Covering your entire case in heatsinks / building a custom H1.S clone frankenchild.

[Phryq]

So I'm wondering, if I used some heatpipes going to the outside of the case, and covered the case in something like this,

Or, simply made the entire case out of heatsinks (I don't really know how I'd do it... I can't weld or anything like that). Wouldn't that be the ultimate passively cooled case? Couldn't you just stick a fan on top of it, and it would be extremely cool? What am I missing?

I'm looking at some heat maps

and it seems to me there are 2 main 'bottlenacks' with thermal dissipation.

First, from the CPU area into the heatsink (in this case it's water cooling. I can understand that the heat will be cooler entering the water, but we should see more than what we do, no?)

In this picture however it seems I'm wrong,

And the heat is moving effectively into the cooling system.

Here's a Before/After adding heatsink comparison
Before After

Though it doesn't really make sense to me. It looks like even the air around the computer is hotter withthe heatsinks added. Is it a matter of hot-CPU=less-efficient and so produces more heat altogether?

And here's another idea

I wonder how difficult / expensive it would be to make something like the H1.S for a Micro-STX from scratch. Aluminum is cheap

The side panels,

http://www.ebay.com/itm/Aluminum-Heatsink-Larger-8-7x1-6x1-6-Projects-LED-Audio-Power-/262926918733

The lid,

https://www.rfparts.com/heatsinks/hs100-6.html

But I guess the H1.S has grooves in the sides to fit the heatpipes into, so heat transfers better into the sides. Maybe it's possible to buy *only* the fins / heatpipes for the H1.S and build your own bottom/top/side plates.

 

[|||]

You may want to take a look at this project.

The thermal pictures you posted are correct.

For the case, the water cooler is about right. The top is probably a plastic and doesn't radiate much heat, as a result. The tubes are red, as they are being heat-soaked. Radiator is orange, as it has a fan or two mounted on it blowing air through it. The white part on the board is hard to tell what is causing it, but it could possibly be a PLX PCI-e switch depending on the motherboard, which would generate substantial heat.

In the Raspberry Pi instance, the scales and color mapping are different. With the heat sink on, the relative difference in the temperature between the processor and the components on the board is less, so it compresses the color scale to show variances at a finer granularity.

 

[zovc]

Streacom makes a lot of fanless cases that use the actual chassis as heatsinks. Hardware Cannucks recently did a video of trying to overclock in one of the cases, they even bought a kit to use additional panels as heatsinks.

The main takeaway I got from that HC video was: no matter how good your passive cooling solution is, actually dissapating heat into the air will always be superior. Just having a fan blow on one side panel of that Streacom case reduced its temperatures like 10C. No matter how good your heatsink is, it'll get saturated and unless you're circulating something over it to absorb heat, it's bound to stop working well.

That said, I like the idea of using large surface-area heatsinks and I do believe that side panels are an elegant/efficient location to utilize as a heatsink. If there were a way to treat a side panel like a hybrid of an air-cooler heatsink and a water-cooling radiator (I understand they're practically the same), where you're forcing air through a small 'heatsink chamber' of your side panel, you might be onto a winning implementation of this idea.

 

[Arboreal]

Here's another project worth looking at in the same vein, with passive cooling and a BIG heatsink

 

[Phryq]

Zovc, I watch that's Cunuc video, at 9:57 "with ice-cold hands, you can actually cool the CPU down by touching the panels". Haha, this is definately me, pushing my computer and trying to cool it down with my cold body parts.

Very good video though. So the 'bottleneck' is actually from the outer-case to the rest of the world... so putting the case in a windy environment, or just blowing some external fan at it, would give some serious cooling, yes? If this is really the bottleneck, the simply increasing the case surface area will have the largest impact on cooling.

Making a case entirely covered in fins shaped like Arboreal's like above

^^ Would give the highest surface area. Maybe something not quite as thick, but have the rivets a bit wider

Here are 2 options, 1 with a fan, and 1 fanless. My preference not to have a fan is more about the required holes (I would use a dust mesh, so maybe not so bad). It would be nice to keep things sealed. However, if having a fan outside the computer is just as affective, I can just buy a big fan, like a 200mm fan, and put it on top of the hole computer.

In theory this design should have 7 times the surface area of an H1.S, and double the heatpipes (stack the heatpipe blocks on top of eachother) and since an H1.S can handle a 65w cpu, this should be able to handle a 7700k and simple GPU, maybe even with overclocking, no? With an exterior fan, it should be able to easily handle any overclocking and any GPU? Way too ambitious?

