so what block is top dog?
 

well whose the best :)

with a 370gph pump, heatcore, and 2 lower flow fans, and 1/2" tubing

Are you in USA or Europe ?

the Q should be commersial, or home brew....


Home brew RULES bigtime...

I dont care as long as its not that expensive and i can buy it from the us. I cant make it because i dont have access to the neccessary tools and basically dont want to put in the effort.
Yes I know im lazy :shrug:

*edit also, could i expect very nice temps from this setup, to oc furthur then my pal 8045 and 40cfm fan?

You can expect lower temps.
OC'ing further is a different problem... If you've reached the max with your Alpha you wont get any further with watercooling.

"Top dogs" in EU: K4.2, CuPlex, Innovatek's "big block".
"Top dogs" in US: TC-4, Maze3, Spiral, and the latest from Swiftech (heard only good things about it)

If you want all 1/2" forget innovatek's block.

You can get the Maz3 in the UK. Also have a look at the Neptune blocks

maze3 and mcw478 have to be the top two commercially available blocks.

for pure performance, see #rotor, fixitt, or jaydee

The best block I have used is the Innovatek rev3. It is especially good at lower volumetric flow rates. I haven't tried a Maze3 yet though.

here you can see the transition

http://thermal-management-testing.com/462UvsInnova.gif

for pure performance listen to hype

that is very interesting data Bill. I have the -UH top on my Swiftech block, and it seemed to be a bit less impressive at higher GPM than the -U apparently is. Is that "stock" or with 1/2" barbs on the Innovatek and with a little piece of copper pipe in the swiftech?

Can I infer that these blocks are both superior to all the labyrinth-types?

as ever, tons of data
but often not quite comparable

STILL trying to find the secret to TIM joint consistency - but closing fast

I do not have a 462-UH, but I sure would NOT drill my -U out
I cannot (yet) demonstrate it, but I'm rather sure that the flow rate/velocity tradeoff is not advantageous
(I did some testing with a 0.625in. inlet on a 462 and it was TERRIBLE)

N.B. Swiftech will deliver what the WCers cry for; a sale is a sale

the fittings used do not matter, eh ? (so long as the wb is not otherwise modded)
- the flow may change but thats what the x axis is all about
[but yes, they were as you surmised)

I do believe so (but cannot quite prove - yet)

Until I learn otherwise I equate the DAI cooling to the "thrust" on the die area . From simple "rocket science " I would expect this to be maximum wth a nozzle the same size as the die .For a 10x10mm die,possibly, a convergent/divergent 10mm/11.5mm diameter nozzle (ye old tapered barb) would be optimum - http://www.grc.nasa.gov/WWW/K-12/airplane/ienzl.html

I would suspect a rocket motor is a bit of a strech as there is no coolant expansion

the tradeoff with a conventional nozzle is simply the head loss associated with the increase in velocity
(due to the reduced cross-section)

for WCers, again, the pump is the limiting factor

BTW, there are companies selling such wbs for chip cooling

I agree with BillA, the rocket model is for a gas, in our case, we're talking about an essentially incompressable liquid.

There are nozzle designs specifically for water/liquid, and they are spec'ed by ASME standards.

Finally managed to view the image(been plagued by the dreaded red cross)
Intrigueing to have a comparison between the heat-spreading of a 30mm(guess) diameter cylinder of the Innovatek with DAI of the Swiftech.
Am left pondering whether the different flow rate response is related directly to velocity or indirectly due to the lower spreading resistance (at low velocities) of the cylinder/bp combination.
Bill,please, do you have an estimate of the bp thickneses?

the designs are nice to compare as they are the opposite extremes
max fins close-coupled to no fins at all

both bps are thick, the 462-U is 0.31in. and I would guess the Innovatek to be something similar
- but no way to measure other than estimating from that old THG image

my suspicion is that:
in the Innovatek the higher flow is passing through the upper fin (disk) area where it is of little benefit
whereas
in the 462-U the higher flow contributes directly to "better cooling" (a lower thermal impedance) via the improved convection rate
- hence also its relatively poor low flow rate performance

to return to your question:
don't think that spreading resistance accounts for the principal differences between these two

and

Yes, but is it any bigger stretch than trying to use a conventional Reynolds No. related cooling effect.
Some "Spreading Resistance" play with Zero TIM Resistance using Waterloo and Kryotherm :-
c/w versus Reynolds No for a 80x80mm plate with 10x10mm heat source

http://www.jr001b4751.pwp.blueyonder...Reynoldlsa.jpg

from Convection coeff v Reynolds No (Kryotherrm):-

http://www.jr001b4751.pwp.blueyonder.../Reynoldsa.jpg

and using the Waterloo Calculator:-

http://www.jr001b4751.pwp.blueyonder...Spreadinga.jpg


Using 6mm square water channel(Kryotherm doesn.t do round) get:-

http://www.jr001b4751.pwp.blueyonder...preading3a.jpg

or:-
http://www.jr001b4751.pwp.blueyonder.co.uk/CWPSIa.jpg

my mind boggles at all the curves

I would observe that the curves in the plot "Calculated C/W vs. Flow Rate . . ", for the lower flow rates, are quite similar to those I actually measure
BUT
it is the offset for a given wb that will define the actual values - and the predictive ability of an equation

how on earth does the coolant pressure figure in this ?
FYI, I run at 5 to 8psi given how I am configuring the flow rate measurement and control
- you are suggesting that if I run at 2psi (not atypical for a WCer) I'm going to see a measurable difference ?

not too sure about this as I used to run this way and see no difference in the thermal results

be cool

1) The "Water Pressure" is the "calculated Head"(?) required to produce the flow when using 200mm of 6mm sq tubing.
The "c/w v pressure graphs" should probably not have been included(or at least qualified) but I did wish to highlight the the pump "Delivery Head" required to produce high flow rates.

2) No. At least not intentionally.