The Most Defining Key Factor of Air Cooler Performance?

[_Hustler_One_]

I'm wondering of what is the key factor that determines an air cooler performance in the same size class.

 

I know that the bigger the fan size the better but let's say every air cooler in 120 mm class. Between all the 120 mm class air coolers, what is the key factor that a cooler can stands over the others?

Is it the heatsink (shape design or size/overall thickness), or the heat pipes (the numbers or the size), or the fan (blade design or high RPM)??

Between those 3 factors with each subfactors, which one is the most defining key factor for an air cooler performance???

 

Let's not to involve price, we are talking about pure performance..

 

So I will have an insight of what to consider when I need to get one, or another one..

 

Let's discuss and share knowledge about it..

 

Feel free to add something that I missed..

[iRileyx]

There is no single defining factor, the heat pipes, fins, contact area(with cpu) all play a part into it. 

Even the fans they come with can make a difference, if you were to run at 100% with a 1200 vs 2400rpm fan for example. 

 

[Bcat00]

From what i recall, the pipes are better off in smaller/medium size with more pipes versus having a couple of giant pipes.

Other than that, nothing else all that much in term of fins aside from molding it to fit the air flow better.

[Pixel5]

as always it depends, there is no single most important thing as these days all decent air coolers have a certain standard like using copper for the contact surface to the IHS and any tower style cooler will have heat pipes.

 

everything else simply depends on many different things, the only thing all have in common is that their performance depends heavily on the ambient temperature.

[Cereal5]

I couldn't say for sure if there's ever been a test on this but I like it.

 

I would think that it falls mostly to the hands of the heat pipes/fins, as that is how the heat of the CPU gets distributed, and is what the fan is actually blowing air on.

 

Now obviously everything you mentioned (and more) plays into how well a cooler cools a CPU. I just like to compare everything within the law of diminishing return. So let's say we have 3 levels of air coolers. These levels, for arguments sake, are:

1) Stock (Intel or AMD)

2) Cooler Master Hyper 212 Evo

3) Noctua NH-D15

 

On one end we have the most basic thing needed to not fry your chip. A heat spreader, a few fins, and something like an 80mm fan. It works.

In the middle, we have the next logical step up (somewhat skipping coolers like the Noctua L9i). The 212 evo is a fully realized cooler with 4 heat pipes and plenty of fins to boot. Also included, a 120mm fan. It even supports mild to medium overclocking, depending on the chip.

And finishing the trio is the NH-D15. Big, beefy, this one really dissipates some heat. 6 heat pipes, more fins, more (bigger) fans. It's the big cheese

 

If all you did was increase one spec on any of the coolers, the cooling capacity would increase. No one really knows how much. You can also compare it to a radiator (whether it be computer or car). Fin density, pipe size/count, blah, blah, blah, etc., etc.

[Enderman]

You need a good combination of all 3.

The main thing is the surface area of the fins, which is what dissipates the heat.

Heat pipes are second most important since they need to carry the heat to the fins.

Then there are like 50 other variables too.

 

The only way to know which is best is to look at benchmarks and tests.

[hereiam]

From an engineer's perspective, everything is defined in equation 1, 2 and 3 in this article: https://en.wikipedia.org/wiki/Heat_sink

 

So basically, in the case of a CPU, Q(dot) is the total dissipation rate, which is equal to the (thermal) power of your CPU. cp is the thermal capacity of air, which depends on various factors you have little control over (inlet temperature, humidity). m(dot) is the flow rate of your fan, which should be proportional (roughly) to its RPM and fan diameter. Rhs is the total thermal resistance of your setup, which is equal to (die to heat spreader resistance) + (thermal paste resistance) + (heatsink package resistance) .

 

So how does all this determine the heat dissipation performance? If you fix the air flow rate, CPU temperature (Ths) and air inlet Tair,in temperature, the only variables you can control to maximize the total dissipation rate are 1) lowering thermal resistance Rhs, which involves bigger heat pipe, better thermal conductive material, better thermal paste or deliding; or 2) increasing outlet temperature Tair,out. The later is achieved either through bigger fins, more fins, increasing total surface area of the fins, or optimizing airflow to maximize heat exchange.

