into the frame they were permanent. Sometimes the mounting location included space and a felt pad as an additional oil reservoir. In some large applications they were press-fit into the frame ends and you could actually remove them and re-install them. Some people believed (not sure whether true or not) you could renew the original oil load. They would remove the bearing sleeves, clean them thoroughly to remove accumulated dirt and microscopic metal, then soak them in hot light oil for days before re-installing. Such porous brass Oilite bearings were supposed to last much longer than plain bare metal sleeves because of the built-in oil reservoir system. A further feature was that the brass wore down faster than the steel shaft, so if the bearing sleeves you had could be removed, when they became so worn as to develop excessive clearance you could replace the sleeves. Of course, this type still did not last as long as ball or roller bearings, and wore out much faster in high-load applications. It appears to me the Fluid Dynamic Bearing and Rifle Bearing designs are just attempts to improve that older design. In fact, I wonder whether the "plain" bearings of today still use that older design I called "Oilite".

We should bear in mind also that much of this discussion is about use of fans with their shafts HORIZONTAL. When fans are used in other orientations, the load of the rotating shaft presses on other surfaces than a simple sleeve around the shaft, and that load-bearing surface needs consideration of a bearing arrangement. This applies also to ball and roller bearings. If you go to buy those designs for heavier loads, one important question is always "axial load or radial?"

 

[Mantrid-Drone]

I mentioned that horizontal or upright fan matter in relation to the FDB types which, being sealed ensures the lubrication oil doesn't leak out. As ^^ said earlier capillary action along with purpose designed channels should mean the whole shaft is kept lubricated, at least when running.

The most interesting alternative technology I've come across is the "POM" bearing ie. polyoxymethylene "self-lubricating" used by some Cooler Master fans.

It has been around for years but has not taken the PC fan manufacturing world by storm.

 

[mindless1]

^ I have to trust that MAJOR fan manufacturers, who make their living selling fans rather than various PC accessories and parts, have done due diligence and researched plastic bushings in the past and found that the slight cost savings isn't worth it - except to coolermaster who is trying to shave pennies off.

Oilite is but one brand of the sintered bronze bushing prevalent in (almost) all sleeve bearing fans these days.

The issue is that if the manufacturer tries to cheap out too much, the bushing may have a bad grain structure, or poor tolerance, or the fan blades are imbalanced. All these things (and more like ambient temp) can effect the serviceable lifespan or interval between which they need relubed. Another way the manufacturer tries to cheap out is use of smaller shaft and bushing, note how Panaflow had used a lager diameter shaft and bushing.

Sealed doesn't really mean much, may even be the worst option. A fan with decent plastic molding tolerance and a good quality rubber plug over the bung hole, won't leak out there, but if you remove that plug/access so it's "sealed" then it can't be relubed (easily) and still has the same more likely leaking area which is the other end which seals against, and wears from friction once low on lube.

Granted this is within the context that if someone cares, they relube their fans every once in a blue moon instead of just running them till they seize up. By relubricating them, you *almost* guarantee they won't ever seize up.

Relubricating a sleeve bearing fan is by far the best alternative. If you do it often you can use a low viscosity oil that seeps into the bushing more, but the longer you wait and the more wear there is, the more likely you will need a larger amount of lube and a higher viscosity to make up for the slack in the bearing, which will still reload itself with oil.

I should clarify that. The higher viscosity should come from the same low viscosity base oil, but more solids from grease additives. This results in the low viscosity that seeps into the bushing well, and the grease solids act as a carrier in the case of cheaply designed PC fans that don't have a sufficient reservoir with a felt/etc material to hold as much oil otherwise.

I have never had a single major brand, bronze bushing fan that was relubed as needed, fail. Heh, I've even had some fans that did fail, seized up, burnt up a PSU that someone brought to me, cleaned them out and relubed, and the fans still work.

I'd rather just have a dual ball bearing fan from a major fan manufacturer. That is what you find in any high end equipment where pennies aren't pinched, with the one exception behing high shock environments that might deform bearings. That tends to be vehicles not stationary objects.

Obviously there is much to disagree about, fans and bearings can't be thoroughly enough discussed in a few forum posts. I suppose I'm back to my main point that I don't see the obsession with seeking a lighter duty fan with marketing pimping it, when the noise vs airflow difference at its normal reduced-RPM speed (keeping the PC components at an acceptable temperature) isn't much.

 

[Mantrid-Drone]

I've been collating some of my fan manufacturer's spec results and no matter how I skew the rating system the best balanced performance Be Quiet (SW3) fan always seem to come out on top by some margin.

