Air Cooler

[Fuzzykins]

The Cooler Master Blade Master that comes with the Hyper 212+ is one of the better fans you can use from what I've read, I'm running two of them on mine and the temperatures max at 67C during normal usage, 76C under Prime95 testing.

Also, the Hyper 212+ is still $35 on Amazon... no idea why it's so much on Newegg now.

I'll just buy it on Amazon... Also, how's the stock thermal paste for CMHyper212+

[Poodles]

...
What are you smoking and can I have some?
The fan failures are from idiots who don't properly affix them. Under no circumstance will your motherboard break if you properly affix the CPU cooler, regardless of it's size.. It would break off the standoffs before the motherboard it's self is damaged, and that just means the idiot didn't have his standoffs properly secured.

Incorrect. I've been doing this for a very long time, and it DOES happen. Then again I build PC's for durability and the ability to ship them if needed.

It's parade float fluff styling more than function that drives these designs. Bigger is better, blah blah, not actual performance.

[Fuzzykins]

Incorrect. I've been doing this for a very long time, and it DOES happen. Then again I build PC's for durability and the ability to ship them if needed.

It's parade float fluff styling more than function that drives these designs. Bigger is better, blah blah, not actual performance.

How can you say that with a straight face? You realize more surface area does equal better performance, right?

[BicycleMafioso]

It's parade float fluff styling more than function that drives these designs. Bigger is better, blah blah, not actual performance.

That's all well and good. But until I see any evidence suggesting otherwise, my Noctua NH-D14 is an extremely potent cooler, increasingly better the higher the temperatures.

My Arctic Cooling Freezer Pro 7 rev 2. is surprisingly good though, and better than a stock-fanned Hyper212+. Usually cheaper as well. In replaced-fan PnP Hyper212+, it's obviously very much cheaper, and performance/price, I like it.

[Fuzzykins]

That's all well and good. But until I see any evidence suggesting otherwise, my Noctua NH-D14 is an extremely potent cooler, increasingly better the higher the temperatures.

My Arctic Cooling Freezer Pro 7 rev 2. is surprisingly good though, and better than a stock-fanned Hyper212+. Usually cheaper as well. In replaced-fan PnP Hyper212+, it's obviously very much cheaper, and performance/price, I like it.

Windowed case, and that thing looks like shit. :<

Wow. Thermalright Venemous X. Matches my case perfectly, has great reviews, and isn't horrifically expensive if I don't replace the fans. (AFAIK, it's not awful with just a single fan either.) Now if only I could fit a black and white EVGA board in my budget. /chinscratch

[BicycleMafioso]

Windowed case, and that thing looks like shit. :<

That's because you think looks matter.

Silence and performance is my forte.

... With that said, if someone would give me a LEVEL10GT from Thermaltake.. I would accept. Despite the discolights and windows.

[Adappy]

Also, how's the stock thermal paste for CMHyper212+

It's not the best but it's not bad either. I've only used it and AS5 and AS5 was only 1-2C cooler when I used it originally (this was comparing the two at 4.4GHz @1.28V, I've only experienced the CM stuff with my current overclock)

[Fuzzykins]

That's because you think looks matter.

Silence and performance is my forte.

... With that said, if someone would give me a LEVEL10GT from Thermaltake.. I would accept. Despite the discolights and windows.

I like refined aesthetics. The black and white theme looks great. Also, it seems a bit small. Logic pretty much dictates that a 120MM fan heatsink will perform better than a 92mm due to more heatsink area.

[Poodles]

Not necessarily true. Heat sink design is far more complex than that.

How can you say that hanging a 3 lb weight off the side of a silicone board is at all smart? The fact that most of these new designs don't fit it many cases shows just how foolish they are.

Do what you want, it's your PC of course, but know that there are better solutions that are more robust and durable and won't risk your hardware.

[Fuzzykins]

Not necessarily true. Heat sink design is far more complex than that.

How can you say that hanging a 3 lb weight off the side of a silicone board is at all smart? The fact that most of these new designs don't fit it many cases shows just how foolish they are.

Do what you want, it's your PC of course, but know that there are better solutions that are more robust and durable and won't risk your hardware.

Such as? If you recommend something, be prepared to be proven wrong through benchmarks galore.

Realize that a motherboard is a lot more solid than you think it is, and the bracket system won't cause a lot of flex on the motherboard, due to the way they're designed. If so, we'd hear about people's NH-D14 falling off, or mutilating their CPU socket/motherboard. I haven't heard of anything like this so far.

Also, really, heatsink designs AREN'T more complicated than that. It comes down to the basic principal of heat disipation. The easiest way to prove this is through water cooling radiators. A 120MM radiator will exhaust heat effectively, but a 240MM radiator will exhaust more heat due to an increased surface area, and this holds true with 360 as well, and so on.

