[SOLVED] NF-F12's or NF-A12x25's for AIO cooling?

[TM1172]

Looking at changing out some QL120's for new fan options on a 3080ti hybrid cooler. My three choices are either the Noctua NF-F12, NF-A12x25, or Arctic P12 PWM PST CO. I ran into some confusing information on Noctua's website about performance.

Noctua itself recommends the NF-F12 for radiator cooling (at least in the chromax and standard lines - the redux line recommended the P12).

The website then included this graphic:

From here, again, looks like the NF-F12 is solid.

Where I got confused was the next graphic the website provided:

This graph appears to indicate that the NF-A12x25 would actually perform better as a rad fan than the NF-F12, getting better static pressure than the NF-F12 at the average impedance of a "typical" radiator. The original page also says: "For best results in all types of application, Noctua recommends the NF-A12x25 ... Please see this page for a detailed performance comparison of NF-A12x25, NF-F12 and NF-S12A as well as further explanations why the NF-A12x25 provides best results in real-world applications, despite being neither a pressure- nor an airflow-optimised model." ("this page" link routes to 2nd graphic above)

Am I reading this wrong? I assume Noctua knows what it's talking about when it recommends the NF-F12, and the NF-F12 does get higher overall static pressure (2.61mm H2O vs 2.34mm for the NF-A12x25). Should I be paying attention to the P/Q curve if my only intended application is a radiator? Should I even be concerned about the "critical midsection" (their words) of the performance curve? Is the performance difference because of the noise-normalized fan speeds?

Then, of course, there's the budget elephant in the room, with the Arctic P12 PWM PST CO claiming to match pretty much every metric that Noctua measured - tech specs have the same nominal noise level (.3 sones, which looks like about 22.6 dBA), and about the same static pressure as the NF-12x25, claiming 2.2mm H20. Plus it's more than 50% cheaper per fan. I've had decent experiences with Arctic P14's, with my only complaint being a little harmonic hum around 1400-1450 rpm.

I have pretty sensitive hearing, enough that the hum was noticeable but not frustrating. I've had no experience at all with Noctua fans, other than reading reviews and specs. Do they really meet their claims on noise production?

If the Noctua's are legitimately better fans than the Arctics, I'm willing to drop the money on them. Performance and noise hold equal importance for me in this scenario. Price isn't so much a factor.

Probably overthinking this one, but I live in a place where "expedited delivery" basically doesn't exist, military mail is the only way to get things shipped over, and I really don't want to deal with returning items.

 

[geofelt]

Since you already have some fans, why don't you first try them out and see how you like them?

 

[Phaaze88]

Yeah, why not just try it if you got room to do it and see what happens? My experience with using the NF-F12 and their IPPC 3000 version wasn't a positive one.

Contrary to what Noctua shows on their own website, the F12s are not good radiator fans.
They don't move much air below 1000rpm, and above that, they start to become annoying audible, preventing me from getting the most out of them.
The IPPC 3000 definitely packs the needed punch, but it's basically an F12 with a greater rpm range, and thus can get even more audible, leading to me running them no faster than the regular F12 is capable of... and that's a waste.


Actually, my experience with NF-F12 lines up TechPowerUp's review of it: https://www.techpowerup.com/review/noctua-nf-f12-pwm-chromax/

 

[TM1172]

Sorry if there was some confusion, I don’t have any of the fans I mentioned above. I only have the QL120’s I want to move away from. There’s no ability to try other fans.

 

[Phaaze88]

AH, sorry.

My experience with NF-F12s says to steer away from them.
I have not tested Artic P12s.
I have not tested NF-A12x25s, but I do have Phanteks T30-120s. The 2 fans are supposedly really close to one another in performance to noise, except with the T30 having another 1000rpm on top... aren't too many folks who'd normally run them that high though.

 

[TM1172]

AH, sorry.

My experience with NF-F12s says to steer away from them.
I have not tested Artic P12s.
I have not tested NF-A12x25s, but I do have Phanteks T30-120s. The 2 fans are supposedly really close to one another in performance to noise, except with the T30 having another 1000rpm on top... aren't too many folks who'd normally run them that high though.

Thanks, that is good information to have - I think the F12’s are out of the running. Kinda puzzling with so many outstanding specs on paper and Noctua’s unequivocal recommendation, but that’s why I posted here.

 

[Paperdoc]

Maybe this will help a little to understand the Noctua details. The PRESSURE spec for a fan is the MAX pressure it can deal with just before it stops pushing ANY air flow. Most fan makers do not provide the details you found in that second graph, so users are commonly forced to ASSUME the relationship between air flow and back pressure is simply linear, when it really is not. What the second graph makes clear is that at large backpressure with almost NO air flowing, the NF-F12 can still push more air through than the NF-A12x25 can. In that scenario you would not get GOOD cooling, but you would get some at least. But in the mid-range of backpressures the NF-A12x25 is better because more cooling air is flowing. Further, the graph says that this mid-range area is where most rad system operate.

