Air Cooling -- Myths and setup tips for the novice performance / gaming builder

[Shrek]

Dang it Shrek, don't do this to me. A 13700K* will go from 20C to 100C, or thermally throttling in an air-conditioned room, in 92 ms. This is equivalent to around 18 million clock cycles for the P-cores, and just shy of 14 million E-core clock cycles at full turbo. Note that this is still less time than it takes you to blink once.

*Assuming pure silicon, a die .5mm thick, PL2, and no heat loss

92 ms is around a tenth of a second, so are we not talking around 300 million clock cycles?

Either way, a tad bigger than 1, so it seems to me we are agreeing.

 

[dirtyferret]

Here is my personal experience with Air vs AiO liquid cooler debates - I don't care, use whatever you want to get the performance you want

90% percent of the opinions and stats posted are useless to other users unless you are running the exact same set up (including exact same case & fans @same rpm) in the exact same environment running the exact same software all at the exact same hardware settings.

too many people flip flop from generic, AiO outperforms Air! to specific, well my stock i7-7700k had better temps on my Noctua U14S than my Corsair H60 therefore Air is better than AiO on any and all CPUs! Some even mix and match, my Artic Freezer II 280 beats any air cooler I ever used!...well if your history of air coolers consists of the intel stock cooler and CM 212+ than I really hope a $100 AiO would outperform them.

too many people confuse spike temps with regular performance temps.

I know top end AiO outperform Air coolers, Ive read all of crazyeyes reviews but I choose air coolers for the following reasons;

1. for the performance I need (more on that below) and the price I'm willing to spend, Air gives me more for my money
2. I don't need to worry about pump failure or tube leaks
3. Noise is part of performance and I don't know where this myth that AiO are silent comes from. A) the pump makes a noise and B) AiO generally come with 2-3 fans that run at higher RPM than Air coolers which generally stick to 1-2 fans at lower RPM. I'm a noise freak so if I have to tone down the fan profile for a top end AiO that means there is good chance I will be paying top end dollars for hardware that won't be giving me top end results (obviously due to me holding it back).

Now, can other people have the exact opposite opinion on all that? Sure.

 

[Bill_Bright]

92 ms is around a tenth of a second, so are we not talking around 300 million clock cycles?

Either way, a tad bigger than 1, so it seems to me we are agreeing.

Come on, Andy! Talk about picking nits. To go from cold to overheated in less than 1/10th of a second is pretty darn quick - which clearly was my point - which was to agree with Mussels and his comment that temps can change "incredibly fast".

 

[Shrek]

Absolutely, but 300 million times slower than your claim... hardly nitpicking.

 

[Vayra86]

Absolutely, but 300 million times slower than your claim... hardly nitpicking.

We ain't in space or something where these orders of magnitude somehow matter

We're in a world where any temp throttling creates a performance penalty, that's the context, and 1 second or 1/100th of a second in that sense is not quite so relevant in it.

The high temp targets are a result of architecture, and almost impossible to circumvent with cooling.

 

[Shrek]

OK, so two orders of magnitude... but eight?

 

[Vayra86]

OK, so two orders of magnitude... but eight?

Context. How fast does it take a liquid or an air cooling setup, to change its cooling capacity? The main differentiator here is fan speed, and it has pretty stringent limits.

These limits can't keep up with how fast CPUs can increase temp.

 

[Bill_Bright]

Absolutely, but 300 million times slower than your claim... hardly nitpicking.

Oh? And what was my claim? I never put a number to how many clock cycles it takes. I said a "few". A "few" is a relative number. And IMO, 1/10th or 1 out of 10 (a number YOU picked) is definitely a "few".

So absolutely, you are being nitpicky and argumentative when CLEARY even you understand "we" were pointing how fast the temps can go from one extreme to the other.

 

[Shrek]

Few ≠ 100,000,000


Your claim

While a CPU can go from cool to overheated in just a few clock cycles (and remember, for many CPUs there are 3 billion+ clock cycles per second) it can cool down almost as quickly, when air flow is properly setup.

 

[dirtyferret]

Context. How fast does it take a liquid or an air cooling setup, to change its cooling capacity? The main differentiator here is fan speed, and it has pretty stringent limits.

These limits can't keep up with how fast CPUs can increase temp.

I recall reading a test several years ago on how an AiO can take temp down faster than an Air cooler due to the larger "area" (not volume) of the radiator and higher RPM fans. Maybe that's still true or not but the real question is, "is the high spike impacting CPU performance?" Obliviously no one wants to be near the limits of of operational temp and get throttled but I get the same CPU performance at 60c as I do at 68c.

 

[Bill_Bright]

Your claim

Yes. My claim said a "few".

100,000,000/3,000,000,000
100 million divided by 3 billion
.03!!!

That's 3/100!

3 100ths

Or 3 out of 100. Yes, Andy, that's a "few".

You can keep being nitpicky, if you choose Andy. That is still incredibly fast!