The problem with that is in the course of my investigations I'd come across a review of Be Quiet fans by a guy who it read like he knew what he was talking about.

He claimed to be a sound engineer and didn't like them for a rather prosaic but nonetheless apparently good reason. The particular low and high frequency noise they, allegedly, produce, just at the edge of the normally audible range for an adult, is enough to cause not just annoyance but actual hearing damage.

I take such reports with a healthy degree of skepticism but what I do know is that the frequency of the noise made by a fan is something other informed fan reviewers say is something to consider.

But how you do that without buying a particular fan or fans and testing it/them against a known fan you're happy with is not explained. But it leaves you having to decide whether to take such warnings from reviewers seriously.

BTW the Noctua P120 IND fan spec figures suggest it has the best balance of performance with regard to dB(A) of their 120mm range. Those figures are actually better than the claimed 'quiet' fans from over 50% of the other (11) manufacturers fan specs I've been looking into.

It is still far behind the Be Quiet SW3 but overall the same score as the Akasa Viper both only marginally behind the the best Corsair, Thermaltake and temptingly cheap Arctic P12 PST.

If you exclude the cost factor the Arctic does not fair so well but at 75% less than the cost of the Noctua fan and 50% - 60% less than most others you cannot ignore them.

 

[mindless1]

^ If the source claims that the minor drone of a PC fan, based on frequency, will cause hearing damage, you should reject everything he stated as unreliable.

At the same time, yes some frequencies are more objectionable to some people, particularly higher frequencies to younger people with more acute higher frequency hearing. At the same time, higher frequency sound is more directional and more easily mitigated based on fan placement or system placement... and really I'm only talking about there being a significance with nearly worn out ball bearing fans, then they start to whine a bit. They do make the whine at lower RPM too, but it is so much reduced that at average system cooling requirements, it is practically inaudible.

Long story short, it seems a lot like you are just chasing numbers on paper and not putting much real world perspective into this, while also expressing a desire for long lifespan which is more likely with a major brand fan that doesn't have the best noise to airflow ratio, because it is optimized for lifespan rather than that, or construction cost.

It comes back to that old saying, pick two of the three: Performance, lifespan, or cost... except that when it comes to PC fan 3rd party marketing agencies, they feel they can charge more based upon the new internet game, of promotion and well placed reviews.

Even so, a reputable PC parts branded fan, priced at a premium, should be expected to have "enough" service life and focusing on noise isn't unreasonable... but I don't think that's going to matter much unless the system in question has large thermal changes where the fan is likely to spin up to full speed often.

 

[Mantrid-Drone]

TBH I'd be happy with all 140mm Akasa Viper case fans and may be a couple of their 120mm (better SAP) on the cooler. That's simply because of my past and continuing satisfactory experience with them.

But I as said wanted to find something noticeably quieter but of similar performance and not bright yellow.

That's the problem because you're doing little more than guessing from the manufacturers' specs supplied that the performance is going to match or be better, have good build quality, be equally reliable, proven longevity and less dB(A).

Short of the unwanted expense of buying a fan of each make/type/size that seem to fit those criteria and testing them myself what else can you do other than collate and compare manufacturers' specs?

Lets face it most people just buy whichever fans they're told are good (typically Noctua). If they get on with them, and its highly likely they will, keeping their system adequately cool they'll be, quite literally, a fan for life.

But the fact is they are out performed by some other manufacturers' products, often at a much lower price too.

Its the inadequate information that is the problem here really. As I also said earlier if every fan manufacturer supplied AF, SAP, CFM and dB(A) figures for an agreed mid-range RPM speed typical of normal use ie. 1000 rpm you'd be able to compare the fans' 'real life' performance much more reliably.

As it is you can have a Phanteks 2000 rpm/3.59 mm H20/93 CFM 140mm fan producing 39.1dB(A) at that full speed being sold as "a quiet fan". How on earth do you compare that with, for example, the Akasa Viper 140mm: 1600 rpm/3.1 mm H2O/110.61 CFM and 26dB(A)?

 

[Tech Junky]

Buy from Amazon test and return what you don't like.

 

[mindless1]

Today it is far easier to just pick the right case and mod it if necessary, more important than fan choice.

Most people buy the fan and then strap it onto a heatsink with insufficient case noise containment, or on a perforated case wall intake/exhaust grill, and the turbulence from that swamps minor differences in fan free-air noise ratings between different fan models (all else equal, size and RPM).

I present a 25 year old case setup that had a lot going for it to reduce noise.

Rear fan hole cut (there wasn't a fan mount there at all) as well as PSU stamped grill cut out... well, the entire PSU case opening was cut out too, formerly used a proprietary smaller PSU and now ATX.

