Cooling My Athlon64 Waterblock!

looks great

 

It's almost been finalized that the base of the production models will be Cu!!!

 

If the top is still going to be aluminium, please make sure it is VERY well anodized, otherwise youe will have terrible corrosion problems.

8-ball

 

8-Ball = back from the dead

MSN me if you've got a spare minute

 

its basicaly a contoured top 64 block copy of Cathars design







btw cathar is the original designer of the whitewater block and his newest block is the SSC ( silver special cascade)

 

Thorilan I resent what you have to say about my block. Sure the R0X0R Block does share some similarities with Cathers block but this comes from design parameters and the laws of thermodynamics. The three barb design has been used in many blocks because of its cooling efficiency. The dimples in the R0X0R Block in no way copy anything that Cather did. Not only is there orientation completely different but there size depth and spacing creates a completely different water flow pattern through the block. The shape of the dimples is dictated by round milling bits. The R0X0R Block was purposely designed to dissipate the large thermal load created specifically by the A64 CPU. All design decisions were based off of the specific thermal characteristics of this CPU, none other. In the end the laws of thermodynamics dictate design.

 

Ferrari,

I think this may not be the correct path to walk.

You are quite correct that the water flow in your block will be different from the flow in Cathar's cascade block. There are many reasons for this.

What I am more interested in is how you arrived at this design based on the "specific thermal characteristics" of the Athlon64 as well as "the laws of thermodynamics".

I'm just curious.

8-ball

 

I am also quite interested in hearing your answer to this my friend.

 

am i the only one who gets red X's instead of pics?

 

Red X's at work and home.

(Is this product still in the works?)

 

roger that. so i'm not alone then =P

but hey, this thread is only a week "old" so i guess it's active in some ways..
might just be a crappyhosting for the pics or something

dammit, ferrari! fix the pix

 

pics are still working for me.

 

It's a great looking block and everything, super work chap... however I don't want to sound like a pecker, but don't you think "the R0X0R Block" is possibly the worst name you could give it? I apologise if you're offended, but if you plan to market it, maybe a more apt, "mature" or perhaps descriptive name would be more ideal?

 

Could be worse.
Could be the r0x0r bl0x0r.




Pics still work here btw.

 

Sure is an interesting looking block but a few concerns the main being..........is the seal going to hold up between the top and base without the use of any "O" ring or gasket with so few bolt downs???

My main concern is the testing...........has it been tested thoroughly before you release it to the buying public????

Be nice to see some independent reviews before we start handing over £££££ $$$$$

Hope that doesn't come across as being harsh and wish you all the best with it BTW..........keep us posted!

Paul

 

Oooh, well noticed coolmiester, I didn't notice that bit about the sealing of the block. I think that needs addressing, even if only to convince the buying public.

However, I'm still a little concerned by the claims you made about it's design, (see my post above)

No doubt, the block will perform well, but I would like a few things before I would consider it.

Either an o-ring or a gasket for sealing the block.

Copper construction, since most radiators and northbridge/gpu blocks are made of copper, and it is not "wise" introducing different metals into a water loop, unless you are prepared to change the water on a regular basis. (Are you going to mention this in your sales pitch, a warning maybe? otherwise you could end up with quite a few blocks corroding.) If you are not going to make it out of copper, will the price reflect this. It should definitely be cheaper than the range of coppper blocks on the market, since copper is quite a bit more expensive and harder to get hold of in the correct alloy than aluminium.

8-ball

 

All your concerns are extremely valid! Ok first of all i wanted to address some of the design aspects of the block to make all the cather people happy. I mentioned the laws of thermodynamics and thermal properties of the CPU itself. I’m not going to go crazy here but basically the die size in proportion to the size of the heat spreader is pretty important. Since there is a heat spreader on the A64 that is another material which the heat from the CPU has to travel though. This means you can focus most of the cooling toward the center of the block (most should be centered here but quite a bit of heat will be transferred horizontally through the heat spreader). This dictated the little dimples in the center under the main inlet barb. Notice how the surface area the dimples cover is larger then the die of the CPU. Also notice how the dimples don’t cover the entire bottom of the base. The block was designed as such because as water moves through the block the volume in the block increases (If you don’t believe me ill make a diagram to better explain what I mean). Unlike most blocks where the volume inside the block is either constant or very large, the R0X0R Block was specifically designed to control the speed at which the water flows and exactly where it flows. Under the center barb the water would be moving the highest velocity if there were no extra volume for it to flow into. The dimples provide this extra volume so the water doesn’t rush right past the center hot spot thus absorbing little heat. The dimples slow the water down by giving it a place to 'rest' they also create great turbulence so the maximum number of water molecules touch the hottest portion of the block. As the water heats up and new water flows in the hot water will flow outward from the center barb until it hits the end of the dimples(the water will flow over the dimples much like the water in rapids, creating turbulence) Once the end of the dimples is reached the waters velocity has been reduced substantially. The step upward onto the flat portion of the block lowers the volume in the block from the end of the dimples to the spot where the water exits the block. If you notice once the bottom is fitted onto the top piece of the block the volume inside the is very small. This creates a sheeting action of 'thin' water where the maximum surface area of water is in contact with the block surface. During this stage in the waters journey through the block the heat in the heat spreader is removed. The sheeting action is a better way for the water to absorb heat from the HS then the dimple design which removes heat from the die as directly as possible.

