My WB: EK-MCW

[Eddy_EK]

The pins at my last picture block are 1,5mm ×1,5mm amd 2mm heigh. As I said before I used Hydor L30 pump.
For now I have estimated the difference in waterflow by look. I could see differences in water flow.
I have head that lumpy channel and R-type WB had different base thicknes. Is that correct and how thick they were?

[Eddy_EK]

OK, now I know, why do you think that pins are better.
Because R-type has preformed better than Lumpy channel WB.
But there are diferences that are preventing to pompare WB and make conclusions out of it:

1. R-type WB has higher pins (6mm) than Lumpy (4,5mm)
2. Lumpy channel has too thick base (1.5mm)
3. Lumpy channel was more restrictive

So my conclusion is that if you compared Lumpy channel and R-Type WB with the same characteristict, Lumpy channel would bo better. And that base thickness of about 2mm would bring the best results.

And please don't think of me as a newbie who wants to be smart here... if you do...

[jaydee]

Originally Posted by Eddy_EK

And please don't think of me as a newbie who wants to be smart here... if you do...

Not at all. Newbs can't make blocks like yours. What I am saying is a properly design pin design will beat a properly designed channel design. Proof is in the numbers. The MCW6000 and the Nexxos XP are pretty good commercial examples of such. If you can beat those blocks with a channel design I will not argue a bit.

You have mentioned the differences in the R-Type and the Lumpy Channel bwhich is a perfect example of why a decently designed pin block will beat a decently designed channel block.

I am not sure why you can't make smaller pins. The equipment that made your blocks shouldh ave no problems at all with it.

I am not slagging your work by the way, just think there is a lot of room for improvment on the performance side.

[Eddy_EK]

I can't make smaller pins than with 1.5mm mill. Because it takes more time, with smaller tool. For me I think the 1.5mm tool is limit. That is why I can't male that small pins like at nexxxos xp or eaven hp.

Anyway Swiftech MCW6002 has bigger pins than nexxxos.

Swiftech has 9mm high pins and 5mm thick base. The only thing that bothers me is base thickness. For my opinion it si way too much... Although I am not sure if the picture shows real base thickness. I know that antarctica has 1.4mm thick base, and nexxxos about 1mm i guess. Of course swiftech should have more thick base because the heat must spread to bigger area...

that is for now...

[Incoherent]

Originally Posted by Eddy_EK

. The only thing that bothers me is base thickness. For my opinion it is way too much...

You need the thicker base for low flowrate performance. Some people (me) like small pumps, restrictive loops. The thicker the baseplate the better in that scenario.

[Eddy_EK]

Originally Posted by Incoherent

You need the thicker base for low flowrate performance. Some people (me) like small pumps, restrictive loops. The thicker the baseplate the better in that scenario.

You might be right...

[Eddy_EK]

OK guys!

I have investigated a little and find out that the shipping cost would be about 11 USD for 1 EK-MCW block with barbs. For more informations pleas contact me by PM, MAIL or read here .

[BillA]

Eddy Ek
after you start bringing your wbs into the US, I will go after your use of "MCW" on a copyright basis
yes, Swiftech speaking

[jaydee]

If I am reading his site right he wants $67.00US for one of those blocks plus $11.00 shipping. To expensive for what you get.

[Eddy_EK]

Originally Posted by unregistered

Eddy Ek
after you start bringing your wbs into the US, I will go after your use of "MCW" on a copyright basis
yes, Swiftech speaking

Oh, hello Bill!
I am huge admirer of yours!

Okey! If you have registered MCW as your trademark that there is no problem, and I will not argue, but as far as I have read it, you have registered only example "MCW6002" and not just "MCW". Therefore there is no conflict.
Anyway, I don't know what your MCW means, I know that my MCW means "Micro Channel Wave". Anyway, when I will sell the block I have right now on stock I wiil sure change the MCW, because I don't want any conflicts, and didn't intentionally name it -MCW. I didn't knew or knowledge the swiftech names of WB.

Jaydee: I am not responsible if the USD is so low. And am not forcing anyone to buy it, but as I saw that some are paying a lot more money for G5 WB, I am feeling kind of funny. :shrug:

[BillA]

no problem, I do not doubt your honest intent

[Eddy_EK]

Originally Posted by unregistered

no problem, I do not doubt your honest intent

I knew you are good person!

Anyway if I got you here, I would like your opinion of my WB.
I know you can't tell much by photos... You can tell me by PM if you don't want it public

[BillA]

Ben here has the calibrated eyeball
looks good, should be 'ok'; doubt will 'best' (or equal) the leaders

[bigben2k]

...and my reticular implant says...

The first design is the best. I disagree with JD about pins versus fins. The "finnage" extends the effective surface area from the baseplate, which is why it works so well.

Now, it might be possible to use a different pin arrangement (i.e. spacing) to obtain similar results to the first one, but I believe that the first design will always outperform pin-style blocks.

Consider the WhiteWater, Cascade and the G5 blocks: the fins are continous throughout the core area. (not obvious on the Cascade/G5, but it's there).

More info here: http://forums.procooling.com/vbb/showthread.php?t=6101 (where it all started).

[jaydee]

Originally Posted by Eddy_EK

Oh, hello Bill!


Jaydee: I am not responsible if the USD is so low. And am not forcing anyone to buy it, but as I saw that some are paying a lot more money for G5 WB, I am feeling kind of funny. :shrug:

And what on earth makes you think your block is anything even remotely close to the G5? The performnce of your block is consistant with $39.00 blocks not anything close to the G5. That is according to your Antartica comment.

