SPCR

Hello Devon,

Since the air volume method has changed, it might be very helpful to retest the best fans from the previous group, so they can be directly compared?

Many thanks for this review.
Currently the "One meter" mp3 links for Global Win 120mm, Fander FX-120W and Globe Fan S1202512L-3M fans erroneously point to one-feet mp3s.

Nice review as usual, thanks SPCR!
IMHO the SilenX fan blade and hub design shows promise, just the shoddy frame seems to hold it back.
May I suggest that SPCR thinks of recommendations for fan packaging?
It's possible that many bearings get more or less damaged during shipping, thus ruining the fan for SPCR purposes.
All the fans I had (that includes my current Nexii) came in el-cheapo packaging, that could do little to prevent bearings damage.

I'm surprised SPCR had trouble getting hold of a new D12SL-12,jab-tech.com & nexfan.com carry them (cheap!).

So does SVC, Case-mod, Newegg, and just about everyone else...

In my experience 38mm SilenX fans are much quieter than 25mm versions.

Comparing mp3 recordings in a multitrack software I'm amazed to see how much ratings differ between comparing 9V and 12V recordings : the best fans when used at 9V aren't necessarily the best fans used at 12V.

I guess my initial reaction is "Err-um?"

Need a D12SL-12 or something? Never even heard of anyone recommending or using the M version.

I like the potential of the enlobal bearing, seems similar to Sunons MagLev bearing, if not the same. The fans are cheaper than Scythe S-Flex's at newegg, I might have to try one out when I have the urge to buy something. Great Review!

Typo patrol on page 2:

Yate Loon manufacturers a special bright orange model for Nexus

Nice review!

Nice review - thanks!

In addition to the above (can't find YLs? they're all over the place), I was a bit curious about the following...

It's also worth pointing out that SilenX has made an effort to force as much airflow as possible out of the Ixtrema Pro. The center hub is unusually small, apparently "for reduced air turbulence" and to allow slightly more air to pass through the frame. As a result, the fan blades appear unusually long and scooped. SilenX' marketing material claims that this gives the fan better pressure characteristics, but how much so is difficult to say. Our airflow measurements did not uncover anything particularly special about the Ixtrema Pro's airflow characteristics... but we are unable to measure the effects of pressure.

...this is a bit puzzling if you look at your data, which shows that the SilenX fan had the lowest measured RPM to achieve 25cfm airflow. By quite a large margin in some cases (~200 rpm)...

SilenX - 820
Fander - 840
NB XL1 - 870
Yate L - 880
EnerM - 950
Global - 970
Arctic - 970
AC/GL - 1020 (!)

...that's quite a range, and also doesn't seem to jibe with your on-going "rpm = airflow" stance. I don't know if your new testing method is producing exactly 'correct' cfm numbers or not, but I think it's as good a method as any and should be useful for comparing fans relative to each other (for the 'open air' case, at least).

Based on your latest testing methods, this would mean that people should be comparing noise levels of a SilenX fan at 7v (for example) to a Nexus fan at ~9v (8.something for 25 cfm). I'm not implying that the SilenX wins that particular comparison (I think the Nexus still wins), but I think it's worth noting. Obviously your "noise at 25cfm" recordings in the older reviews are geared towards that end, but not presented in this review for some reason (I assume you left them out until you can correct the sound files in the old reviews to reflect the new cfm ratings).

You did mention a few fans being less efficient (you'd have to run the Arctic Cooling and Marathon fans at full speed to get 25cfm), so I guess I was just surprised that you seemed to not notice that the SilenX was more efficient than the rest of the pack. It's a shame they used such cheap materials and charge so much, but it seems that the small hub design and maybe the blade design (that comes up onto the hub) helps get more air through for any given rpm.

That's Canada, all right. Dozens of lumberjacks roaming the SPCR lab and touching all those delicate, flimsy fans with their hands the size of skillets and fingers the size of sausages.
Some time ago I had a chance to compare an S-Flex 1200 against my Fanders and I was unable to tell the difference. My oldest Fander (92mm) still runs fine, and both my 120mm versions also show no sign of wear or damage. They're all almost one year old.
Since Fanders are (strangely enough) popular around Polish dB geeks there's always lots of talk on them and the issue of faulty bearings is, naturally, one of the subjects. There was talk of how early Fanders were better than the newer revisions but it's impossible to tell, really. Yeah, hard to say for an ordinary consumer like myself, but it seems to me there's just as many faulty Fanders as AC's or S-Flex'es or Noctuas around here. From time to time a clicking unit will show up and it will be replaced by the manufacturer. No big deal.
Anyway, I'm happy with all my Fanders and see no reason to change them for anything else.

Oh, BTW - another great review!

As mentioned above, I would be interesting in seeing the updated numbers for the Nexus and Noctua 120mm fans to make the comparison easy.