Arboreal The heatsink / radiator he's using looks great. It makes me think... Aluminum has the highest conductivity per weight (copper beats it by size, but alumunum wins because it conducts more per weight). I wonder if a radiator like his would benefit by having those 'wobbles' even wider... I guess it doesn't matter as no one makes something like that.

 

[zovc]

Very good video though. So the 'bottleneck' is actually from the outer-case to the rest of the world... so putting the case in a windy environment, or just blowing some external fan at it, would give some serious cooling, yes? If this is really the bottleneck, the simply increasing the case surface area will have the largest impact on cooling.

Kind of. I think you might be getting confused a little bit. Increasing your surface area makes it easier for you to dissipate your heat. However, increasing your surface area does not dissipate your heat. The reason heatsinks have fins is not only because they are an engineering solution that allows you to occupy lots of volume, but they also create lots of turbulence for air. By pushing (relatively) cool air through (relatively) warm heatsinks, the cool air conducts heat from the heatsink, this then carries heat away from the part.

Demitri putting his cold hands on the case is a similar example, the heat is dissipated from the case into his hands at which point the case is actually cooled. Prior to his intervention, the case has become 'saturated' with heat and is unable to dissipate any of it. The fan reducing his temperatures by as much as 10C is the same sort of case but it is harder to visualize because there's not the tactile/visual aid of physically cold hands and a warm chassis. The air in the room is presumably cooler than the ~50C(?) the case is at, so heat will conduct into the air. But, the case isn't so drastically warm that it will create turbulence on its own. When you introduce the fan and it starts pushing room-temperature air into the case, that air is relatively cooler so it picks up some heat as it's moving around the case.

This is me getting a bit long winded, but suffice to say: increasing your surface area isn't going to help nearly as much as introducing some form of circulation.

In theory this design should have 8 times the surface area of an H1.S, and since an H1.S can handle a 65w cpu, this should be able to handle a 7700k and simple GPU, maybe even with overclocking, no? With an exterior fan, it should be able to easily handle any overclocking and any GPU?

The issue with this design is that you're not really utilizing your increased surface area at all. Your fan is blowing against a flat panel, so it's not moving air through all those fins. You'll notice that CPU coolers are normally designed to have fans blowing through the fins of them rather than against a flat surface that'll stop any airflow.

 

[Phryq]

Right, so either a fan with something like this as the lid,

and a dust mesh on top. (I don't really like this idea).

Or my first fanless concept, and just have an 'exterior fan' blowing at the entire case.

The exterior fan should be just as effective as any typical interior fan settup, because the heat has traveled effectively to the surface, yes? In fact, it should be much more effective, because the heatsink has way more surface area. It would be as good as an open-concept / caseless computer with a giant heatsink.

 

[zovc]

Ohh, I didn't consider the gaps in the fins in the image. In the design with the fans, you'd have the top panel somehow open to allow air from the fan(s) to push through it? That's a pretty interesting idea. I think you should try to follow in that heatsink's general design direction and have your heatpipes go towards where the fan is blowing rather than to the sides like your image suggests.

I think dust filters are a little overrated, especially for an oddball design like this. If your case is designed to be this open, it's not that hard to take some canned air to it every now and again.

 

[Phryq]

Ohh, I didn't consider the gaps in the fins in the image. In the design with the fans, you'd have the top panel somehow open to allow air from the fan(s) to push through it? That's a pretty interesting idea. I think you should try to follow in that heatsink's general design direction and have your heatpipes go towards where the fan is blowing rather than to the sides like your image suggests.

I think dust filters are a little overrated, especially for an oddball design like this. If your case is designed to be this open, it's not that hard to take some canned air to it every now and again.

Yes, I was thinking to put as many holes between the fins as possible. I was thinking to have the second row of heatpipes (see how there are 2 stacks?) go to the top, however I didn't know how to draw it (or really how to implement it) so that the pipes go past the pan into the lid.

I would want a dust mesh. I hate dust, which is why I'd like the 100% enclosed idea if practical for a 7700k. The other problem is, how would you get the motherboard in in the first place. I guess the bottom would have to come off, and you would attach it upside down. If the heatpipes are attaching to the lid, and way of sliding the system out of the case would detach them from the lid/thermal paste.