Ben, There is more water to copper surface area with pins. There is more copper mass with channels. When you mill a groove to make a pin in a channel block there is additional surface area on the base. More copper Mass is not what you want. You want as much water to copper surface area possible. This is the very same concept the G5 and Cascade follow hence the cups in the base. By drilling down he removed copper mass and replaced it with copper to water surface area. Copper to water surface area combined with impingement is why the G5 works. Also the same reason the Nexxos XP works so well.

[BillA]

ah jd, now you're doubting Ben's eye as well
good to see your critical organ is well

[FL3JM]

The fin vs pins argument is kind of hard, since it all depends on what designs you use when you compare the two diffrent styles. Ofcourse you CAN get a bigger effective surface area with pins no matter what fin block you put up, as you can just slice the fins in it, thus making it more of a pin style block. But at some point you will have too little copper in the WB taking the heat out.
Lets say you have the MOST OPTIMAL design in a fin style block with a certain amount of surface area. If you go by the rule "more surface area = better" then you can take that fin style block and chop up the fins so that it gets more of a pin style block, but you HAVE TO take into concideration what the negative effects are aswell! If you take that optimal fin design and chop the fins to resemble pins you take away copper! Don't forget that the copper is needed for an effective thermal conductance to the water, if you chop the fins to make pins of it (to increase the surface) there is less copper in the block, thus actually decreasing the performance rather than increasing.
Now this theory only works if you have a very very good fin design!! Take a less well designed fin block that has too thick fins and chop it in to pins and of course it will perform alot better!
The same goes to pin blocks, i think lots o blocks out there with pin design would perform better if it had fins instead, because the pins are just too thin.

In theory i think the optimal pin design block will have exactly the same performance as a fin design, and i also belive that they will have allmost equal surface area aswell.

Please comment my thoughts, i just had to ventilate my brain a bit, lemme know what you think.

[bigben2k]

Originally Posted by jaydee116

...Ben, There is more water to copper surface area with pins. There is more copper mass with channels. When you mill a groove to make a pin in a channel block there is additional surface area on the base. More copper Mass is not what you want. You want as much water to copper surface area possible. This is the very same concept the G5 and Cascade follow hench the cups in the base. By drilling down he removed copper mass and replaced it with copper to water surface area. Copper to water surface area combined with impngment is why the G5 works. Also the same reason the Nexxos XP works so well.

True, but there's a fine balance between surface area and the mass (of the fins).

As with any heatsink design, the purpose of a fin or pin is to effectively extend the surface area of a flat baseplate. (sorry for the rambling, I'm just adding that for everyone else's benefit)

A pin-style heatsink is nothing more than a continuous fin design that's been chopped up. Chopping up the fins has one effect: the temperature gradient will increase up the pins, as less of the heat is able to leave the baseplate. When the temprature gradient increases, the performance drops.

Consider this: if you start with a flat plate, and figure the area that is extended by fins/pins with both designs, you'll find that more of the baseplate is extended with continous fins. (hint: there are some interesting things that can happen from looking at it that way).


The Cascade/G5 "works" because it maximizes turbulent flow against the base as well as the fins, just about as efficiently as can be had.

I'd like to see the efficient flow geometry of Cascade/G5 applied on a pin-style block design, but I really doubt that it would beat a continous fin design. It ought to be close though.

Here's another crazy thought: an "air" version of Cascade/G5. (the fins would be 3-4 times thinner).

[bigben2k]

Right on, FL3JM. I think your explanation may be clearer than mine... (it just gets fuzzy after a while; I'm ready for a beer now...).

[jaydee]

Ok guys, I disagree and that's were I will leave it.

[maxSaleen]

I think we can look to our aircooled counter parts for a solution to this debate. First let me say that I agree with JD, but I'm going to save my reason for later.

Go over to frostytech.com

Compare the Alpha PAL-8592 to any "fin" heatsink you can find on the list. Just make sure you discard the ones with heatpipes as they have an advantage that is... how can I say this... almost mechanical. Just use some common sense when looking at the charts (take into acount airflow and noise level).

Now for why I agree with JD. We are dealing with water cooling. Water has a very high specific heat capacity (1cal/gram degree C). Copper has a very low specific heat capacity (.0093cal/gram degree C). The goal here is to get the copper to transfer its heat to the water as quickly as possible. From that respect we would like to have as much surface area as possible with the least amount of copper possible. This is why direct die cooling is so effective (though I must note the inherent risks and impracticalities).

I made it sound simple though it really is not. There is a point where too many fins or pins will kill flow rates and performance.

[bigben2k]

Originally Posted by jaydee116

Ok guys, I disagree and that's were I will leave it.

I think that all we were looking from you is that you recognize the limit on the other side. Otherwise what we have is "you can't have enough pins!".

Curent heatsink designs are mostly chopped fins; there's an argument on your side!

[BillA]

several of you are confusing economics (existing products) with intrinsic performance capability
consciously designed products have a price/performance 'target'
all are not race cars, though all will make that claim

[Cathar]

Actually, by volume, copper has a heat capacity not that far removed from water.

Water has a specific heat capacity of 4186J/kgC, while for copper it's around 385J/kgC, but copper has a density of 8.9.

So per 1cm³ of each item, water sits at 4.186 J/C, while copper is actually at a pretty healthy 3.427 J/C.

None of this really matters when we're talking about conduction and convection of course, but I just wanted to highlight the point, that given two equal volumed bodies of copper and water, the copper holds about 82% of the heat capacity of water, and really is right up there in terms of offering one of the higher volumetric based heat capacities of most any material you care to name, mostly due to its very high density.

[jaydee]

Originally Posted by bigben2k

I think that all we were looking from you is that you recognize the limit on the other side. Otherwise what we have is "you can't have enough pins!".

Is that a wrong conclusion?

Curent heatsink designs are mostly chopped fins; there's an argument on your side!

And the best one's are pins aswell.