 

As you might expect, all those optimizations have diminishing return, so the best heatsink will try to achieve all those optimizations simultaneously. 

 

TL;DR: there's no single most important factor deciding the performance of a heatsink. There's no point using huge heatpipes on heatsink with insufficient fins area. Similarly, huge heat sink with insufficient heat pipes to carry the heat to the fins is also pointless (unless you want to intimidate spectators).

[EsaT]

2 hours ago, _Hustler_One_ said:

I know that the bigger the fan size the better

Bigger fan can move same airflow at lower RPM, but that means greatly decreased pressure.

So it's not that straightforward and needs compensating in heatsink design.

 

14 minutes ago, Enderman said:

The main thing is the surface area of the fins, which is what dissipates the heat.

And same applies also to waterpipe coolers.

Compact and/or slim radiators simply don't have excess of surface area.

Which is why with additional noise sources those often struggle against good heatpipe coolers in cooling per noise.

 

In similar type coolers weight is pretty good consultant for how much surface area there can be.

Though making fins thinner gives more area per weight...

While increasing thermal resistance of fin in spreading heat from heatpipes to whole surface.

 

Set heatsink size also affects to how much surface area can be crammed into it.

Bigger dimensions (or multiple fin stacks) allow more surface area, but that leads fast to issues with DIMM/PCI-e slot clearances.

So some heatsinks pack fins lot denser than others.

While that gives more surface area per cooler dimensions, it causes consequencies elsewhere:

Densely packed fins have higher impedance for airflow and need more pressure from fan to have any airflow.

So even if they have more surface area, they can do worser with quiet low speed fan, than lighter heatsink with less surface area, but with less airflow impedance.

 

Alternatively heatsink can use offcenter design moving fin stack backward to make space for fan without it covering DIMM slots.

For example Thermalright uses that widely with their big dimensions low density fin stacks good for low airflow or even passive cooling.

Also Scythe Mugen 5 uses that design for great effect challenging many bigger (and lot more expensive) coolers in cooling performance per noise.

 

 

As for part of transferring heat to fin stack there's limit for how much heat single certain heatpipe can transfer.

So higher end coolers often use more heatpipes or in some cases bigger ones.

Though while thicker heatpipe has big increase in cross sectional area, their surface area doesn't increase as much.

And that surface area is needed for contact with fins and moving condensed heat transfer liquid back to heatsink base.

That's likely why 6mm heatpipe diameter has stayed as standard.

 

Though you don't need ten heatpipes to do well, unless heat load is huge.

For example Cryorig H7 has three heatpipes and performs great for its size and especially price.

Its otherwise good design simply compensates that single heatpipe difference to many cheap four heatpipe coolers.

Its heatpipes also likely have good design/construction capable to moving liquid fast back to base area.

 

Third part is getting that heat out from CPU to heatpipes.

Good firm contact between CPU and base of heatsink is crucial for good heat conduction.

So any cooler using those crappy plastic push pins is in major disadvantage to same size cooler using mounting to backplate behind motherboard.

And actual heat producing silicon die of CPUs is lot smaller than size of heatpsreader.

So naturally it's better to have heatpipes densely packed and as close as possible to center of heatspreader.

 

All traditional design coolers have heatpipes tighly packed and soldered into base of cooler.

But many so called "direct contact heatpipe" coolers even have space between heatpipes, pushing other than centermost heatpipes away from hot spot.

That hinders heat conduction to those, even if they're actually soldered on their place.

There's lots of marketing hype about direct contact heatpipes, but in the end its just cost cutting design:

If it actually performed better expensive high end coolers would certainly use it!

 

Also all around penny pinching Intel has problems in making heatspreaders flat and many of them are actually slightly concave.

That easily leads to increase in TIM layer thickness possibly greatly hindering heat conduction.

Hence some coolers have tiny bit convex base to increase chances for good contact in center where it matters most.

Not sure if bases of direct contact heatpipe coolers can be machined that way.

[_Hustler_One_]

4 hours ago, EsaT said:

Bigger fan can move same airflow at lower RPM, but that means greatly decreased pressure.

So it's not that straightforward and needs compensating in heatsink design.

 

And same applies also to waterpipe coolers.