I'm personally on the lookout for the quietest fan that produces roughly equivalent airflow to my TriCool (on low). Its the fan that came with my Antec Solo. According to an Antec pdf I found, its about 39CFM at 25db. Obviously I'd prefer 40CFM at 22db

K9-Cop wrote:As mentioned above, I would be interesting in seeing the updated numbers for the Nexus and Noctua 120mm fans to make the comparison easy.

I'm personally on the lookout for the quietest fan that produces roughly equivalent airflow to my TriCool (on low). Its the fan that came with my Antec Solo. According to an Antec pdf I found, its about 39CFM at 25db. Obviously I'd prefer 40CFM at 22db

SPCR's measurements (for comparitive purposes) put the Antec Tri-cool on low setting (5v) at 20dbA@1m with a newly calculated airflow of 21 cfm. So if those numbers are all correct, most of the fans tested could be substituted at some voltage setting - take your pick .

When I first came to SPCR, the only fans at the Quiet PC shop were Acoustifan. (http://www.quietpc.com/nz/casefans.html) This web page lists two Acoustifan product lines, the transparent 'C' series (tested by SPCR) and the opaque 'Dustproof' series. From the specs on this page, a silencer would be mad to go with the 'C' series rather than the 'Dustproof' series. (For example, 120mm, 7.5V, similar airflows, C is 26.3dB, Dustproof is 15.6dB.) (The difference is not just the material - they run at different RPMs.)

So I was at first excited that an Acoustifan would be tested, but then disappointed to see it was the transparent 'C' series, which I'd never have bought in the first place.

I never did try the Dustproof. I first bought a ThermalTake fan branded as being quiet, and was very disapointed (it was just a noisy fan with a speed controller). Next I got a Nexus mail-order from USA, which I still use. (Single fan system, never runs above 60% fan speed.)

Nexus are now available in New Zealand, and cheaper than the Dustproof, so this is no longer such a big issue for me. I would still be curious to see the Dustproof tested, however.

jaganath wrote:I'm surprised SPCR had trouble getting hold of a new D12SL-12,jab-tech.com & nexfan.com carry them (cheap!).

This is more an issue of time than supply. I didn't want to delay the article yet another two weeks while we waited for new samples to arrive and be tested.

@Spanki:

There's a couple of hidden issues here I think. First and foremost is the issue of back pressure in our test box. Different fans deal with back pressure differently; some drop significantly in speed, while others barely budge. I think I mentioned somewhere that the worst-case drop was ~200 RPM at full speed.

However, we do not measure RPM with the fan installed in the airflow box. This was done to counter potential issues with people asking about bad samples when our measured RPM conflicts with the rated RPM, which has occurred in the past. This means that the fans that drop the most under pressure have an unfair disadvantage in a straightforward RPM:CFM analysis. As a result, I had to allow for a rather wide margin of error when I did my analysis, especially for fans that I knew were affected by pressure.

The second issue is that my evaluation was based on more than just the data at 25 CFM. My comparisons were done based on average RPM:CFM ratios across all measured points, with a slight emphasis on CFM at ~1,000 RPM, since this is one of the points we looked at when we re-evaluated our test method. This identified the Enermax and Arctic fan as having low airflow, and, looking at the data again, I probably should have flagged the Global Win as well.

While the SilenX did do well in my analysis, it didn't stand out the way the Arctic Cooling and Enermax did. Given the large margin of error I was allowing myself and the problems that arose the last time I recommended a fan on the basis of high airflow, I didn't feel confident that the data supported concluding that the Ixtrema was necessarily higher airflow.

That said, I did get the impression that it might have been a little better, which I why I made the comment about pressure in the first place, and recommended experimenting. I'm not confident that our data is conclusive, but I do think that further testing and analysis might bring a difference to light. Of course, the issue then is whether the difference is big enough to make a practical difference...

To all who are waiting for updated data to past reviews: Keep waiting, it's coming.

Devonavar wrote:Of course, the issue then is whether the difference is big enough to make a practical difference...

This brings us right back to the crux of the matter, which is that all of our care on trying to get "correct" CFM readings may be for naught. We've discussed this at length in past discussions, the gist of my argument being that it takes a very significant change in airflow to effect even a tiny change in monitored temperatures in a PC. I refer to both case and HS fans. In general, we're looking at >20% airflow difference to effect something like a 1-2C change. The cost in turbulence noise alone is generally much more significant than the improvement in cooling, which is much better achieved with either a better heatsink (that does well with low airflow) or reduced case impedance.

OTOH, regardless of airflow, there's no substitute for a fan that has a nice acoustic signature.

Spanki wrote: ...this is a bit puzzling if you look at your data, which shows that the SilenX fan had the lowest measured RPM to achieve 25cfm airflow. By quite a large margin in some cases (~200 rpm)...