Lets talk more about this step and change in volume a bit more. Since the dimples slow the water down and create a space for the water to ‘rest’ the water will stay there as long as possible. This design makes the r0x0r block differ from other block designs because the time in which the water stays in a location in the block is controlled by the shape of the block. Most blocks just create turbulence and have larger volume making it difficult to control where a water molecule is at a given time. The step up may, at first, seem like it would speed up the water but it doesn’t it keeps the velocity constant. You may be concerned that this seemingly low volume inside the block would restrict water throughout your system, and I must say even after all my calculations I was a bit skeptical the first time I tested the block. Low and behold the block preformed magnificently having an extremely high flow rate. This time management of a waters location is the key to allowing the water to absorb heat and transfer it away from the cpu.

The center barb design was chosen because of the laws of thermodynamics. The amount of thermal energy that can be transferred through any object is directly proportional to the difference in temperature from the hot to cold ends. The center barb allows the coolest water to contact the hottest spot of the cpu allowing MAXIMUM thermal transfer. Not only is temperature difference responsible for heat transfer but the amount of mater between the heat source and where you are trying to move it. The dimples minimize the amount of material between the cpu and cool water. The heat spreader complicates all of this though because as I mentioned above allot of heat will transfer itself horizontally through the heat spreader. It will do this because as you move away from the die the heat spreader will become cooler due to the fact that it wont have a thermal load being exerted on it. Removing the heat from the surrounding heat spreader is just as important as removing heat from the center of the heat spreader where the die resides. Cooling the sides of the heat spreader creates the difference in temperature we talked about previously allowing heat to quickly move away from the die in all directions, both horizontally in the heat spreader and vertically through the heat spreader directly into the block thus into the water. Encompassing the die and heat spreader with different, specialized, cooling solutions inside the block creates maximum cooling for the cpu.

Some mentioned the gasket-less design in concern. The r0x0r block was designed to have the top and bottom portions clamp together creating a water tight seal. I must say once the block has been clamped together it is really difficult to separate. The block I used to test on my t-bird used this metal on metal seal as the only water seal. As a backup the production blocks are additionally sealed with locktite gasket forming sealant which permanently bonds the two metal portions together. The block also boasts 6 screws to firmly and uniformly clamp the two pieces together. Leaks will not be a problem with the R0X0R Block.

You want some reviews from an independent source, so do I! At the moment I am in talks with several review sights who are extremely interested in reviewing the block. Finally some people complained about the name of the block. I admit it is kind of cheesy and unprofessional but while I was searching for a name I showed the block to a friend who commented that the blocked “r0x0red” and the name just kind of stuck. I apologize for the delay in this post but I have been very busy talking to review websites and finalizing certain aspects of the blocks design.

 

Most integrated heat spreaders in the past actually do very little in the way of spreading the heat. They are too thin, and provided you can efficiently remove the heat from the die area, it won't need to spread.

I'm not too sure about the "sheeting" bit, since it is not a better way to remove heat from the heat spreader. No matter how you look at it, a flat surface with water flowing across it is the worst possibe geometry for heat exchange.

However, this doesn't really matter, since this region won't be removing as much heat as you think.

Also, with the dimples, there will be a great deal of interaction and turbulence in the water at the opening to each dimple. I think you may struggle getting water to enter the dimples. If you have a force of water pressing down on the cavity of the dimple, which is already full of water, where is the existing water going to go, to allow new water to enter.

Essentially, you will get a lot of stagnant water at the base of the dimples.

So, how can you state that the dimples are so effective at removing heat.

I'm sorry for doubting you like this, It jus seems like you haven't thought this through entirely.

The ONLY reason the dimples work in Cathar's block, (and this is just about the only successful block I have seen to use dimples) is because he solved the problem of getting water to the bottom of the dimple while allowing a free path for the water to escape. Now, I see neither of these features in your block, which makes me concerned.

Another thing for you to consider is that a new processor does not require any new designs of cooling. The only thing that needs changing is the mounting. As far as cooling, the same principles apply regardless of the heat load, die size or anything else you care to consider. There are basic rules governing the performance of a heat exchanger, and they don't change from one processor to another. Previous waterblocks are perfectly adequate, all that is needed is upgraded mounting hardware.

Sorry for being obstructive

8-ball

 

Very nice work man. Any advice on manufacturing a block? I've got a Pentium 4, 1/2 Od one in planning. Also, what keeps the seal between layers?

 

mm from what I've seen of Cathar's Cascade block this one looks like a half assed effort to knock off the design...

They BOTH have the distinctive dimples, and the same internal shape....
http://forums.overclockers.com.au/showthread.php?s=&threadid=169230&perpage=15&pagenumber=1

But given the choice I'll take Cathar's block..... He has proven track record with his White Water, LRWW and Cascade SS blocks.
http://forums.overclockers.com.au/showthread.php?s=&threadid=169230&perpage=15&pagenumber=1


[EDIT You may also like to take note of the date of his post 14th May 2003