26 120mm Fans Tested and Compared

[Vapor]

Good, bad, and ugly update.

Just playing with the new stuff now...accidentally fried my DMM, so I only have 12V and ~7V to test for now Still fun to be back and testing fans though

New location from last time, so very, very different noise personality....old results will not be directly comparable....I WILL retest many of the fans however

Due to a WCG-inspired processor swap, the CPU temp comparison will also not be directly comparable (went from a Celeron D...GREAT for this purpose, not for WCG...to a Conroe). For various reasons actually....I'm considering flat out dropping the CPU temp test. Results were largely consistent with restricted CFM test and this test took freaking FOREVER (~1hr a fan for the first pass [some ~25% of the time, another 1hr 2nd pass was needed to verify/rectify data]). Near the end I could pull a test time down to ~30min after I got in a groove, but that was only because the Celly D was so forgiving with going passive on the TTBT. And I've also repurposed the thermal probe I was using for ambient temps....and the Conroe isn't nearly as flexible as the Celly D was for stability at high temps. So the long and short of it is that the CPU tests future are in jeopardy. Talk me in/out of it....I'm on the fence, leaning toward ditching them.

I've been playing with my three new fans, the Sharkoon, the FN121, the U1A, and the SHE Tri-blade.

1) the San Ace is just incredible, freaking incredible. At ~7V it preliminarily pushes more air quieter than the Sharkoon does at 12V. At 12V, it pushes ~115CFM, not the 103 that it's rated. I can't get it to drop to ambient noise levels....but that's okay considering it's a 115CFM fan, lol. Compared to the restested U1A, it pushes 1CFM more at 12V and at 1dBA lower. At 7V, it's 7CFM ahead and 1.5dBA quieter. FWIW, the U1A is getting extremely similar results to what it had before...and it was an 'All-Star' before too.

The Panaflo LH is old school to the max....old school noise, old school design....and old school performance :-/ At 12V, it's around 5.5dBA louder than the 7V San Ace, and pushes the same amount of air. Compared to the Sharkoon, it pushes ~1.5CFM less at 12V and 3.5dBA louder. The Sharkoon also fell victim to an approaching thunderstorm that increased ambients a lot ...so that dBA data may be on the high side.

The Comair is the worst rated fan I have tested to date...they say it's a 102CFM fan, but I'm getting 129CFM out of it....but it's also loud. Quickly compared to the Delta SHE Tri-blade, it was a little louder.

The Silverstone FN121 is still putting up identical numbers from before....the Yate was in a case and I was too lazy to remove it, so no retest numbers on that yet. The Sharkoon is getting ever-so-slightly worse dBA numbers (for now...thunderstorm and one piece of electrical noise may have effected it)....but still better at 7V than the FN121 is at 12V.

Anyway, so yeah, a thunderstorm has rolled in so noise testing is useless....and CFM results compared to last time haven't changed even a tad on any of the fans.

[CPLB]

Thanks for the update Vapor, I'm running 2 of those San Ace fans and they're great.

[Spanki]

Which Sharkoon is that this time?

[Vapor]

Same exact fan....Silent Eagle 2000 "Golfball"

[Spanki]

Re-test.. ahh, cool.

[Vapor]

Ah, yeah...sorry for any confusion. The Sharkoon, FN121, U1A and SHE are all retests.

[Spanki]

the San Ace is just incredible, freaking incredible. At ~7V it preliminarily pushes more air quieter than the Sharkoon does at 12V

..but I take it that it's still somewhat loud at ~5v? Sounds like it may not have enough lower range.

Edit: hmm... just referenced your original chart again and it looks like the U1A gets down to YL 12v dbA range at 5v, so maybe there's some hope, if it's comming up quieter than the U1A at 7v.

[Vapor]

At the very bottom of its working voltage range (4ishV), it was still audible...not loud, but if it were the only thing in the room, you'd notice it.

Can't get better/official numbers until I get a new DMM :-/ (or a new fuse, but I think it's time I get a new DMM anyway)

EDIT: just realized that made no sense above....lol. Anyway, with the thunderstorm, audio tests are useless....and if I wanted to plot the U1A vs. the San Ace, I'll need to get my DMM working anyway.

[serialk11r]

Ouch your multimeter broke...I hope it wasn't a good one i have a 10 dollar one

[Spanki]

No problem . On the cpu testing... that was one of the features that I really liked about your review. I'm not sure that the radiator numbers line up exactly (must be difference in the amount of impedence)...

