Airflow vs static pressure fans. Are they a scam?

[The red spirit]

*that is moving the goalpost.*
Don't censure someone else not recognising your own fault. Radiator tests should be the industry standard, not the other way around.

Why? Most people don't use them. I built quite a bit of systems and only once with radiator.

 

[mtcn77]

Why? Most people don't use them. I built quite a bit of systems and only once with radiator.

You aren't doing that for water cooling, rather to demarcate the fan's properties.

 

[The red spirit]

You aren't doing that for water cooling, rather to demarcate the fan's properties.

Fan properties wouldn't be the same if you weren't water cooling.

 

[VSG]

Fan properties wouldn't be the same if you weren't water cooling.

Exactly. This is why I mention as much in my fan reviews which are part of the water cooling review category here

 

[The red spirit]

Exactly. This is why I mention as much in my fan reviews which are part of the water cooling review category here

I think that the most practical and easiest to implement change would be to add CFM per dBA graph. No need to collect new data, just do some calculations and reader gets some nice very relevant data. That radiator test is actually a bit genius, due to how it tests pressure and air velocity as all once, but obviously it doesn't give lots of data if fan is used in other conditions.

 

[VSG]

I think that the most practical and easiest to implement change would be to add CFM per dBA graph. No need to collect new data, just do some calculations and reader gets some nice very relevant data. That radiator test is actually a bit genius, due to how it tests pressure and air velocity as all once, but obviously it doesn't give lots of data if fan is used in other conditions.

I did that before elsewhere and it quickly became a nightmare with overlapping graphs. This is the most accessible approach at this time.

 

[The red spirit]

I did that before elsewhere and it quickly became a nightmare with overlapping graphs. This is the most accessible approach at this time.

Could you link me an example of that?

 

[formula383]

Each fan has a type of use, so if i was to make a review those fans would be tested in there specific use case. But again most fans are pretty similar for most use cases. so not really a huge deal imho.

And you could always throw in other use cases mix matching for fun and depth too. (you know someone is going to want to see more data no matter how much you put out)

 

[The red spirit]

@VSG
I opened latest fan review:

CORSAIR iCUE SP120 RGB ELITE Fan Review

CORSAIR adds to their ever-growing fan family with the much-requested SP RGB ELITE fans. Previously only sold with RGB variants of their cases, these new fans offer excellent PWM control over a relatively long RPM range for case fans, coupled with low noise, directed-airflow performance, and the...

www.techpowerup.com

And calculated cfm/dBA ratio of all fans at 1500 rpm. I found some interesting differences. The best fan in this metric was Noctua NF-A1225 PWM, with whopping 0.945 cfm/dBA ratio, meanwhile the worst fan in this metric was Silverstone SST-FM121 (and very close to it was Noctua NF-F12 PWM) it had 0.745 cfm/dBA ratio (0.749 for Noctua F12). Surprisingly enough, that 140mm Silverstone, which imo should be disqualified from here due to not being 120mm fan, while managed to get above average cfm/dBA ratio, it failed to beat Noctua, as it reached 0.905 cfm/dBA.

I personally think that this metric should be in reviews as it gives a very valuable insight in now fans perform, imo it's more important than just cfm or just dBA. I still don't get it how it ended up being graph hell for you, but if you recalculate it for each tested rpms and then post it as separate graphs, I don't see how it will be problematic.

 

[FireFox]

I built quite a bit of systems and only once with radiator.

Shame on you

 

[The red spirit]

Shame on you

I would only consider water cooling again, if it was custom, without pump, without fans and still would cool stuff quite well. Basically, a big car rad with lots of coolant. For other sane, cost conscious and silent builds, absolutely nothing can touch air cooling. Something like Scythe Mugen 5 PCGH is where it's at.

 

[mtcn77]

@VSG
I opened latest fan review:

CORSAIR iCUE SP120 RGB ELITE Fan Review

CORSAIR adds to their ever-growing fan family with the much-requested SP RGB ELITE fans. Previously only sold with RGB variants of their cases, these new fans offer excellent PWM control over a relatively long RPM range for case fans, coupled with low noise, directed-airflow performance, and the...

www.techpowerup.com

And calculated cfm/dBA ratio of all fans at 1500 rpm. I found some interesting differences. The best fan in this metric was Noctua NF-A1225 PWM, with whopping 0.945 cfm/dBA ratio, meanwhile the worst fan in this metric was Silverstone SST-FM121 (and very close to it was Noctua NF-F12 PWM) it had 0.745 cfm/dBA ratio (0.749 for Noctua F12). Surprisingly enough, that 140mm Silverstone, which imo should be disqualified from here due to not being 120mm fan, while managed to get above average cfm/dBA ratio, it failed to beat Noctua, as it reached 0.905 cfm/dBA.

I personally think that this metric should be in reviews as it gives a very valuable insight in now fans perform, imo it's more important than just cfm or just dBA. I still don't get it how it ended up being graph hell for you, but if you recalculate it for each tested rpms and then post it as separate graphs, I don't see how it will be problematic.