I'm done here.

 

[Space Lynx]

And the bottom line is:^^^This!^^^

My final setup: 3x Arctic Storm 120mm exhaust fan, 2 at top and 1 at rear. and 2x200mm intake fans up front. and my noctua nhd15s in the middle with its 1x140mm noctua fan.

should be a pretty good rig overall as far as airflow and cooling goes. keeping it nice and simple.

 

[Bill_Bright]

To me, you likely have more than you need - but for me, that is due in part because I hate fan noise. At least with that many, you can move massive amounts of air while keeping your fan rotation speeds slow - thus more quiet.

 

[claes]

My final setup: 3x Arctic Storm 120mm exhaust fan, 2 at top and 1 at rear. and 2x200mm intake fans up front. and my noctua nhd15s in the middle with its 1x140mm noctua fan.

should be a pretty good rig overall as far as airflow and cooling goes. keeping it nice and simple.

Too much exhaust if you ask me, but if negative pressure is what you’re after then go for it

 

[Count von Schwalbe]

Too much exhaust if you ask me, but if negative pressure is what you’re after then go for it

It is positive pressure - the area alone of a 200mm fan is almost a great as 3x 120mm fans, and he has 2. Also, a larger fan moves more air for its area at the same speed, as the tips of the fan blade are moving faster.

 

[Bill_Bright]

Don't forget, unless the PSU has its own intake, it is adding to the exhaust side.

Negative is fine - if your case does not have air filters because then the vacuum will pull in the extra that it needs through every unfiltered crack, crevice and port. This allows dust inside the case, but also allows the dust to get packed into uncovered ports, like unused USB and audio ports. If you have an optical drive, it will also allow dust to collect near the lens area - not good.

But if the case is filtered (and you don't want to bypass those filters), you want the slight over (positive) pressure as that pulls in the air through the air filters, and pushes out any extra air through those cracks, crevices and ports - good.

 

[Space Lynx]

I'm not opposed to switching it around. My case is the Montech Air X.

Any suggestions welcome.

How about 1x extra fan attached to the NH-D15S + 2 top exhaust and no rear exhaust?

and this case has dust filters everywhere, the PSU draws from the bottom of the case fan down

 

[freeagent]

If you use good fans, you don't really need to use an exhaust fan. I do have one installed for now, just for old times' sake. It's a 140mm fan that snaps right into the Torrent Compact, don't even need to use screws lol. It's a TY-143 that is limited to 75%. Funnily enough, my TY-147 will not fit and they are pretty much the same fan, at least in shape. NF-A14 iPPC will never fit. So, it's just there for the novelty, not out of necessity. I do have 2x NF-A14 iPPC at the bottom of my case, but they are limited to 37% lol...

When I stopped using my Meshify C, I had the top blocked off and was just using 2x140s and a 120 in the front. They were iPPC 3Ks, but they still worked great. Loud AF though if you don't limit them.

 

[Sithaer]

Don't forget, unless the PSU has its own intake, it is adding to the exhaust side.

Negative is fine - if your case does not have air filters because then the vacuum will pull in the extra that it needs through every unfiltered crack, crevice and port. This allows dust inside the case, but also allows the dust to get packed into uncovered ports, like unused USB and audio ports. If you have an optical drive, it will also allow dust to collect near the lens area - not good.

But if the case is filtered (and you don't want to bypass those filters), you want the slight over (positive) pressure as that pulls in the air through the air filters, and pushes out any extra air through those cracks, crevices and ports - good.

Thats exactly whats happening with my In Win 101 case, I have it set up the default way based on In Win's instructions but this way I get air sucked in at the right and left side of the glass panel cause there is a small gap. 'its not a completely shut panel'

Bottom 3 fans intake, side 2 exhaust and also the backside 1 fan exhaust but since this case has a top PSU mount it also exhaust from inside of the PC.
Only filtered at the bottom, I thought about either removing the side 2 fans or turning them intake but that area is not filtered so it would take even more dust inside the case..

Luckily the gap between the glass and the case 'well the case's side next to the gap' catches most of the dust and I just clean it off every now and then but it does allow some dust inside the case so a full case cleaning with an air compressor is needed every few months/half year and around once a month glass panel cleaning.

As much as I like the look of this case its a bit of a pain in the ass to build in/keep clean but I don't really feel like buying a new case and to be honest most case designs nowadays doesn't catch my fancy or they are simply too expensive for my budgets. 'at least the temps are fine'

 

[Vayra86]

I recall reading a test several years ago on how an AiO can take temp down faster than an Air cooler due to the larger "area" (not volume) of the radiator and higher RPM fans. Maybe that's still true or not but the real question is, "is the high spike impacting CPU performance?" Obliviously no one wants to be near the limits of of operational temp and get throttled but I get the same CPU performance at 60c as I do at 68c.

Roger but that's the problem with recent top end CPUs, they push so much power through the chip you do get temps at throttle target. AMD even designed to maintain near it.

