Hypothetical question...

I was thinking about this while I was watching my toilet flush...

From what Cathar said, he always tries to make a design that enables all the water molecules to touch the cooling surface..

Working off of that, what if you had a block with a very small propeller that spins inside the WB and creats alot of turbulance...?? This will enable maximun contact between the cooling surface and the water molecules. or does it?

In thereory i see where your going but just think of how many more leak spots that could have to deal with and putting a propeller in something that small .. well it would have to be pretty d**m small it's self. I think just get'n better turbulance designs for ridges in the block would suffice.

yes that was my major concern and lets just say that it DOES NOT leak... Hypothetically

This isn't a new idea. I believe it was Volenti that was playing with a block that doubled as the impeller housing for his pump. I'm not sure he ever got it to work as well as he was expecting though.

Electrecitial engines produce electro-magnetic radiation.
If not that it would be no greater problem to put kinda fan in water block.

aaaaaaa electro-magnatic radiation I forgot about that. Ok guys Ill drop the toilet idea

I'd be more worried about vibration myself...

Vibration is no great deal - you can make proper mounting for block...

Anyone thought about direct injection? I've seen pics of direct injection phase-change block (by berkut, injection - bosh, probably from car), so tere could be some stuff like that but with water.

---Comments plaz---

Can you explain what direct injection phase-change is????

basically - a freon rig with not simple inlet, but direct injection - like in car engines (same parts tho). refrigerant is sprayed into block in some foggy form, which evaporates extremly rapidly, chilling block. Anyway pics of that gave me idea about not so extreme rigs - water direct injection.

How about a propeller power by the water coming into the block, like a windmill concept.

Ven0m said:

Anyway pics of that gave me idea about not so extreme rigs - water direct injection.

Click to expand...

Volenti's done that too. Among others.

Typically the CPU dies within days, weeks or months.

Volenti has also done the pumpcpuwaterblock too.

I have too tried direct die cooling. dead in a week, same day i was gonna take it off.

Jon

Ven0m said:

basically - a freon rig with not simple inlet, but direct injection - like in car engines (same parts tho). refrigerant is sprayed into block in some foggy form, which evaporates extremly rapidly, chilling block. Anyway pics of that gave me idea about not so extreme rigs - water direct injection.

Click to expand...

so you would use a fuel injector to do this... Thats very interesting. and use a fuel pump perhaps? I remember there was a thread about Bosch Fuelpump before... Hmmmm thats interesting. Fuel injectors are designed to atomize the fuel mixture so I dont think it will hit the cooling surface with any force like Jet Impringement but it would be interesting to see.

Basically, I didn't want direct die. If I wanted, I'd use some low-melting-point metals to cover core - similiar to direct die but harmless and more surface.

omg, I've got an idea - came just now.

- you get low-melting-point alloy, for example Wood's alloy
- very small copper sticks
- mount something around core, have stick vertically on core
- pour that metal on whole thing to cover core, but the thinnest layer possible - just to cover thing and hold sticks in place
- a) when alloy is stilla fluid, mount cover
- b) or just use some o-ring
- with direct inejction pump or just cascade-style


other version - I'm not sure if it's in English, but to use silver amalgamate
- prepare top
- place thingie on CPU, give it proper shape
- mount top

This way layer is close to direct die, there's no thermal paste - we all know that even AS5 is a thermal insulator comparing to thermal-conductive metals - and shape is more complex than direct die.

Problem with that first is that low-melting-point alloys aren't good thermal conductors, and silver thingie could be hardly removed. Still you can easily use low-melting-point metals to connect water block / heatsink to CPU.

My English cannot be considered perfect, but I think that you cna understand.

but would your cpu temps be low enough to keep these low melting point metals from melting? i mean, how low is "low melting point" and how cool do you expect the design to make the cpu? obviously, any overlap in melting temp and and cpu temp spells disaster.

sandman001 said:

How about a propeller power by the water coming into the block, like a windmill concept.

Click to expand...

Thats what I was thinking too..........it would be much easier to make

Low melting point alloys - we can assome that <200*C melting point is low.

Wood's alloy has the lowest melting point I know and it's around 70*C, if I remember correctly. There are many alloys melting <100*C, so they can be melt even with hot water and using them on CPU is no temperature risk.

for example:
make Wood's alloy powder, place on CPU, put water block, put some hot (close to boiling) water into block, let alloy melt and then become solid

I bet that it's better than tharmal compounds used now

And how good it can be? With this way of mounting block, I expect better results with all blocks - Cascade would be even better than now.

And if I'm not mistaken, rig with that silver thingie, could be better than any water block used now. I don't tell that it will be better, as blocks like Cascade would be strong competition, but it has chances in my opinion.

You definitely have some good idea Venom but I think that the risk/difficulty might seem high for some people hence causing people to stay away from it. If this can be done and is proven succesful Im taking off my IHS ASAP and going with Wood's alloy lolz

But I think the problem still lies with tep of core being too high. If the metal can be melted using boiling water, wont the core have the chance of melting it?? so I think whatever is casuing the meltdown of the metal has to be substantially higher than the core it self. If that's the case I think the difficulties of this project might be too high...

My idea - if core temp is always bleow X*C, use ally that melts over X*C, for example X+5 or more

If only I can find some low-melting-point alloys, I'll definietly try them. I think that using one shouldn't be harder, or might be even easier than applying thermal paste.

Another thing - this way heatsink is soldered to CPU, not with solder but still sticking hard. Anyway it may increase safety of transporting comp - lower risk of CPU edge damage as block doesn't move relatively to CPU.