What you are saying is just like quoting this one article where it says Ice Giant is the latest and greatest invention. It is anectodal. We can't do everything by a single benchmark. There are outliers with every reference point like Ice Giant not being able to cool a 3970X to silicon lottery volts.

 

[freeagent]

Scythe Mugen 5 PCGH is where it's at

What makes you say that?

 

[FireFox]

I am confused about this:



@VSG what rad did you use for the test?

 

[VSG]

I am confused about this:

View attachment 208176

@VSG what rad did you use for the test?

Swiftech MCRP120QP, based on previous tests to determine a radiator that is average for airflow restriction as well. FYI, the testing details are here: https://www.techpowerup.com/review/darkside-gentle-typhoon-1450-black-edition/4.html

As for the CFM/dB charts, I misunderstood what @The red spirit said and I get it now. It does seem like a good idea, let me see what's the best way to implement it in the future.

 

[The red spirit]

What makes you say that?

800 rpms or less cools anything on the market.

 

[FireFox]

Swiftech MCRP120QP, based on previous tests to determine a radiator that is average for airflow restriction as well. FYI, the testing details are here: https://www.techpowerup.com/review/darkside-gentle-typhoon-1450-black-edition/4.html

I dont know if i am deaf or what but i have 8x Corsair LL120 RGB and when they run at 750 Rpm i don't hear a thing

 

[mtcn77]

Swiftech MCRP120QP, based on previous tests to determine a radiator that is average for airflow restriction as well. FYI, the testing details are here: https://www.techpowerup.com/review/darkside-gentle-typhoon-1450-black-edition/4.html

As for the CFM/dB charts, I misunderstood what @The red spirit said and I get it now. It does seem like a good idea, let me see what's the best way to implement it in the future.

He wants dBA matched temperature charts, not rpm matched temperature charts basically.

PS: @VSG do you have, or plan to test any 38mm(nfa12×15 with nfa12×25) tests in the mean time? I find optimaltech is up to something with the two piece deep section fan assembly.

 

[VSG]

I dont know if i am deaf or what but i have 8x Corsair LL120 RGB and when they run at 750 Rpm i don't hear a thing

You also probably don't hear these fans from 6" away in an anechoic chamber as those measurements are taken for good SNR.

He wants dBA matched temperature charts, not rpm matched temperature charts basically.

PS: @VSG do you have, or plan to test any 38mm(nfa12×15 with nfa12×25) tests in the mean time? I find optimaltech is up to something with the two piece deep section fan assembly.

I have no plans to do temperature charts for fan reviews, it just adds more variables and I'd rather focus on testing purely the purpose of the fan in this aspect which is to blow air through the radiator.

The EK-Meltemi and the Silverstone fan were the only thick fans I ever planned to do, no one really buys 38 mm fans for PC watercooling.

 

[The red spirit]

As for the CFM/dB charts, I misunderstood what @The red spirit said and I get it now. It does seem like a good idea, let me see what's the best way to implement it in the future.

I had some time to kill and I used that same Corsair fan review to pull data from and I made this Excel file with graph:

Dropbox

www.dropbox.com

Most fans are pretty much the same, but some stand out. Higher diameter Silverstone fan stood out by offering a lot more airflow at lower rpms for lower noise level, but at high rpms was overtaken by Noctua NF-A12x25. Phanteks PH-F120MP while close to other fans, it was second best after Noctua. All fans seemingly become more efficient at higher rpms, which is something nice to know. Noctua NF-F12 PWM was the worst of the bunch. And for some reason RGB fans seemingly perform worse than their non RGB models. I thought that it had no impact on performance, due to most fans just making transparent ring around blades, but if blades themselves light up, then it's reasonable to expect lower performance by small margin (it probably doesn't matter, but I expected no difference).

I have no plans to do temperature charts for fan reviews, it just adds more variables and I'd rather focus on testing purely the purpose of the fan in this aspect which is to blow air through the radiator.

That's a good idea, because temperature changes air density and viscosity and you get some rather unpleasant variables to deal with.

 

[mtcn77]

The EK-Meltemi and the Silverstone fan were the only thick fans I ever planned to do, no one really buys 38 mm fans for PC watercooling.

I mean, not really, just connecting them in the same axial frame. One 15mm fan, one 25mm...
Like this;

I think they haven't tested two different section fans in the same socket, just did a regular push pull setup.

 

[VSG]

I mean, not really, just connecting them in the same axial frame. One 15mm fan, one 25mm...

Spoiler

Like this;
View attachment 208181View attachment 208182

I think they haven't tested two different section fans in the same socket, just did a regular push pull setup.

Seems quite impractical aside from a one-off "What happens" style article. But between keyboards, audio, networking, and everything watercooling, there's no time at the present moment sadly.

 

[freeagent]

800 rpms or less cools anything on the market.

I don’t know what you mean by that?

 

[FireFox]

I don’t know what you mean by that?

That the Scythe Mugen 5 PCGH at 800rpm or less cool anything on the market?

Maybe i am wrong, just guessing

 

[The red spirit]

I don’t know what you mean by that?

That's the maximum rpms of that thing and it's good enough to cool anything that you can buy.