SilenX - 820
Fander - 840
NB XL1 - 870
Yate L - 880

Not that much, tho... and 20rpm out of 840rpm is 2.4% variance-- is that within production variance of fans? I have no idea, personally.

About horizontally mounted fans:
- why are sleeve bearings not so good?
- what is the best fan for horizontal mounting?

Thanks.

Good review.

Interesting reading about the Fander.... the article says it is made by Xinruilian, a company I hardly ever come across. I have two 80mm sleeve bearing fans made by Xinruilian (they actually say "Xinruilian" on them), one I bought used from a junk pile (presumably a case fan) and the other came with an old Landmark case I still own.

Both are odd asymetrical frame designs, but they are both amazingly quiet at 5V despite their age and the amount of dust caked on them. Even though they have the same part number one spins noticably faster, making it "noticable" at 7V, but the other one is still very quiet at that setting. If they weren't so fast running they would be among the smoothest, quietest fans I own.

It's really too bad that your Fander samples developed a chuffing noise. I had a couple 120mm Coolermasters do that to me while they were sitting around. Maybe see if running them for a while or "working" the bearing around a bit by hand helps? One of mine improved after doing that and the chuffing hasn't returned (yet).

Devonavar wrote:...However, we do not measure RPM with the fan installed in the airflow box. This was done to counter potential issues with people asking about bad samples when our measured RPM conflicts with the rated RPM, which has occurred in the past. This means that the fans that drop the most under pressure have an unfair disadvantage in a straightforward RPM:CFM analysis. As a result, I had to allow for a rather wide margin of error when I did my analysis, especially for fans that I knew were affected by pressure.

...

While the SilenX did do well in my analysis, it didn't stand out the way the Arctic Cooling and Enermax did. Given the large margin of error I was allowing myself and the problems that arose the last time I recommended a fan on the basis of high airflow, I didn't feel confident that the data supported concluding that the Ixtrema was necessarily higher airflow.

Thanks for the clarification. I think I'll have to re-read that again, but it does explain a few things.

Devonavar wrote:Of course, the issue then is whether the difference is big enough to make a practical difference...

@ MikeC as well...

I think we're all after the same thing, ultimately, which is to get the best 'cooling' fan, for some noise level/quality that we're comfortable with - regardless of airflow numbers, rpm numbers or dbA numbers.

Where we might (might not) disagree is in the amount of cooling needed for some task (my situation may not be the same as yours) or and the noise level we're willing to accept to get there, but that doesn't change the goal.

If fan A falls short of doing the job, then it doesn't matter HOW quiet it is - it's useless.

If fan B does the job, but is X amount louder and...

If fan C does the job and is also X amount louder, but has a more pleasing/bearable sound quality...

then I choose fan C - and I think you do too.

If fan A works great as a case fan, but falls apart with some impedence in front of it (radiator), then I need to figure out what fan/plan B is - that's all .

utimately, everyone is looking for a free lunch: lots of airflow with minimal noise. TANSTAAFL-there ain't no such thing as a free lunch. If you have a hot computer, you will have to make a trade-off between noise and cooling, and there is no magic fan that can make that compromise go away. So SPCR can flag up the fans that they think produce most CFM dB, but there is no getting round the laws of physics I'm afraid.

jaganath wrote:utimately, everyone is looking for a free lunch: lots of airflow with minimal noise. TANSTAAFL-there ain't no such thing as a free lunch. If you have a hot computer, you will have to make a trade-off between noise and cooling, and there is no magic fan that can make that compromise go away. So SPCR can flag up the fans that they think produce most CFM dB, but there is no getting round the laws of physics I'm afraid.

Yeah, I assumed all that went without saying .

Well, it does need saying, because we still get too many people who take specs like 74CFM, 11dB at face value.

Is anyone trying to say Tt's marketing dept is not to be trusted?

Fair enough... if you hear what they're smoking, let me know - I could use a good laugh.

Mike - If possible please send 120mm models our way at my cost to check them. I'm very curious what happened as postal packages handling is sometimes brutal. I'll also send You 3-4 brand new ones to compare with old pcs. I've sent You my address on PM.

This test had the Yate Loon SM - NOT the SL . Few places carry the SM for some reason. I tend to like fans that can really move some air on full speed yet be reasonably quiet when slowed down with a fan control of some type.

Copy of messages sent to author:

Hello Devon,

I also found this 40 page PDF that covers the Global Win fans in depth:
http://www.globalwin.com.tw/products/nc ... nglish.pdf

--------------earlier message-----------------

http://www.jab-tech.com and sometimes http://www.petrastechshop.com have the Yate Loon SM model.

.bh.

thanks for another great article. my only disappointment was that a new D12-SL12 wasn't reviewed so that I could compare it directly to the X3 more expensive nexus.