 

[TM1172]

Maybe this will help a little to understand the Noctua details. The PRESSURE spec for a fan is the MAX pressure it can deal with just before it stops pushing ANY air flow. Most fan makers do not provide the details you found in that second graph, so users are commonly forced to ASSUME the relationship between air flow and back pressure is simply linear, when it really is not. What the second graph makes clear is that at large backpressure with almost NO air flowing, the NF-F12 can still push more air through than the NF-A12x25 can. In that scenario you would not get GOOD cooling, but you would get some at least. But in the mid-range of backpressures the NF-A12x25 is better because more cooling air is flowing. Further, the graph says that this mid-range area is where most rad system operate.

Thank you, this is an excellent explanation. I wish more companies included information like the second graph, but I can see how it would potentially get misconstrued. I wound up buying a pair of the NF-A12x25’s, really looking forward to installing them.

 

[Karadjgne]

Throw out all airflow charts, Noctua recommendations, he said-she said suppositions, it's all useless info.

Has only half to do with the fans. The other half belonging entirely with the rad in question. Contrary to popular belief all rads are not the same. Different amount of fins, different thickness, depth, even different designs, offsets, etc. Noctua tested their graph on just one rad, unless you use the same rad, the numbers will be totally different.

The fan needs to match the rads needs. Some rads need high speed, high sp fans and only get better with more. Some rads are better suited to higher airflow fans at lower speeds and their performance curve tops out very shallow, where more equates to worse.

The A12x25 does better on the low density, higher airflow rads at lower rpms, the F12 does better on higher density rads because of its forced airflow design ( almost 90° straight flow) vs the A12's wider cone. The straight flow doesn't suffer from nearly as much turbulence, it goes straight through the fins, wider cones bounce the air at an angle to the fins. Wider the cone, worse the fan for rads at higher rpm.

In their hay-day, the 2 single best fans by a long shot for an aio rad were the F12 and Scythe Gentle Typhoon. There were fans with higher rpm, higher sp, higher cfm, but those 2 fans hit about the exact right balance of all three vrs those rads and had tight enough cones to maximize throughput and minimize backpressure. It's the single biggest reason most rgb/argb fans suck compared to non lighted fans, they simply are not optimized for the rad.

Radiator Review Round Up 2016 - ExtremeRigs.net

360 Radiator Review - hardware labs, koolance, ek, xspc, aquacomputer, watercool, water cooling, black ice, liquid cooling, phobya, alphacool, coolgate

www.xtremerigs.net

A good read to understand more about the relationship of fans to rads and how different rpm, cfm, sp affects different rads, differently.

 

[TM1172]

Throw out all airflow charts, Noctua recommendations, he said-she said suppositions, it's all useless info.

Has only half to do with the fans. The other half belonging entirely with the rad in question. Contrary to popular belief all rads are not the same. Different amount of fins, different thickness, depth, even different designs, offsets, etc. Noctua tested their graph on just one rad, unless you use the same rad, the numbers will be totally different.

The fan needs to match the rads needs. Some rads need high speed, high sp fans and only get better with more. Some rads are better suited to higher airflow fans at lower speeds and their performance curve tops out very shallow, where more equates to worse.

The A12x25 does better on the low density, higher airflow rads at lower rpms, the F12 does better on higher density rads because of its forced airflow design ( almost 90° straight flow) vs the A12's wider cone. The straight flow doesn't suffer from nearly as much turbulence, it goes straight through the fins, wider cones bounce the air at an angle to the fins. Wider the cone, worse the fan for rads at higher rpm.

In their hay-day, the 2 single best fans by a long shot for an aio rad were the F12 and Scythe Gentle Typhoon. There were fans with higher rpm, higher sp, higher cfm, but those 2 fans hit about the exact right balance of all three vrs those rads and had tight enough cones to maximize throughput and minimize backpressure. It's the single biggest reason most rgb/argb fans suck compared to non lighted fans, they simply are not optimized for the rad.

Radiator Review Round Up 2016 - ExtremeRigs.net

360 Radiator Review - hardware labs, koolance, ek, xspc, aquacomputer, watercool, water cooling, black ice, liquid cooling, phobya, alphacool, coolgate

www.xtremerigs.net

A good read to understand more about the relationship of fans to rads and how different rpm, cfm, sp affects different rads, differently.

Thanks for this resource and your insight. EVGA doesn’t publish the FPI of their hybrid GPU radiators that I’ve been able to find (I plan to measure tonight once the offspring have gone to bed) but just visually it doesn’t look like a crazy dense radiator. Appreciate the response.

 

[Paperdoc]

Karadjgne makes very good points. Certainly all rads are different, and we really do not get any info to help on this. The one relevant piece of info from Noctua in that second graph is the label that the central oval area covers most "typical" rads and CPU finned heat exchangers. Of course that does not guarantee it relates to all. I suspect further that many AIO complete system designs are NOT optimal because the makers do not attempt to match exactly the fans supplied with the rad. For the most part they are fan makers who use the fans they already sell. This factor of matching becomes nearly impossible when you are building your own liquid-cooled system from separate components. So "fully optimal" won't happen, but you can get a really good system to exceed your needs with careful attention to detail.

 

[Karadjgne]

Hey, we are Corsair and we just made 10milluon 120mm SP's, and they won't sell. Oh, I know, let's throw them on the AIO's!