Thick 1.0 - 1.2mm metal, and plastic clad combined so less vibrations, but not a lot of passive holes to let interior noise escape.



Fans upper limited by 5 cent resistors in series



Make airflow most effective by entering only at the opposite end and flowing through the HDD rack, instead of just tossing a half dozen giant fans everywhere.



No forward facing grill for sound to escape with line of site/ear to the user, rather intake up through the bottom, and perforated case wall, grill area cut out.

 

[Mantrid-Drone]

Buy from Amazon test and return what you don't like.

But I'm good guy.......I can't do that it would be dishonest.

Also I'd have no way of testing CFM or SAP, I'd still be relying on the manufacturers' specs for those figures unless there is some DIY solution.

 

[Tech Junky]

To test things you can get output values using linux / sensors and a vareity of other apps to display vital stats on the system.

CFM usually means good temps under load which would mean using some load producing apps which are easily installable on livecd just like they would be if you installed Linux w/o the need to use a spare drive to do so. As soon as you reboot you're back to your normal system.

If it really just comes down to the color of the fins though just spray them whatever color you want them to be and be done with it.

 

[mindless1]

Keep in mind that decent fans are precision balanced and it is difficult to spray paint curved objects evenly, making it likely you would upset the balance some, and increase noise.

 

[Mantrid-Drone]

Interesting you should mention that - there's a Cooler Master fan type: Blade Master, very difficult to find in the UK now except as part of on eof their cooler/fan packages so I didn't bother to look at its specs too closely. It also uses very similar 'S' shaped to the Akasa fans.

Anyway one of the things that caught my attention was one of the fan blades looked to be etched, deeply with a branding logo.

The first thing I thought was: doesn't that affect the balance? Surely it must effect the efficiency creating disruption in the airflow over that particular blade?

Blade Master 120 | Cooler Master

www.coolermaster.com

 

[Mantrid-Drone]

Interesting you should mention that - there's a Cooler Master fan type: Blade Master, very difficult to find in the UK now except as part of one of their cooler/fan packages so I didn't bother to look at its specs too closely. It also uses very similar 'S' shaped to the Akasa fans.

Anyway one of the things that caught my attention was one of the fan blades looked to be etched, deeply with a branding logo.

The first thing I thought was: doesn't that affect the balance? Surely it must effect the efficiency creating disruption in the airflow over that particular blade?

Blade Master 120 | Cooler Master

www.coolermaster.com

 

[mindless1]

They can just design the mold to have that little bit extra material on that side to offset the text, though something else I've seen on some fans is some type of epoxy or adhesive added to the inner hub to precision balance it.

Yes it would cause a little turbulence, but apparently coolermaster didn't care that much to begin with because they chose a softer plastic with a textured molding... note how the blades aren't perfectly smooth. That can help reduce vibrations a little if used on the frame but not so much on the blades.

 

[Mantrid-Drone]

Just finished collating my ranking results using the manufacturers' spec for their 'best' 120mm PWM balanced performance fan and it is quite interesting. The mistake I made, then corrected, was treating the RPM range as a positive feature so the fans with higher top end and a low bottom end speed were unduly favoured.

The correction simply used the top speed (which is used almost exclusively for all their AF/SAP/dB(A) specs anyway) as a negative ie. the higher the top speed the lower the ranking. This seems fair treatment.

The precise detail and scorings are not really important but I think I've been fair and the results are useful (to me at least) comparisons of the fan manufacturers' figures.

For those interested my top three balanced performance 120mm PWM fans based on the manufacturers' published data are:-

1. Be Quiet Silent Wings 3 HS (BL070).
2. Akasa 12cm Viper (AK-FN059).
3. Noctua NF-F12 IND (NF-F12-IND-PPC-2000 PWM).

Now to do the 140mm PWM ones.

Just finished doing that:-

1. Akasa 14cm Viper (AK-FN063).
2. Scythe Kaze F Round (KF1425FD18-P).
3. = Be Quiet Silent Wings 3 HS (BL071).
Thermaltake Toughfan 14 (CL-F18-PL14BL-A)

The Akasa walked away with it in the 140mm tests - miles ahead of the nearest opposite with, surprisingly, the Scythe coming second mainly due to its reasonable pricing and relatively low dB(A) rating at top speed.

The Noctua FN-A14C came nowhere even when the high cost was factored out.

The Thermaltake costs more in the UK than any of the fans compared but when that is factored out it would move up to clear second place relegating the Scythe to third and the Be Quiet out of the top three.