New aquacomputer cuplex XT waterblock.. what do you think?

[idleuser]

http://snt-systems.com/catalog/images/aqua-computer/cuplexxt_promo.jpg

Personally i think it looks like a hybrid Alphacool NexXxos XP but with more bling bling factor! I wonder how it preforms hmmmm....

[thegreek]

can you say restriction?

It looks like it will perform well with a strong pump BUT why dont they put the inlet barb right over the middle to save some flow... I guess they are more concerned with looks then performance.

[MaxxxRacer]

flow killer of the highest order. give me a couple of hours in autocad and I could have that flowkilling monster giving you more performance with more flow and it wouldt cost any more to manufacture than it does now.....

[idleuser]

hehe I have no idea how much it'll cos tbut i'm thinking at the 80+ range msrp, seeing how much aquacomputer stuff usually cost :) But I agree it def looks like a flow killer but doesn't it look bling bling :)

[MaxxxRacer]

lol.. bling bling.. when in hell has XS been about bling.. look at cascades.. they are pretty ugly... covered in foam.. but they cool better anything this side of ln2..

[The Mofo]

Funny thing is, you can goto Hard Forums and read a 50+ page thread on this crap. 6mm lines, multiple blocks, god, and the price. Terrible performance. I guess if you got a rig you dont really care about overclocking the :banana::banana::banana::banana: out of, this may be an option for 300% more $. To each their own. WOnder why they dont concentrate on performance for as much as they charge for this stuff?

[MaxxxRacer]

who knows.... hard forums?? as in the [H]??? they couldnt be into that stuff there...

[RaptorRaider]

There are a lot of people only cooling their videocard.
They don't care if the block is a flow killer, as long as it performs great.
Almost every other German video card blocks are flow killers anyway; I don't think Germans care much.

BTW why doesn't Cathar make video card water blocks?

[MaxxxRacer]

no he doesnt. only the storm right now... He had thought about it and designed a block a while back but I think it was stolen and used by silverprop. Dont quote me on that though. I think its just rumor.

But the thing is that the blocks take alot of pumping power to perform great... and germans use TINY pumps... its not like they are using iwakis... like the nexxos xp performs amazingly well with and MD20.. its one of the best blocks on teh market if you use that pump.. but that means that you have nothing else in your loop, and you have fittings to convert from 1/2inch or 3/4 to the tiny tubes that the nexxos ues..

[Butcher_]

Aquacomputer, it's what happens when you let graphic designers make blocks. 100% bling, 0% performance.

[idleuser]

more pics :)
http://www.aqua-computer.de/prodimg/nd_products/cuplex_xt_duese_500.jpg
http://www.aqua-computer.de/prodimg/nd_products/cuplex_xt_boden_500.jpg

[caLume]

hm, cooler looks, hm, "different" :D
mainly i hate this ac stuff.
but it looks quite good.

any temps ?

[Butcher_]

Is that an aluminium top BTW?

[idleuser]

yes sir you are correct :) that is an alumnium top and it retails for 69 Euro which is about 90 dollar USD right now!! There aren't any temps yet since it's not even out I suppose it should due in a few weeks.

[stu_allen]

All you people who keep banging on about flow and 0% performance - have you actually ever used any of this stuff? Or are you just regurgitating the same old rubbish?

Tell me why i need more than 0.7 litres a minute flow. I never see more than 46 deg cpu when loading cpu and gpu together. Why would i care if it could be 5 degrees cooler?
(4 gpu coolers, nb cooler, cpu cooler, 2 6800 coolers, flow sensor, res, rad, pump (the 12v converted 1046 @70hz) in the loop)

you dont *need* any more performance, unless its for bragging rights.

[Butcher_]

Al+Cu = corrosion, you'd think people would have learned by now. :(

[Butcher_]

All you people who keep banging on about flow and 0% performance - have you actually ever used any of this stuff? Or are you just regurgitating the same old rubbish?

Tell me why i need more than 0.7 litres a minute flow. I never see more than 46 deg cpu when loading cpu and gpu together. Why would i care if it could be 5 degrees cooler?
(4 gpu coolers, nb cooler, cpu cooler, 2 6800 coolers, flow sensor, res, rad, pump (the 12v converted 1046 @70hz) in the loop)

you dont *need* any more performance, unless its for bragging rights.
You don't *need* water cooling at all, people here (in the main) water cool because they enjoy it as a hobby. They want to get the maximum out of their comp - to get the maximum you need the coolest temps, 5C does matter
in that situation.
As for 46C, I'm assuming that's off a mobo sensor and thus just a pseudo-random number so rather meaningless.

