Specs:
Intel Core i3-2105 CPU w/HD3000 graphics
Foxconn H67S motherboard
Scythe Kozuti HSF w/80mm fan
Crucial 256GB M4 SSD
8GB Crucial DDR3 system memory
150W mini-box.com PicoPSU
Microsoft Windows 8 Developers Preview

More here...
http://slipperyskip.com/page33.html

Impressive. As always.

A stylish case indeed!

But could you provide temperature readings? Many wonder about running Sandy Bridge dual-cores fanless but few do...

Thanks Vicotnik!



Thanks. The case is fanless but the Scythe HSF uses an 80mm fan (it's hidden underneath the heatsink).

Oh well. I was hoping somebody finally got around to trying fanless...

Modern Intels work OK with a single fan. In my experience you only need a good fanless case with the stock low-profile heatsink. Your heatisnki is way better.
It'd be interesting if you tried your case with a fairly common fanless board one day to compare the efficiency of your design with the commercial fanless cases.

I'm working on a design similar to what I did with my Pico Bayard several year ago.



By exposing the heatsink to the outside of the case I'm thinking that going totally fanless might be a little more feasible. If not....it might look cool.

A marvel to look at, and a little draft is always a good idea. If the motherboard is good at regulating the fan the sound level ought not to be a problem.

Where does the drive sit? Edit: Oh I see, under the motherboard. Clever.

Though this system featured in the top of this thread is not completely fanless, there are people who made them, this guy for example provides all the details of his system including components used and temperature readings: Silent Mini-ITX build

Thanks for the link.

The system in that link is a god awful looking piece of crap. Compromises have to be made for silent computing but I don't think suffering that kind of comical ugliness should be one of them. Visual pollution versus sound pollution I guess. I think I would rather turn up my Bob Marley a little than have to endure that.

Yeah, it's preposterous. But it has a kind of zany dragster appeal, I think. Kinda cool.
The trouble is that it's probably unsafe. I wouldn't want this on my desk and certainly not as a server.

The builder wanted to run a 2100 at full speed. But you should really be underclocking (and undervolting as well ideally) to run a mainstream desktop CPU fanless. Then you wouldn't need such a huge heatsink.

I have had in mind an industrial design that integrates heatsinks as part of the theme. Shouldn't be too hard to do. I would mount the motherboard upside down so that the heatsinks hang down into dark shadows. Use part of the support structure to block off unwanted views. From what I understand, Ivy Bridge will make fanless cooling much easier.

I'm not sure defying gravity is such a great idea. If you're not using any fans, convection is important.
My "thermodynamic intuition" has failed me berfore but I'm afraid the CPU wouldn't cool well and would cook the board if you forced hot air to rise against the board instead of above it.

The way to do it properly if you want to put in the work is to use the metallic parts of the case as a heatsink by using heatpipes to link a smallish CPU heatsink to large plates (or a full shell) surrounding the board. Then you could cool a powerful CPU without any fans.

"...cook the board if you forced hot air to rise against the board instead of above it." A bit of an exaggeration don't you think? The footprint of a mini-ITX board won't impede much of anything let alone a little warm air. To be honest, you are preaching to the choir because I have been ridiculing the whole "front-to-back" fan positioning and airflow scheme for years.

Most of my work has been based on leveraging the ever improving thermal efficiency of each generation of CPUs to do things that have never been done before. Heatpipes to finned aluminum casings are cool but it is expensive and...well...it's been done. I have lost count of all the times naysayers have told me I can't do something.

You'd be amazed what can impede airflow if you have no fans, like a mesh with holes which are a little too small.
Like I said, I've been wrong before. You've got to try if you want to know how many hours it would take before temperatures stop rising. But my advice is to test the principle with a cheap board...

What is the difference between natural hot airflow hitting the motherboard and that of a downward draft HSF that exhausts hot airflow directly onto the motherboard? My guess is...not much. Some would suggest that the airflow is good for the cooling of certain motherboard components.

On another forum I posed the question of whether I should reverse the Scythe Kozuti fan to exhaust out of the case instead of onto the motherboard like it was designed. The consensus was no. Remember that the Kozuti fan is beneath the fins/heatpipies.

I have built two custom systems that mount the motherboard upside-down and have inadvertently tested both of them for CPU fan failure (don't ask). My experience didn't result in any "cooked" motherboards.

Did these systems really have no fans whatsoever and did they run for hours?


The temperature of the air obviously.
I suppose it doesn't heat the board faster at first. But the heat builds up over time...

Mike has more experience with this stuff than I do (I run low-power hardware without fans, not hardware designed to be used with fans) and he says it got an excessive board failure rate under ideal conditions. But maybe less-than-ideal conditions don't kill boards much faster.

The total amount of energy (heat) directed at the board is the same regardless of the temperature of the air leaving the heatsink. The lower temp is an indication of the efficiency of the heat exchanger. A 65W CPU will produce the same amount of heat regardless of how it is cooled. Turning a fanless motherboard upside down shouldn't be anymore detrimental than at downward draft HSF.

The amount of heat is the same (not quite: there's also a positive feedback with inefficency but let's ignore it). How the heat is dissipated isn't the same.
The equilibrium temperature is not a simple function of the heat generated but the temperature at which the amount of heat dissipated reaches the amount of heat generated.

The temperature of the air flowing from the heatsink is a function of airflow as well as heatsink temperature.
Fast-moving 40C air will stop heating a part before it reaches 40C. Slow-moving 60C air will keep heating it up. Of course it's not going to heat the part as fast as if a fan was somehow blowing 60C air but the equilibrium temperature will be higher. (Numbers are for illustration only.)

That's because fans cause more a lot more heat to be dissipated through the temperature difference between the air entering the case and the air leaving it. Without fans you dissipate more heat through infrared radiation and air cooling of the outside of the case which implies a hotter case (as well as hotter parts).
Airflow in the case plays a part in determining how hot individual parts will be. You want to keep the board as cool as possible relative to other parts of the system because some parts like the top of the case are going to be unusally hot without fans.

LOL

Interesting debate.

This is what it's all about in the near future. Design for fanless operation. When Ivy Bridge arrives I will go fanless and I suspect I will never go back. The main potential problem with fanless operation will be how the motherboard will cope with close to zero airflow. I would hesitate to mount the motherboard upside down out of fear of heat pockets.

Heat pockets? Might be a problem if it were mounted in a standard case but my cases have never been deemed "standard". It is not the ideal setup but I don't think it is the end of the world either. Think about the conditions a laptop motherboard operates in. Not ideal but they aren't made of ice either.

For the record here...just because I build pretty things doesn't mean I leave my engineering degree at home.

Very nice work, slippery... once again !

You rule ^^