Swiftech's new CPU waterblock MCW5000

[gone_fishin]

Originally posted by Les
Yes the agreement is somewhat scary and probably unreal but worthy of of further investigation.

Wanted to use two values for r
(1)the min value(2.5D) allowed,this gives the max h value,to use as as an "effective impingement area. area".
(2)an r value representative of the bp area (30mm?) to get by manipulation a value to use as "background convection"*.
However as 2.5D(36.25mm) in the 0.57"(14.5mm) case is larger than the bp area I decided,somewhat arbitarily, to use the same r value of 25mm. This 25mm was intended as representative of a 50x50mm bp.Unfortunately mid way thro calcs memory said a more realistic bp was 60x60mm and used this in Waterloo calcs. Probably should redo with r=30mm(and correction of D= 0.0145 not 0.014)but ... ..and quick look showed little difference.
Apologise if this is as clear as mud but .................
Can (and have 1/4 done,somewhere) the two r (and hence two h) sums when nozzle diam approaches 3mm.

Using "the real value of H is 0.013m" get
Get the same for "the -U values are: D = 0.010, H = 0.005m".
Will "Press (Calculate Now) to calculate anyway"for both cases,but allow a little time ........ slow and oft befuddled brain.
Dunno whether to change r to 30mm.



.* As used in last my post here (but called Wall Jet) http://forums.procooling.com/vbb/sho...5&pagenumber=7

The Ht. should be between 2 and 12 times D, you are calculating for 1.3XD, decrease D or increase H. Since H is real world I presume, then you can more easily decrease D. As you increase H, r also increases (how to precisely calculate though?), As you decrease D, r should also increase? Overload on the logic circuits

[gone_fishin]

Originally posted by gone_fishin
The Ht. should be between 2 and 12 times D, you are calculating for 1.3XD, decrease D or increase H. Since H is real world I presume, then you can more easily decrease D. As you increase H, r also increases (how to precisely calculate though?), As you decrease D, r should also increase? Overload on the logic circuits

Edit, r is the radius of the jet as it hits the bp, not the total size of the bp as I understand it. Obviously as you manipulate the ve, d, and h the size will change. Observance and actual measurement (of r) is required here?

Edit to the edit, food for thought, click

Edit, edit, edit: Model of jet impingement

[BillA]

ooohh, ooohh
I'm disabled, I'm disabled

great link g_f

and the math is all done, not too bad at all
(other than having to write a program to do it)

but
is the jet submerged ? or not ?
(a LOCA implies not)

[gone_fishin]

Originally posted by unregistered
ooohh, ooohh
I'm disabled, I'm disabled

great link g_f

and the math is all done, not too bad at all
(other than having to write a program to do it)

but
is the jet submerged ? or not ?
(a LOCA implies not)

I do not think it was encapsulated (the two plates).
Here's an excerpt,
"If the gap size increases at the same velocity, the spreading flow width slightly decreases. Also, the spreading width greatly
decreases at the bottom of the flow film, if the pipe diameter increases from 1.7cm to 2.2cm."

The principle (size of r related to other variables under discussion here) should be the same but with added factors like fluid resistance in a closed liquid filled environment. The absolute numbers would definately have been different but the relationship of the variables as it pertains to the size of r is interesting.

Edit:
BillA, if you liked the last links, this will cause you to take a cold showerclick

[BillA]

keep googling, you're hot g_f
it was only the first link that was good,
and its quite 'off' not being an immersed jet
(but the calcs should be adaptable)

that last lab makes my feel ok about mine,
need to find some grant $ for my DAQ system

[gone_fishin]

Most of the current stuff is pay per view

Care to pay $35 per article to tickle the neurons? click, 284 and a few others

Thankfully most of this stuff isn't current

[BillA]

for $35 I can drive over to Rice U

[gone_fishin]

A may have to take that last statement back, according to these people we're examining the cutting edge

[BillA]

while I would, and do, agree that they are at the cutting edge,
it is equally certain that we are not

the stuff that is of interest to us (and that we can even understand) is probably 10 to 20 yrs 'old'

a great of the research that is done today involves the transient response analysis,
which really is quite beyond my ability in terms of both the performance and data reduction
- very deep

[gone_fishin]

Agreed. We are seeking an easy answer to submersed jet impingement to which there is no online calculator. But it's always fun to google.

click comparable to your conclusions?

[BillA]

yea, Aavid has the same setup
their advantage over mine is the multiple TCs so they can extrapolate the face temp and actual offset

while I could set up the die that way, obviously the wb would still not be

results the same, but mine lack that precision

[gone_fishin]

Synthetic jet array in air cooling
Browse around there to get the jist. Induced impingement through pulses. I wonder if that could be incorporated into pulses over a slow moving water stream?

Edit:
Uncanny pin grid array resembling swiftechs latest.

[bigben2k]

Nice link indeed, G_F!

Now I can't wait to get to work, so that I can print it (I'm actually looking forward to Monday, thanks!).

I was going to say something about it not being submerged, but BillA beat me to it.