[The Mofo]

All you people who keep banging on about flow and 0% performance - have you actually ever used any of this stuff? Or are you just regurgitating the same old rubbish?

Tell me why i need more than 0.7 litres a minute flow. I never see more than 46 deg cpu when loading cpu and gpu together. Why would i care if it could be 5 degrees cooler?
(4 gpu coolers, nb cooler, cpu cooler, 2 6800 coolers, flow sensor, res, rad, pump (the 12v converted 1046 @70hz) in the loop)

you dont *need* any more performance, unless its for bragging rights.

So you are willing to pay twice as much for half the performance? Why? Thats like you showing up at a drag race with an Integra vs my Lightning. Here's a video of this very sitiuation. Evo 8 Lancer, 70 shot of nitrous, talking crap about my " pickup "... The end result... Equals This < Click Here > (http://www.venomousmofo.com/mofo04/Kevin.MPG)

Evo = AC... G4 Storm/Silver Prop GPU = My Lightning

[EnJoY]

Mofo, whatchu got under that hood?

[thelostrican]

who knows.... hard forums?? as in the [H]??? they couldnt be into that stuff there...
:( not everybody of course.... :cool:

[thelostrican]

All you people who keep banging on about flow and 0% performance - have you actually ever used any of this stuff? Or are you just regurgitating the same old rubbish?

Tell me why i need more than 0.7 litres a minute flow. I never see more than 46 deg cpu when loading cpu and gpu together. Why would i care if it could be 5 degrees cooler?
(4 gpu coolers, nb cooler, cpu cooler, 2 6800 coolers, flow sensor, res, rad, pump (the 12v converted 1046 @70hz) in the loop)

you dont *need* any more performance, unless its for bragging rights.
why watercool then? silence? i can understand that....
one thing though, come here to the caribean, and you will see why you want more cooling power.....

[Ancient_1]

Your truck would of given my old small block regal a good run. My berst in street legal trim was 12.005 at just under 117. with open exhaust would of probably done high 11.70s with good air.

[saratoga]

All you people who keep banging on about flow and 0% performance - have you actually ever used any of this stuff? Or are you just regurgitating the same old rubbish?



I've looked at the numbers and done the math. Have you? It certainly doesn't sound like it.


Tell me why i need more than 0.7 litres a minute flow. I never see more than 46 deg cpu when loading cpu and gpu together. Why would i care if it could be 5 degrees cooler?


I don't know why you need a computer, let alone watercooling, but in that flow range, you're getting dangerously close to aircooling performance wise.


(4 gpu coolers, nb cooler, cpu cooler, 2 6800 coolers, flow sensor, res, rad, pump (the 12v converted 1046 @70hz) in the loop)

you dont *need* any more performance, unless its for bragging rights.

Why not just use air cooling? Like wasting your money? Bragging rights indeed (particularly the flowsensor).

[iboomalot]

.7 * 60 = 42 liters per hour / 3.7855 liters per gallon = 11.09 gallons per hour

or .18gallons/min That so low it won't show up on a flow chart which starts at .25 gall/min

why even have a flow meter??? I think you answered your own question. So why do you watercool your puter???

[quicksilverXP]

I think it performs the same as my German Nexxos XP (supposedly it has less restriction than my XP, and the older AC Cuplex Pro was somewhat on par with the Nexxos XP). I've pretty much stuck with 3/8" tubing since 1/2 is unnecessary and ugly. My setup consists of a Black Ice Pro II radi, 2 Papst 120mm fans, a 3/8" Dangerden Acetal Block, and an MCP350 along with that Nexxos XP block from Alphacool. The FX55 at 1.68v was kept cool, never went below 48 degrees Load, at 3.0Ghz.

For testing purposes, I switched to a 1/2" setup with an MCP650, a 2x120mm Evercool fan setup with a bonneville 77 rad, same Acetal DangerDen CPu block, and a MCW6002 block. Difference in temps were 3 degrees, never went above 45 degrees Celisus. Change in overclock? Nothing.

I'm telling you, you guys worry too much about the watercooling differences between German and U.S. flow... where the real purpose in mind is overclocking.... and frankly, a 1/2" setup isn't worth the noise/fat tube ugliness with huge American rads.

This block looks nice, and you're already judging it as a "flow restriction" with comments that are so typical like "Can you say... flow restrictor?" Ohh please... watercooling now has diminishing returns, I'm reviewiing a Thermaltake Bigwater now that has the same overclocking results as a high-end 1/2" setup. If you're so worried about those small change in temps, get a Mach 1 for 350 used, or 400 brand new.