If you look at the all-stars, adjusted db on radiator, at the 27-28db range (where the low speed fans top out), the FN121 is measuring just under 30cfm, the bottom end of the U1A is roughly the same (but a bit louder), and the YL is pushing ~4cfm more than either of those. But then when you look at the final chart (cooling processor, adjusted dba) at that same 27-28dba range, the FN121 is doing a better job on the heatsink than the YL (which was doing better on the radiator). It's only a couple degrees, but the radiator chart kinda paints a different picture, as it were .

Maybe if the heatsink test is a pain, there's some other way to do a second test on the radiator with more/less impedence?

...of course the more interesting fans (to me, at least) are the medium speed ones, which seem to perform well in either case at that dbA level (with some breathing room on either side of that speed).

[Vapor]

Maybe if the heatsink test is a pain, there's some other way to do a second test on the radiator with more/less impedence?

I like this idea...I'll see what I can rig up for a rig with more impedance (I have a mesh grate I'd like to use for a low-restriction setup).

TBH, I really liked the CPU tests as well...shows just what the extra noise can gain (in that scenario at least...). It did mostly corroborate, but you're right, a few key parts were different. One key thing about it is that it took fan geometry into account....especially hub size and things like that. No doubt it's useful the more I think about it, but damn it's a PITA....lol.

I'll sleep on it.

Heck, I have a spare TTBT, I could do a CFM test on that too (EDIT: actually, logistically, I can't....so nvm).

[serialk11r]

I have an idea. Instead of looking at the temps on the CPU, strap the heatsink on and measure the airflow. That way you can see which fans perform well with restriction, and without the hassle (because obviously, more airflow=better performance).

[Vapor]

See, the thing about the heatsink test was to show deficiencies in certain designs.....yeah, some fans may get a lot of airflow, but if the hub is too big or the airflow is concentrated to only certain parts of the HSF due to funky blades, then it may not cool as well as a diffuse pattern of airflow or a small-hubbed fan.

Again, I'll sleep on it.....I may build yet another dedicated rig for this now that I think about it more....lol.

The things we do for that extra 1%

[Spanki]

TBH, I really liked the CPU tests as well...shows just what the extra noise can gain (in that scenario at least...). It did mostly corroborate, but you're right, a few key parts were different. One key thing about it is that it took fan geometry into account....especially hub size and things like that. No doubt it's useful the more I think about it, but damn it's a PITA....lol.

Good points as well. Blade and Hub geometry are kinda actually the big question here. Some folks claim that pretty much any fan turning at X rpm wil move the same amount of air as any other same-size fan spinning at X rpm... even though thier own testing data doesn't really support that claim.

I "don't know" if there's any truth to that claim or not, so I'm interested in finding out . Some fans have really large hubs, reducing the blade size... some have really small hubs (like the Scythe 100mm fans), allowing for larger blade sizes and then some (like the SilenX) have blade wrapped up onto the hub. Then you get to the 'scooped' fan blades, like used on a Cooler Master Hyper Tx, with a rounded/knob hub-cap, then you have your Golfball blades (with a fairly large hub), etc.

Is that all 'just' marketing gimmicks? Or does it make some measureable difference (ie. if it makes it quieter or at least more pleasing, maybe I can run it faster for more cfm compared to some other fan, etc.).

[Spanki]

The things we do for that extra 1%

...true, but it's another data-point, that could be the (hypothetical) difference in:

"Fan A sounds better than Fan B AND it can make this thing 4C cooler in the process"

vs.

"Fan A sounds better than Fan B, but probably cools a little worse, given that data on the radiator test".

...more data points help make better informed buying decisions .

[Vapor]

Okay, I'm keeping the CPU tests, but only going to do them at 7,9, and 12V.

Also of note is that today I started playing with my SilenX's and a few 140mm fans. The SilenX's aren't too bad....I've been having problems with inaudible (higher frequency than I can hear) ambient noise from SOME source that wasn't present last night that I have yet to find so no 'official' dBA numbers at all on these.

The 120x120x38 that claims 90CFM pushes 60CFM in my tests, while the 120x120x25 that claims 72CFM pushes 55.9ishCFM in my tests. They seem pretty quiet and undervolt fairly well (no extraneous noise). BUT, the motors get bloody hot....much hotter than even the monster Delta 190CFM EHE. They're aesthetically pleasing, though very light....kinda feel cheap because of the lack of weight. They are absolutely not magic fans.