[gone_fishin]

Your welcome bb2k. Be sure to keep us updated on any developments

Here's a testers link, click

I'm just about googled out

[BillA]

jesus christ g_f,
that is a phenomenal paper - but 43 pages of equations ??? makes my head hurt !
a fascinating spin through the IHS world,
but backwards as they are going for constant temp control

some tidbits:
apparently Intel CPUs are 'graded' at 85°C
so that would imply a throttling set point of ??
(can they disable it during testing ?, but if so then the rating would be pointless)
- my guess is that the setpoint is somewhat above 85°C as its effect is instantaneous

IHS thickness of 1.8mm
Rt = 0.42 cm²K/W
hc = 1200 W/m²K
and a bunch of info on lateral heat spreading (not so much)

good find g_f

[Les]

Originally posted by unregistered
the title of the article is "Waterblock Bench Testing Results"
to see it send an e-mail to Joe Citarella and ask him when it will be up

an excerpt just for you:
An interesting situation exists with the Swiftech 462-U and the 462-UH as they are the same wb, but with different connection sizes. Comparing their "C/W"s it is easy to see that flow per se does not do the 'cooling', else their positions would be reversed. To look at the effect of inlet flow velocity (more accurately die area impingement velocity), in place of the very large bore (0.57 in. ID) inlet connection initially tested with the 462-UH, a Swagelok barbed connector with a 0.39 in. ID was substituted. The outlet was not changed and the standoff from the bp surface was the same for both (but greater than the 462-U). It can be seen that the performance of the 0.39 in. ID inlet is substantially improved over the larger connection having lower impingement velocity.

Chart 4

Redone Flomerics and Waterloo sums.
For Swiftech 462-UH
Flomerics H=0.013, r =0.03, D=0.01 and D=0.0145
For Swiftech 462-U
Flomerics H=0.005,r =0.03,D=0.01.
Waterloo for both: Die 10x10mm,bp 60x60x8mm.

[gone_fishin]

Thanks, thought you would find it interesting.

One last gem in the rough. Let's see how long it takes someone to copy this baby
click

[BillA]

hot damn Les, don't know what to say

ok
what does a D=0.006m yield (H=0.013 and 0.006m)

I'm fishing here, obviously

g_f = Google_Finder

EDIT:
I've seen similar, and the patent apps (5 yrs old ?)
no one will make for a CPU, 300psi supply ??

[Cathar]

Originally posted by gone_fishin
One last gem in the rough. Let's see how long it takes someone to copy this baby
click

I was experimenting with these multiple jets like this (but not to this level of detail of course). The problem I found was that the jets in the middle tend to interfere with the performance of the outer jets, presumably the volume of the water being pushed out from the middle jets is seriously disrupting the flows coming out from the outer jets.

The design also reminds me very much of Volenti's direct-die cooling jet setup in a basic way.

[BillA]

because your pressure was too low/volume too high

I've seen them reported where the metal erosion was a near term problem

works quite well but is hardware intensive, not quite what we're doing

[gone_fishin]

Yea, its got a 5cm (2") inlet pipe. Would need a boiler pump or bilge pump to get that pressure.

The other pulse induced concept has me interested. Would need a stronger pulse but the same concept could apply. In theirs there is a constant flow of cool air being passed through then the pulses go off and induce a whole lot of cooling mayem.


Edit:
"ok
what does a D=0.006m yield (H=0.013 and 0.006m)"

Now you are in the defined parameters (>2X).

[Cathar]

Originally posted by unregistered
because your pressure was too low/volume too high

I've seen them reported where the metal erosion was a near term problem

works quite well but is hardware intensive, not quite what we're doing

That makes sense now. Yes, I can now see how trying to do this with a hobby pump would yield what I was seeing.

As for metal erosion, I can believe it. I was reading with interest a little while back about machining metals with ultra high speed water jets.

[Gabriel Rouchon]

For those of you who do not like marketing trash, I have streamlined and clarified the offending statement, which now reads:"outperforms all previously released Swiftech water-blocks"... and you can take that to the bank !

In response to "letting customers fill-in the blanks":
not true and unfair. What is true on the other hand, is that I prefer to let hardware review sites define the performance of my products, as independant (and hopefully impartial) third parties.

Focused flow vs patent pending DPM (Diamond Pin matrix): for those of you who have been involved with liquid cooling for a long time, you may recall that Swiftech was the first Company to introduce Focused Flow. I used this technology because I didn't want any channels in my blocks.

The problem with this configuration is the inherent difficulty in bleeding the block, as pointed out by Steve [H], and others.

Relocating the inlet and outlet as they are now, allows flawless bleeding in any position, while the DPM structure really does a magnificient job at absorbing the heat load, and yes, superior to the MCW462-UH by at least 1.5C (all tests equipment/conditions being equal).

My bottom line is this (no marketing here, just facts):

The product performs better than the best block we have ever made, it is more convenient and safer to use, it installs without removing the motherboard, and it is cheaper

I'm sure Bill A. will have a blast playing with it too ...

Happy holidays to all..

Gabe

[BillA]

Hi Gabe,
good to see a mfgr setting the record straight

no debate being initiated,
but in truth there is precious little guidance available to the user as to how to 'set up' the MCW462-UH
while you may "prefer to let hardware review sites define the performance of my products", I will roundly characterize these sites as being incapable of even the most rudimentary measurements
- and such in any case does not address the even more difficult task of 'optimizing' the 462-UH, which is a quite excellent wb with the 'right' connectors
-> and to my mind it is the mfgr who should be providing such guidance (not some ‘know-it-all tester’)

enough of that
"superior to the MCW462-UH by at least 1.5C (all tests equipment/conditions being equal)"
whoo whee, I wonder if you have any idea just how well the 462-UH can (be made to) perform ?
EDIT: the preceding sounds a bit strident;
what is meant is that with a whole bunch of testing to establish an 'optimum', the 462-UH is a very good wb indeed
- the challenge is accepted, 462 vs. 5000

"I'm sure Bill A. will have a blast playing with it too ..."
ahem . . . can we say 'testing' ?
yes, of course - just waiting . . . .

[pHaestus]

Gabe:

Thanks for responding. I probably came off a bit harsh before, it was not my intention to bash your company or your products. It is good to see wb mfgrs posting in threads without actively promoting.