So what about those water blocks?

[EliteRetard]

I see now that the ones you posted have etching on them (Ive got a 24" 2048x1536 so the tiny thumbnail didnt have detail untill I enlarged it). One even looks like its cut through? How do those work, I saw one block looked like it only had a single inlet...wheres the water go?

So if its all about high flow and turbulance then my idea still seems to stand, a large block with lots of thin dimpled fins (like I said, kinda like a heatpipe tower) so a large ammount of water can pass through but there is also more surface area to transfer heat. Im not talking mass, big difference.

And that thing about nozzles forcing water into a plate? Is that increasing pressure or something, how does that help? Dont get to technical, Im trying to actually understand this stuff lol. I can understand that a thinner material between the actual CPU and the water would be good, as that material is going to trap some heat. But you cant go to thin or it will flex when you mount it.

All of this is making me want to expiriment with my setup. My pump says 12v, 400L/Hr, 3/8" hose (inside diameter?), ceramic bearing, 70,000 MTBF, where does that rank in the grand scheme of water pumps?

 

[aigomorla]

Wow... this is going to take a long time to explain..

first off let me start here:

The difference between an air sink and a water block is the location at which the heat is offloaded into ambients.

A modern heat sink uses heat pipes to move the heat to the top where the funs then cool the heat pipes and vice versa.

The movement and amount of heat an air sink can do is based on whats inside the heat pipes, as well as the flow of the liquid, in relationship.

In a heat pipe as the liquid boils, it travels up to the fins, where its cooled down and goes back to collect heat.

Now you've seen how small and tiny heat pipes are, they are all done by capillary action.


Now for a Waterblock.
You have a medium called water.
The average custom water system will contain a lot more water then a heat pipe will.
Water has a great holding capacity.
Water is basically picking up the heat from the block and moving it to your radiator.
Since the holding capacity of water is so great, you want the least amount of resistance at the heat source to pick up heat.

Your probably saying isnt that true with heat pipes.... the answer is no.
The reason is due to the limitations of heat the heat pipes can carry before you oversaturate them.
Also since the heat pipe is done in a capillary fashion, having a dedicated pump move liquid is more efficient.
Also water works on a principle of equalibrium, meaning your water temps will not have a crazy delta thoughout the entire loop, unless you have a small nuclear reactor somewhere. 480GTX is what i consider a pico reactor...add 2 more and you got yourself a small reactor.

So less resistance @ the block end...
This means... you want a thin base... because less metal for the heat to saturate, means your entire variable will be on your holding capacity for water and your radiator.

Even flow... along the base plate... this allows a more uniform cooling pattern.
Remember equalibrium? The water inside the block will be the same.
This means short fins... u dont need long fins.. because u will have no gradiant.. You want surface area hence micropins.

Turbulance... ever blow into a hot spoon of soup?
Notice how if you agitate it with a strong blow to cause ripples, it cools down faster then you just blowing over the top?

Same can be said inside the waterblock.
More turbulance hitting the plate with even form = greater efficiency.
More molecules you got smashing into the plate and picking up heat, the greater the efficiency.

In H2O Cooling, we dont look at just the Ltr/Hr. We look at the entire pump graph to see how it reacts in relationship to pressure.
So that number is usually represented in psi, or feet head... meaning the height in which the pump can push water up vertically.

When your playing with resistance.... feet head or head pressure is what gives you a better representation, when looking at the overall system after resistance has been taken into eq, then you look at liter/hr.

 

[Schmide]

My pump says 12v, 400L/Hr, 3/8" hose (inside diameter?), ceramic bearing, 70,000 MTBF, where does that rank in the grand scheme of water pumps?

That sounds like Thermaltake Bigwater pump to me. Depending on who you talk to you're going to get a fair amount of bias towards that pump, not all of it really justified. There were a bad batch early in its release so it got a bad reputation. The long and short of it, it's a pretty good pump more than enough to drive most systems to their full potential.

If I'm right the big issues with that class of cooling is radiator. It really doesn't provide enough of it for the amount of heat created by modern overclocked systems. Sure it will reach a balance, but it will be higher than most common water cooled systems. No matter how much you experiment with the various components and flows of the system, nothing is going to make up for radiator area. Yeah there are some differences in quality, density, efficiency and flow, but those only effect performance marginally.

Relative to radiator, pump and block are not as dramatic in terms of performance variation. I am in no way saying there isn't a difference in classes of components, the best of the best will give you the best, but a decent block/pump will still give you great performance.

One of the things I've noticed about recent block design is it's really compact. You only need a small area of turbulent water action to efficiently move heat. Large fins and monolithic blocks are a thing of the past.

 

[EliteRetard]

Interesting, so if I just happened to have a heater core or something laying around...then just take a thinwalled pipe and a roll of mesh inside, flatten it out and stick it to the CPU. Connect a big ass pump and voila!

Im thinking the block I have is the main limiting factor, even the single 120mm radiator looks like it has more surface area than most heatsinks.

Whats the cooling difference between copper aluminum and steel? I thought copper and aluminum were rather close (but your not supposed to mix the two) but just curious how that relates to steel, 150% 200% better heat transfer?

 

[aigomorla]

look up specific heat...

That should answer the last paragraph in your question.

 

[Schmide]

Interesting, so if I just happened to have a heater core or something laying around...then just take a thinwalled pipe and a roll of mesh inside, flatten it out and stick it to the CPU. Connect a big ass pump and voila!

This seems a bit reductio ad absurdum to me. There is a point where a simple waterblock is unable to keep up with the heat generated from a source.

Im thinking the block I have is the main limiting factor, even the single 120mm radiator looks like it has more surface area than most heatsinks.

Not knowing what radiator you have makes this kind of hard to explain. However, if you do have a big water kit, it most likely has a single pipe design. This is far inferior to the type of radiator you see with most custom kits where they use thin-multi-pass parallel channels.

For the most part that design of radiator 120mm is good for <100w loads. (120.2 <150w loads). While a good 120mm thick radiator of the good design can handle 200w loads.

xspc rx120 example