My 140mm experience has been lackluster so far....I just can't get any of the Aerocools or even the Yate to pull their claimed CFM (the Yate being the furthest off). The Aerocools fall right at the +-10% they use as padding....so that's okay....but the Yate is puzzling me. The 14SM-12 claims 62CFM when I can only get ~43CFM out of it. It is exceedingly quiet....much quieter than the claimed 29dB (getting around 26.5 in the screwy ambient....). I'm wondering if I have a mislabeled and rare 14SL-12.

EDIT2: On-board RPM reading says 1100RPM on a sensor that over-reads by ~6% on every other fan I've tested (IDK how though....isn't that one of the simplest circuits?) Looks like I have a D14SL-12....ugh, lol

[Ar3s]

well you will see the exact same thing with the D12SM. i don't know why just my opinion but they don't push as much as their claim. 70cfm please....

[Vapor]

I've also doubted the 70CFM @ 1650RPM (esp considering the D12SL does 48CFM@1350 with the same geometry [I think])....but 62@1400 (on a 140mm) is perfectly fine and expected. It just doesn't spin much faster than 1035RPM or so....and is frankly one of the nicest sounding fans @12V I've ever (not) heard. Has to be a mislabled D14SL-12....which is nice and all, but the SM variant is what I wanted to test....looks like I'll go fan hunting again.

[Vapor]

Just got my PTS package today....Scythe Minebeas, Nexus 120mm, a couple Sunons, their version of the YL D12SL-12 clear blue, and BOTH the YL D12SH-12 and D12SM-12. The SM has been OOS everywhere, even PTS, but I talked to Alex and he had an extra and we agreed to (because PayPal is a PITA) order an SH and then swap with an SM at his side. Well Quoc included both and handwrote a nice note

In other news, I'm a dumbass.

Yesterday the USPS guy left one of those "I came, I rang, You didn't sign" notes on my door saying that the Jab-Tech package needed to be signed for....so I go to the Postal Office today at 1:30 and they're closed

In yet other pieces of news....the YL SM at 12/7V does 55.3/35.8CFM and the SH does 76.6/49.3CFM. My ear and unofficial numbers say they're nice and quiet

Oh, and the SM has the SL's blade geometry (I think...my SL is in a case and spinning right now, so I can't 100% verify that)...the SH has a different one.

Oh, and the Nexus 120mm D12SL-12 is INCREDIBLE in my prelim testings...

Oh, and on the Scythe Minebea's....on all three for 12V, I'm getting within .1CFM accuracy to their claimed CFM. Kudos to Scythe for having realistic CFM ratings! Not that mine are super-scientifically accurate, but .1CFM difference (at most) on all three fans....that says something, IMO. The dBA ratings, that's anothers story, BUT....very different environements as well.

[Ar3s]

yeah i checked out jab-tech large image view. I have SL & SM i looked they look exactly the same. hahaha good to know that the plastic bag test still works for me. OH then i should get SH instead since they will be having the SM ratings instead then.

Scythe Minebea's are by NMB Minebea. Scythe didn't even relabel the fans NMB deserves the credit there if thats true. Amazing stuff and thanks to scythe for providing the quieter smaller versions for computer users.

on a dumbass note from me. i got myself a scythe kamabay. the stock fan that comes with it spins at 800rpm. the kamabay has a mesh and a filter. so there is NO air coming thru. this is the most useless fan in the whole casing !

[skane]

what do you mean with the nexus is incredible? higher cfm than expected, quieter than expected,...?

[Vapor]

what do you mean with the nexus is incredible? higher cfm than expected, quieter than expected,...?

higher CFM, very, very little noise and what noise there is has to be the least obtrusive sound a a fan has ever made.....undervolts well too

[Spanki]

Run across any horizontal mountable (non-sleeve bearing) fans yet? I'm starting the hunt for something quiet to replace my Antec Tri-Cool.

[serialk11r]

Run across any horizontal mountable (non-sleeve bearing) fans yet? I'm starting the hunt for something quiet to replace my Antec Tri-Cool.

If the enermax marathons turn out to be good they might be an option

[Spanki]

If the enermax marathons turn out to be good they might be an option

Yeah, I was thinking the same thing... but I thought I read (on SPCR?) that the Enlobal bearing was subject to the same sort of issues as sleeve bearings when not running vertical.