[Poodles]

Yes, it actually is more complicated than that. The math behind it is extremely complex so I won't get into it, but as the contact area with the CPU can't be increased, no metal is a perfect conductor (i.e. it has resistance to moving heat through it), and variables over airflow and ambient, like I said, it's not necessarily true.

Also have to deal with the issues in heat pipes, but that complicates matters worse. There's a reason most server coolers do not run heatpipes still...

If you have a setup that allows the motherboard to lay flat, large tower coolers work fine. I would never move the PC assembled though.

[Cantii]

Yes, it actually is more complicated than that. The math behind it is extremely complex so I won't get into it, but as the contact area with the CPU can't be increased, no metal is a perfect conductor (i.e. it has resistance to moving heat through it), and variables over airflow and ambient, like I said, it's not necessarily true.

Also have to deal with the issues in heat pipes, but that complicates matters worse. There's a reason most server coolers do not run heatpipes still...

If you have a setup that allows the motherboard to lay flat, large tower coolers work fine. I would never move the PC assembled though.

Why don't you back up what you're saying with CPU HSF suggestions and benchmarks of real-world (or synthetic) testing?

There really is no complex math behind a CPU cooler. The more surface area of the cooler, the more heat is dissipated, plain and simple. If that CPU HSF has a perfectly flat surface, as well as the CPU (read: lapped to a flat, mirror finish), that increases surface area as well as allows for more metal-on-metal contact, which also increases cooling efficiency.

[Saithes]

Also, the Hyper 212+ is 626g aka 1.3lbs :P Not 3lbs! Servers use lesser heatsinks because its all they need. No one in their right mind would overclock a vital server so all they need is a basic heatsink that will suffice the TDP. Not to mention the ambient in server rooms or datacenters is a constant low without any major fluctuations. Goodluck overclocking on any passive heatsink for a Server.

[Nab]

I got the CM Hyper 212 + and my temps never went higher than 60C and I'm at 4.2Ghz. Not much but it's a 1Ghz difference from stock.

[Plasmon]

... There really is no complex math behind a CPU cooler. The more surface area of the cooler, the more heat is dissipated, plain and simple. ...

Maybe you're getting that idea because they all look kinda similar these days, but...wow, I can't believe you said that... clearly you've never taken heat transfer at the university level. Not that there's anything wrong with that, but if you did you'd understand how wrong it is to assume such an extreme oversimplification of the heat transfer mechanics. It's almost insulting to the people who put their PhD's and their careers into this kind of stuff, and the engineering profession.

The rest of my discussion has absolutely nothing to do with the original topic, but I want to say a few things about what's behind CPU air-cooler design since the topic came up and there is a dispute about large coolers.

The amount of math and science behind the optimization of a modern CPU cooler is massive and extremely complicated. In a typical undergraduate heat transfer course the most you learn how to calculate is using the most basic heat sink (like a passive rectangular one attached to a north bridge 8 years ago) and the math involves so many assumptions and simplifications that it is near useless in the real world when applied to modern CPU coolers. Those basics is about as far as I've done in heat transfer classes, but I'm fully aware of the additional complications that make the math impossibly difficult.

First, the selection of the material is complicated, it's not just "pick the highest thermal conductivity" when you have to consider material cost, corrosion properties, mechanical properties such as elastic and shear modulus, toughness, creep, thermal expansion, fatigue, weight, manufacturing properties such as formability, ability to be molded, extruded, bonding to other materials and any diffusion issues that may arise at that interface, etc. There's also no such thing in the real world as a part/component made of just "copper" or "aluminum". Which alloy? There's dozens, if not hundreds and they all have different properties.
You see a lot of copper for heat pipes and the contact base, but not often for large cooling towers. One of the reasons is because copper is 3 times heavier than aluminum, so a large copper tower with the same fin efficiency would be very heavy in addition to being much more expensive, and less corrosion resistant (why it's often nickle plated). Material selection involves math and the choice of material has ramifications for the other aspects of the design such as the thickness of each cooling fin.

It's definitely not just surface area that matters, although yes it's very important. You could make a heat sink with a massive surface area that would completely fail because of the many other properties and trade-offs that matter just as much. Different heat sink fin shapes and fin arrangements have different fin efficiencies even if they have the same surface area. Thickness of the fins matters; thicker is actually better when considering individual fins, but thinner allows the same amount of material (cost, weight) to be distributed to more fins. So there's a trade-off and an optimum thickness has the be determined. The mechanical properties are also dependent on fin shape & thickness and the cooler has to be designed not to break, bend, or crush easily.
The spacing between each fin is not obvious either because a wider gap is better when considering the heat transfer from individual fins, but finer spacing allows for more fins to be packed within the same volume. The trade-off here is the effect that it has on the airflow, which is not simple either... the field of computational fluid dynamics says hello. By making the spacing thinner, less air can fit through the gaps, the smaller volume of air per fin has less heat capacity to absorb heat from the fin, it's more difficult to coax the air to take that difficult path through the fins instead of around the heat sink, it takes a lot more energy to move air through which means more power fans which increases the noise, cost, and wear on the fan bearings, the presence of dust also becomes more detrimental in the smaller gaps, you've also gotta consider the effect of the fin design on the Reynolds number of the airflow, etc. All of these factors make the math involved in the design of a CPU cooler to be extremely complicated, to the point where calculations are impossible and everything is instead approximated with empirical relations and solved by expensive finite element 3D mathematical modelling software.