Compact and/or slim radiators simply don't have excess of surface area.

Which is why with additional noise sources those often struggle against good heatpipe coolers in cooling per noise.

 

In similar type coolers weight is pretty good consultant for how much surface area there can be.

Though making fins thinner gives more area per weight...

While increasing thermal resistance of fin in spreading heat from heatpipes to whole surface.

 

Set heatsink size also affects to how much surface area can be crammed into it.

Bigger dimensions (or multiple fin stacks) allow more surface area, but that leads fast to issues with DIMM/PCI-e slot clearances.

So some heatsinks pack fins lot denser than others.

While that gives more surface area per cooler dimensions, it causes consequencies elsewhere:

Densely packed fins have higher impedance for airflow and need more pressure from fan to have any airflow.

So even if they have more surface area, they can do worser with quiet low speed fan, than lighter heatsink with less surface area, but with less airflow impedance.

 

Alternatively heatsink can use offcenter design moving fin stack backward to make space for fan without it covering DIMM slots.

For example Thermalright uses that widely with their big dimensions low density fin stacks good for low airflow or even passive cooling.

Also Scythe Mugen 5 uses that design for great effect challenging many bigger (and lot more expensive) coolers in cooling performance per noise.

 

 

As for part of transferring heat to fin stack there's limit for how much heat single certain heatpipe can transfer.

So higher end coolers often use more heatpipes or in some cases bigger ones.

Though while thicker heatpipe has big increase in cross sectional area, their surface area doesn't increase as much.

And that surface area is needed for contact with fins and moving condensed heat transfer liquid back to heatsink base.

That's likely why 6mm heatpipe diameter has stayed as standard.

 

Though you don't need ten heatpipes to do well, unless heat load is huge.

For example Cryorig H7 has three heatpipes and performs great for its size and especially price.

Its otherwise good design simply compensates that single heatpipe difference to many cheap four heatpipe coolers.

Its heatpipes also likely have good design/construction capable to moving liquid fast back to base area.

 

 

6 hours ago, iRileyx said:

There is no single defining factor, the heat pipes, fins, contact area(with cpu) all play a part into it. 

Even the fans they come with can make a difference, if you were to run at 100% with a 1200 vs 2400rpm fan for example. 

 

 

Right.. I think The Thermalright Grand Macho is an ideal example which simply apply all the factors I asked by stay being traditional in design, doesn't it? Like back to basic..

 

It has really vast surface area, dense fins, 7 heatpipes, big fan, yet it performs on top of the 140 mm list. But I found the opposite in 120 mm class which is BeQuiet! Shadow Rock 2. In my PoV it's simply the miniature of the Macho, it has vast surface area and dense fins, 4 big heat pipes, strong fan, but why it performs relatively worse among the other 120 mm coolers?

 

I'm thinking of how if Shadow Rock 2 heatsink paired with Noctua NF-F12 fan.. Will it make it better? Or there is other factor that makes it perform not that good regarding the basic properties that it has like the Macho has???

 

Is there someone can explain this?

 

 

These 2 similar coolers but difference in relative performance in their own class yet has same properties is questioning.. This case is stuck in my mind recently..

 

 

UPDATE:

So I found about the Scythe Mugen 5 which has almost as thick as Shadow Rock 2 but slimmer with 6 heat pipes on it and performs better, both are paired with 120 mm fan but Shadow Rock 2 fan max RPM is 1600 rpm and Mugen has lower max RPM at 1200 rpm. With this, can we now say that the number of the heat pipes is the key factor?

 

 

[Enderman]

Just ignore what the heatsink looks like and go look at benchmarks.

Looking at it will never tell you how it performs, there are hundreds of variables.

Everything from what grade of copper the base plate is made of, the heat transfer rate of the heat pipe wick, the surface coating of the heatsink fins, the thermal connection between the heatpipes and fins, etc etc etc.

You're wasting your time if you think you can determine the best by looking at only what is visible in a picture.

 

Tests like this is the only way to know:

[EsaT]

5 hours ago, _Hustler_One_ said:

Right.. I think The Thermalright Grand Macho is an ideal example which simply apply all the factors I asked by stay being traditional in design, doesn't it? Like back to basic..

 

It has really vast surface area, dense fins...