 

[Udyr]

I'm not opposed to switching it around. My case is the Montech Air X.

Any suggestions welcome.

How about 1x extra fan attached to the NH-D15S + 2 top exhaust and no rear exhaust?

and this case has dust filters everywhere, the PSU draws from the bottom of the case fan down

I got the same case with the 2x200mm (intake) in front and the 120mm (exhaust) in the rear, but I added two Noctua fans I had from a previous build, 1x120mm at the bottom as intake (yes, you can put fans there, even if it's not officially supported) and 1x120mm on top of the CPU cooler (cooler has 2x120mm fans) as exhaust. The PSU is set to exhaust with the fan always on.

 

[Bill_Bright]

If you use good fans, you don't really need to use an exhaust fan.

I think that really depends on the case and what it has for exhaust vents. If the exhaust vents are limited, the cooling may need a little help from a fan to help push the heated air out. Otherwise, you just keep pushing air in but it has nowhere to go but build up and become stagnant - and very warm.

I see this on a bigger scale for my main bathroom. It has a furnace vent pushing warm air in, but the return vents are out in the hallway. If I want a toasty bathroom in the morning, I close the door so the only "flow" is through the gap at that bottom of the door. The warm air from the furnace builds up, and heats up that room nicely. If I forget to close that door, the heated air from the vent goes straight through.

 

[Space Lynx]

I think that really depends on the case and what it has for exhaust vents. If the exhaust vents are limited, the cooling may need a little help from a fan to help push the heated air out. Otherwise, you just keep pushing air in but it has nowhere to go but build up and become stagnant - and very warm.

I see this on a bigger scale for my main bathroom. It has a furnace vent pushing warm air in, but the return vents are out in the hallway. If I want a toasty bathroom in the morning, I close the door so the only "flow" is through the gap at that bottom of the door. The warm air from the furnace builds up, and heats up that room nicely. If I forget to close that door, the heated air from the vent goes straight through.

how would you recommend i set up the montech air x case?

I got the same case with the 2x200mm (intake) in front and the 120mm (exhaust) in the rear, but I added two Noctua fans I had from a previous build, 1x120mm at the bottom as intake (yes, you can put fans there, even if it's not officially supported) and 1x120mm on top of the CPU cooler (cooler has 2x120mm fans) as exhaust. The PSU is set to exhaust with the fan always on.

I don't think I will do this, why is another intake needed when you already have 400mm of it?

I'm thinking just 2x200mm intake, then 1x rear exhaust and 1x top exhaust nearest to the rear one, and leave the other exhaust blank... just to give it a directional diagonal flow at least? Cause the heat from the gpu will be pumping into the case as well.

 

[claes]

It is positive pressure - the area alone of a 200mm fan is almost a great as 3x 120mm fans, and he has 2. Also, a larger fan moves more air for its area at the same speed, as the tips of the fan blade are moving faster.

I don't think I will do this, why is another intake needed when you already have 400mm of it?

Why are you guys measuring airflow in terms of fan size? CFM is what matters here — the area of a fan frame has no bearing at all on how much air it can move…

More, with intakes you have to account for the chassis frame, filters, and any other obstructions, all of which will significantly reduce airflow. Even a fine mesh filter can cut CFM in half. Combine that with low speed, low pressure 200mm fans (this is where size matters) and you have very limited intake.

I'm thinking just 2x200mm intake, then 1x rear exhaust and 1x top exhaust nearest to the rear one, and leave the other exhaust blank... just to give it a directional diagonal flow at least? Cause the heat from the gpu will be pumping into the case as well.

Your intakes are rated for 90 CFM, so 180 CFM - obstructions. The Air X seems to have a fair bit of them, so let’s say 60% of 180, or 108 CFM. I tried to find an Arctic Storm 120 but no dice, so I can’t help there.

Looking at your case I’d probably add an intake at the top front if you need it and just leave the single, rear exhaust.

 

[Space Lynx]

Why are you guys measuring airflow in terms of fan size? CFM is what matters here — the area of a fan frame has no bearing at all on how much air it can move…

More, with intakes you have to account for the chassis frame, filters, and any other obstructions, all of which will significantly reduce airflow. Even a fine mesh filter can cut CFM in half. Combine that with low speed, low pressure 200mm fans (this is where size matters) and you have very limited intake.

Your intakes are rated for 90 CFM, so 180 CFM - obstructions. The Air X seems to have a fair bit of them, so let’s say 60% of 180, or 108 CFM. I tried to find an Arctic Storm 120 but no dice, so I can’t help there.

Looking at your case I’d probably add an intake at the top front if you need it and just leave the single, rear exhaust.

Antec Storm, sorry, my bad.

"Provide the High Performance max. 66.56 CFM, max. 2.7 mmH2O through narrower frame and bigger air inlet area."

Montech Air X has a TPU review, and it was reviewed very well for airflow. I guess I can just look at pictures of the review setup and go from there