Heat pipes are no simple design either, they aren't just solid copper tubes. They are hollow and typically contain a fluid inside that changes phases from liquid to gas during it's temperature cycle to take advantage of the latent heat of vaporization. How do you determine the dimensions of the pipes? Should some be different to others? (see Super-pipes on Twin Frozer coolers) What shape? How far away before meeting the fin array? (clearance issues with RAM & other components) How many heat pipes are ideal? How do you attach them to the base and to the heat sink fins without creating thermal contact resistance between those material interfaces? What arrangement of the heat pipes within the cooling tower will maximize heat transfer to the fins and not impede the airflow too much? These questions are also solved with complex mathematical modelling.

Design of the fans? That's not simple either, but I think I've made my point and should have stopped typing long ago. Just look at the websites of the best fan manufacturers for some insight into the innovations in bearings and fan blade design, all to maximize airflow through fin arrangements while keeping audible noise to a minimum.

Something to consider: Imagine working for Noctua as a design engineer and your boss asks you to design a higher performance CPU cooler without increasing product cost, without increasing audible noise, without requiring a major overhaul of the factory manufacturing process, without blocking the RAM slots or other motherboard components, and without being too big for common case designs, does that sound like an easy problem to solve with no complex math involved?

How can you say that hanging a 3 lb weight off the side of a silicone board is at all smart?

Silicone... is not even close to what motherboard PCB's are made out of... but ignoring that, you're still wrong in thinking that motherboards aren't tough enough for modern air cooler towers. The laminated composite materials used to make PCBs results in an extremely tough board and it takes a lot more than a couple pounds hanging off it to cause mechanical failure. I guarantee you are applying way more stress to the motherboard than the CPU cooler does when you press down to seat the RAM into it's sockets or when you pull out a RAM stick, when you flex the board to force the holes to actually line up with all the standoffs, when you push in or pull out the 24-pin connector, and when you screw in the CPU cooler and it's brackets. Also, don't forget that these big tower coolers are attached to a large metal bracket which is firmly held in place by a large metal backplate behind the motherboard, so there is some heavy reinforcement support distributing and taking the load. I think the only likely case where a fracture would occur after hanging a CPU cooling tower is if there was already a smaller fracture in the PCB for whatever reason... Maybe due to mishandling during shipping, abusive installation practices like over-tightening the screws or putting your weight on the seated motherboard, etc. Crack propagation requires much less stress than the initial fracture does due to stress concentration and structural weakening.

I'll admit that it's possible to conceive of air-cooling designs that aren't as heavy, put less stress on the motherboard, and don't reach out all the way to the case wall, but what we've got now is the result of modular design. The cooler is designed by someone else as a separate part but still has to fit a wide variety of motherboards and cases, must do so without blocking the RAM or any other motherboard components, must operate near silently, and all for a reasonable price. If computers were less modular (not that I'm advocating for this at all), then the motherboard + CPU + cooler + case could all be custom designed to something that looks completely different. But there are far too many companies out there making different kinds of parts for the industry to make a massive paradigm shift away from modular designs of desktop PCs any time soon, and many consumers and OEM builders probably wouldn't be happy with that concept either. It works in portable devices though, but that's because it's necessary to achieve the small profile and high efficiency that define smart phones, netbooks, tablets, and even laptops.


/end nerd rant. I should have gone to bed instead of doing this... oh well.

[Fuzzykins]

HOLY SHIT FUCKING WALL OF TEXT

Almost all air coolers out right now are the same, as already said. The cooler that wouldn't stress the motherboard would be almost impossible to create due to the fact that you still wouldn't get the surface area. I realize it's not the only factor, but it's one of the most important. The contact of the cooler, the material, the spacing of the heatsink fins, etc, all have an impact as well, and I'm not saying they don't. However, what I am saying is that heatsinks at this point operate under the general principal that as the price increases, the performance increases, and the heatsink area increases. If you want to dispute this, think about the NH-D14. It's essentially 2 CM Hyper 212+'s for twice the price.

If you've seen Ayako's pictures, the heatsinks coming out in CompuTex are fucking massive. I mean, 3-4 separate Cooler Master Hyper 212+'s daisy chained together. This has been the trending pattern for awhile in heatsinks.