All Thermalright Machos have fin spacing in looser end of heatsinks, their surface area comes from big dimensions.

Hence air flows easily through them without high speed high pressure fan.

Also general quality is good.

 

Quote

But I found the opposite in 120 mm class which is BeQuiet! Shadow Rock 2. In my PoV it's simply the miniature of the Macho, it has vast surface area and dense fins, 4 big heat pipes, strong fan, but why it performs relatively worse among the other 120 mm coolers?

That Bequiet has extremely dense fin spacing with 51 fins. (Mugen 5 has 39)

And not only that but it's also very long in direction of airflow. (unlike another 50 finner Noctua NH-U12S)

Hence it has super high airflow impedance and likely very inefficient airflow through fin stack.

Unless given two high speed high pressure fans in push-pull configuration sounding like aircraft in take off.

 

Also like I said actual surface area of heatpipe doesn't grow as fast as its cross section.

So efficiency of heat conduction from heat pipes to fins and transfer of liquid back to base might be lower.

Higher number of smaller heatpipes would have more surface area to contact with fins per same cross section.

Plus with higher number of heatpipes allowing more even transfer of heat to area of fins.

I mean 8mm heatpipes have stayed in niche, so there must be some clear drawback in them.

 

Then come other things, like construction/quality of heatpipes, or thermal contact between heatpipes and fins.

That weight means lot more material cost in fin stack compared to typical heatsinks, so who knows if costs were cut elsewhere.

[WoodenMarker]

7 hours ago, _Hustler_One_ said:

So I found about the Scythe Mugen 5 which has almost as thick as Shadow Rock 2 but slimmer with 6 heat pipes on it and performs better, both are paired with 120 mm fan but Shadow Rock 2 fan max RPM is 1600 rpm and Mugen has lower max RPM at 1200 rpm. With this, can we now say that the number of the heat pipes is the key factor?

Heatpipes play a role but it's hard to know how big of a difference it makes. It could also be a matter of mounting pressure which often makes or breaks a cooler's performance. 

Here's an example of how much difference mounting pressure for a single cooler can make for performance: https://www.overclockers.com/scythe-fuma-heatsink-review/

When comparing different coolers with different mounting hardware, the difference can be even bigger. 

[_Hustler_One_]

4 hours ago, Enderman said:

Just ignore what the heatsink looks like and go look at benchmarks.

Looking at it will never tell you how it performs, there are hundreds of variables.

Everything from what grade of copper the base plate is made of, the heat transfer rate of the heat pipe wick, the surface coating of the heatsink fins, the thermal connection between the heatpipes and fins, etc etc etc.

You're wasting your time if you think you can determine the best by looking at only what is visible in a picture.

 

Tests like this is the only way to know:

I did see the benchmarks off all 3 coolers, that's why I asked, why in such similar design but there is different performance (by seeing benchmarks).

I can't tell each performance of the coolers without see their benchmarks tho, because I don't have them, so I won't talk non-sense about their performance without references.

[Enderman]

1 minute ago, _Hustler_One_ said:

I did see the benchmarks off all 3 coolers, that's why I asked, why in such similar design but there is different performance (by seeing benchmarks).

I can't tell each performance of the coolers without see their benchmarks tho, because I don't have them, so I won't talk non-sense about their performance without references.

If you just want the best performance then get an NH-D15.

[_Hustler_One_]

1 minute ago, Enderman said:

If you just want the best performance then get an NH-D15.

I just wanted to know something that makes this difference. Knowledge, you know..

I use NH-U12S which is works like charm already, and I'm very satisfied by its performance.. Also my case cant fit D15, and I would prefer Macho than D15 if I must to change the U12S but unfortunately is a "no", so thanks..

[Enderman]

6 minutes ago, _Hustler_One_ said:

I just wanted to know something that makes this difference. Knowledge, you know..

I use NH-U12S which is works like charm already, and I'm very satisfied by its performance.. Also my case cant fit D15, and I would prefer Macho than D15 if I must to change the U12S but unfortunately is a "no", so thanks..

Not even with a mechanical engineering degree in thermodynamics can you predict the performance of a cooler by just looking at it.

As I said before there are hundreds of variables.

