Block project

Which one do you think will be the most effective?



PS. Sorry for quality, it was made in excel

I'd say definatly not the first one because of the right angles. The second one has more channel wall surface area but it is questionable how much heat actually convects that high up the walls. So I'd go with the last one because of the shorter streamlined water path, in and out quickly for good flow rate. They would all work acceptably but that's my opinion why which one would be best.


WELCOME TO THE FORUMS!!!!!

I would go for the first one....even if it has angles...who cares..buy a bigger pump so you get more pressure....The first one seems to have the larger area exposed of water...big chanels for water are not the best...(if they would than why not make only two chanesls ....something like a Z shape)

lol i was actually thinking the second one looked good, guess everyone has different opignions. But i would get a mix of the second and third, because the 2 has that division in the center when it flows from the center but has a too sharp of turn. I hope you understand what i mean, not sure if i can myself when reading this, if you need pic let me know.

Here's another vote for the square design. Right angles give turbulence which is good for heat exchange (it is bad anywhere else in your system.) Design one also wins because it has the most surface area for heat exchange.

few more questions:
what's better: deep'n'narrow (19mmx5mm) vs. shallow'n'wide (9mmx10mm) channel (I assume similar waterflow)?
is 1mm thickness of the base ok?

I'd go for deep and wide but that's just me. High flow rate and velocity is what causes good turbulence. High flow rate is also better in the rad so don't limit the flow at the block if it can be helped. Base thickness may be design dependant to a degree. I think some recent testing at another site found 3mm to be optimal but that was with a labyrinth design. 1mm though I think would be safe to assume to be too thin and limit heat spread in the block.

gone_fishin:
but AFAIK what we want is to transfer the heat to water (and out), not to spread the heat to the block. however i think i go 2mm . And maybe I use 18mmx8mm channels ...

Is there some dependency between the length of the channel, base thickness and and the velocity?

What I think is that if you have short, high speed channel the base should be thinner as there is no sense to spread the heat to further parts of the block. but I could be wrong

If you make your base plate very thin, the heat cant's spread troughout the block. This means that only the area directly above the cpu will be warm, this implies a reduced surface area and that's never good. in the ideal situation the entire block is the same temp as the cpu ...

heres what i think,

you make a dual channel deep and narrow, dual channel for dual surface area.....

WELL FROM MY TESTING AND I HAVE BEEN USING WATER TO COOL SINCE THE CELELON 366MHZ WERE OUT. SPRIAL IS THE BEST.I HAVE MADE A FEW MAZES AND STILL DO BUT THEY DO HAVE A DISADVANTAGE OF SPRIAL DESIGN...YES THEY DO HAVE 90 DEGREE BENDS AND NORMALLY YOU WOULD THINK THAT WOULD GIVE MORE turbulence BUT AT A COST.FOR EVERY BEND YOU MAKE THERE IS A POCKET RIGHT BEHIND THE BEND THAT WATER DOESNT MOVE AS MUCH.JUST LIKE A FISHING STREAM, BEHIND THE BEND YOU WILL ALLWAYS HAVE A POCKET OF STILL WATER.SO YOU LOSE ALOT OF YOU COOLING .SO SPRIAL WINS WITH NO BENDS MOVING WATER CONTACTS THE COMPLETE CHANNEL. NOW THE BASE NEEDS TO BE ABOUT 3MM THICK TOO THINK IS NOT GOOD.3MM GIVES THE BEST BETWEEN THICKNESS OVERALL TO USE.
I HAVE A SITE I BUILT A LAST YEAR SHOWING SOME OF MY COOLING HARDWARE. I HAD TO LEARN THE HARDWAY IN THE EARLY DAYS OF OVERCLOCKING.
http://www.fortunecity.com/skyscraper/perl/709/

A Journey to 2.0 G MHz

Click to expand...

Mines almost Identical to the top block and it does just Grooovy .. hehe

Of course with a hand mill I didnt have the option of curved channels .. sigh ....