Even a huge company like noctua resorts to trial and error and physical testing to try to improve their products.

[_Hustler_One_]

5 hours ago, WoodenMarker said:

Heatpipes play a role but it's hard to know how big of a difference it makes. It could also be a matter of mounting pressure which often makes or breaks a cooler's performance. 

Here's an example of how much difference mounting pressure for a single cooler can make for performance: https://www.overclockers.com/scythe-fuma-heatsink-review/

When comparing different coolers with different mounting hardware, the difference can be even bigger. 

This cooler is surprises me honestly. Scythe Fuma, I haven't read about this one before. I see this cooler is a 120mm dual towers but can stand in tier 2 on your tier list thread. In the benchmarks it stands above the D15 performance by 4^oC lower. This really surprise me even more, regarding it's a 120mm fans and the height is only 149mm which is can fit inside any ATX case. I would like to have it but unfortunately it is unavailable in my country. Scythe as a brand in general doesn't sell their product where I live.. 

[hereiam]

On 9/9/2018 at 9:35 PM, _Hustler_One_ said:

It has really vast surface area, dense fins, 7 heatpipes, big fan, yet it performs on top of the 140 mm list. But I found the opposite in 120 mm class which is BeQuiet! Shadow Rock 2. In my PoV it's simply the miniature of the Macho, it has vast surface area and dense fins, 4 big heat pipes, strong fan, but why it performs relatively worse among the other 120 mm coolers?

That's because the Shadow Rock 2 is of the very low noise (and low airflow) variant. The 140mm variant of the Pure Wings 2 fan has an airflow rate at max 104 m³/h, whereas the Grand Macho's 140mm fan has an airflow rate 125 m³/h top, about 20% increase.

20% increase in airflow is a huge deal. Not equivalent to a 20% increases in thermal sink rate, due to the aforementioned diminishing return, but still quite significant.

The Noctua 140mm fan has an airflow rate of 140m³/h top, so you might want to use that if you want even better cooling performance. 

On 9/10/2018 at 10:04 AM, _Hustler_One_ said:

This cooler is surprises me honestly. Scythe Fuma, I haven't read about this one before. I see this cooler is a 120mm dual towers but can stand in tier 2 on your tier list thread. In the benchmarks it stands above the D15 performance by 4^oC lower. This really surprise me even more, regarding it's a 120mm fans and the height is only 149mm which is can fit inside any ATX case. I would like to have it but unfortunately it is unavailable in my country. Scythe as a brand in general doesn't sell their product where I live.. 

The Fuma has the Slip Stream 120mm fan, which has a 134.20 m³/h airflow rate, higher than even the 140mm variant of the Grand Macho, and is 50% more than the 87 m³/h of the Shadow Rock 2 120mm Pure Wings 2 fan. That's an incredibly large airflow for a 120mmx25mm fan. Plus the cooler has two of them, increasing noise even further.

As a rule of thumb, for most fans higher airflow equals louder noise. That's why BeQuiet! usually uses low airflow and low noise fans, and performance suffers as a result. 

 

As a side note, it is incredibly hard to have apple to apple comparison between coolers. There are so many variations to consider: fan airflow, pwm-airflow curve, (as previously mentioned) contact thermal resistance, etcetera. Even the geometry of the inside of your case can affect (theoretically) the cooling performance.

[WoodenMarker]

1 hour ago, hereiam said:

That's because the Shadow Rock 2 is of the very low noise (and low airflow) variant. The 140mm variant of the Pure Wings 2 fan has an airflow rate at max 104 m³/h, whereas the Grand Macho's 140mm fan has an airflow rate 125 m³/h top, about 20% increase.

The SR2 comes with a 120mm PW2 at 87 m³/h and not a 140mm. A 140mm fan would probably much better utilize the heatsink but would compromise ram compatibility. 

[hereiam]

29 minutes ago, WoodenMarker said:

The SR2 comes with a 120mm PW2 at 87 m³/h and not a 140mm. A 140mm fan would probably much better utilize the heatsink but would compromise ram compatibility. 

Sorry, I forgot to mention that I used the 140mm variant of the SR2 fan to have a more apple to apple comparison, since I could not find the exact 120mm variant of the Macho fan. For some weird reason, Thermalright has three 140mm fan variants (TY-147A, TY-147B, and TY-147ASQ), but only one 120mm fan variant (the TY-127), making apple to apple comparison quite difficult. 

You're right though, don't just plug a 140mm fan into a heatsink made for a 120mm fan, most of the extra airflow would not actually flow through the heatsink and ended up getting wasted. 

[_Hustler_One_]

4 hours ago, WoodenMarker said:

The SR2 comes with a 120mm PW2 at 87 m³/h and not a 140mm. A 140mm fan would probably much better utilize the heatsink but would compromise ram compatibility. 

So basically SR2 drawbacks are its fins stack which is too dense and its low airflow fan?

 

Also what makes Scythe Fuma Rev.2 thermal performance is worse than the original Fuma? The specs and the design looks identical from my perspective.

 

Why Scythe uses airflow optimized fan than the static pressure fan?

[MEC-777]

@_Hustler_One_

 

It's surface area for heat transfer - in a nutshell.

 

More pipes are better, but only if the fin array is large enough to allow sufficient heat transfer, otherwise the whole thing will "heatsoak". 

 

So: Large fin array with higher number of pipes (not bigger pipes), and have the pipes running up through the fin array out at the sides where air movement from the fan is the strongest.  

[WoodenMarker]

6 hours ago, hereiam said:

Sorry, I forgot to mention that I used the 140mm variant of the SR2 fan to have a more apple to apple comparison, since I could not find the exact 120mm variant of the Macho fan. For some weird reason, Thermalright has three 140mm fan variants (TY-147A, TY-147B, and TY-147ASQ), but only one 120mm fan variant (the TY-127), making apple to apple comparison quite difficult. 

You're right though, don't just plug a 140mm fan into a heatsink made for a 120mm fan, most of the extra airflow would not actually flow through the heatsink and ended up getting wasted. 

The 120mm fan included on the Macho Rev. A for example comes at 78.5 m³/h.

I was saying the opposite for the SR2 since a 120mm fan doesn't directly blow over much of the large heatsink. Most heatsinks of a similar size use 140mm fans. 

2 hours ago, _Hustler_One_ said:

So basically SR2 drawbacks are its fins stack which is too dense and its low airflow fan?

Also what makes Scythe Fuma Rev.2 thermal performance is worse than the original Fuma? The specs and the design looks identical from my perspective.

Why Scythe uses airflow optimized fan than the static pressure fan?

The deep and dense fin stack probably requires more static pressure than what the stock fan can provide to perform optimally. The Fuma on the other hand can get away with less static pressure since it uses two fans for two shallower towers. 

The Rev. 2 comes with spring mounts that prevent overtightening. This was done to prevent damage to cpus with thinner pcbs like Skylake.

[_Hustler_One_]

1 hour ago, WoodenMarker said:

The 120mm fan included on the Macho Rev. A for example comes at 78.5 m³/h.

I was saying the opposite for the SR2 since a 120mm fan doesn't directly blow over much of the large heatsink. Most heatsinks of a similar size use 140mm fans. 

 

The deep and dense fin stack probably requires more static pressure than what the stock fan can provide to perform optimally. The Fuma on the other hand can get away with less static pressure since it uses two fans for two shallower towers. 

The Rev. 2 comes with spring mounts that prevent overtightening. This was done to prevent damage to cpus with thinner pcbs like Skylake.

If SR2 heatsink is being paired with Noctua NF-F12 ippc fans either 2000/3000 PWM which has much higher pressure do you think it will move the SR2 rank up to which tier? Or NF-A14 iPPC is the more ideal to be paired on it than the F12?

If the pairing is done, presumably could it be equal to which other cooler? I'm really curious about this. I want to test it but don't want to waste money for it either.. Need save some money for my kids..

[WoodenMarker]

1 minute ago, _Hustler_One_ said:

If SR2 heatsink is being paired with Noctua NF-F12 ippc fans either 2000/3000 PWM which has much higher pressure do you think it will move the SR2 rank up to which tier? Or presumably can be equal to which other cooler? I'm really curious about this. I want to test it but don't want to waste money for it either.. Need save some money for my kids..

Probably 1 tier, possibly 2 tiers.