*edit*.. and I agree.. you don't even need watercoolign anymore with the heatsinks we have now... watercooling is geared more than ever now towards noise rather than performance. With the new 6xxx P4s that run cooler, and 90nm AMDs, high-end heatsinks are more than enough. The science of watercooling has already reached its limits in terms of performance.

[MaxxxRacer]

wow.. iboomalot.. I didnt even think to convert that... but PH doenst even start his test until like .5gpm.. lol... .18gpm is so low you might as well be air cooling.. that is a TOTAL waste of money....

[idleuser]

Yeah since watercooling has reached it's limit then I suppose we can say we can all throw away all our gear and go back to the original heatsink! I suppose all the watercooling company should just pack up and stop making new designs and we'll all stick with the original maze design block and cathar design wouldn't have to be copied :x

[MaxxxRacer]

your right idle. you have hit upon a grand point here.

[The Mofo]

Mofo, whatchu got under that hood?

Kenne Bell Blowzilla @ only 16 PSI and a couple of bolt on's :banana:

[MaxxxRacer]

mofo, time for a port job

[The Mofo]

Your truck would of given my old small block regal a good run. My berst in street legal trim was 12.005 at just under 117. with open exhaust would of probably done high 11.70s with good air.

They gave me a bunch of :banana::banana::banana::banana: for no rollbar, so i had to throw the 3" pulley on and disconnect the bottle for the runs. 1.89 60ft, my 28" slicks couldnt bite there since it had rained the night before. Best run is 11.41 with a 50 shot and 18 psi :)

[MaxxxRacer]

mofo, you would have had a fun time dragging my barbers 400hp integra.. that little mofo was f'ing fast... 18psi hks turbo, 1.8liter fully ported with custom crank, pistons, and rods, 3inch hks exhaust (the thing sounded like a v8 not a honda).

but the nice thing about that car was that it was a sleeper. It looked completely bone stock except for the 3inch exhaust (which was well hidden), and some nice rims.

[Ancient_1]

I had a little better 60' (in the 1.70s) with my M&H 275 60 (would of been a 26 x 10.5 slick) but i had single 4bbl no nox. It is what you would call a sleeper other than a little tappet noise and rough idle. I could of used a little more gear tho since I crossed the line at 5000 in 3rd (1-2 shift at 6700 and 2-3 at 6500) but still left good with the 3.06 1st gear of the 700r. For me wasnt only the roll bar I didnt have floating axles either, c-clips dont work.

[The Mofo]

mofo, you would have had a fun time dragging my barbers 400hp integra.. that little mofo was f'ing fast... 18psi hks turbo, 1.8liter fully ported with custom crank, pistons, and rods, 3inch hks exhaust (the thing sounded like a v8 not a honda).

but the nice thing about that car was that it was a sleeper. It looked completely bone stock except for the 3inch exhaust (which was well hidden), and some nice rims.

Ive seen a few clean imports but they are far and few between. Lot of " ricers " which make the true hot rod imports look bad. We usually have a private track day every 3 months in bakersfield at fomoso.. Usually some really fast imports show up. Ill let you know the next time we go out :)

[MaxxxRacer]

mofo, he sold his integra due to the iritation of not having AC (the huge hks turbo and intercooler replaced the AC) and the sheer noise of the exhaust. Now he has an M3 and he is modding that right now.. All i can say about the m3 is i want one.. that thing is amazing.. we were drivng at 60mph and he nails it and it gives you whiplash...

[idleuser]

hehehe i'll take that m3 with my friends t67 :P

[The Mofo]

He's got the new M3 right? With that SMG stuff? Those are nice, i looked at one in the beginning of last year. My buddy who tunes my truck has one and its wicked. I just cant believe the prices for mods on those things. I gotta by a house before i start dumping $$ around again.. Oh and i just realized, we just crapped all of this guys thread. LOL.. Hit me up on aim dude, VenomousinHB

[Craig]

All you people who keep banging on about flow and 0% performance - have you actually ever used any of this stuff? Or are you just regurgitating the same old rubbish?

Tell me why i need more than 0.7 litres a minute flow. I never see more than 46 deg cpu when loading cpu and gpu together. Why would i care if it could be 5 degrees cooler?
(4 gpu coolers, nb cooler, cpu cooler, 2 6800 coolers, flow sensor, res, rad, pump (the 12v converted 1046 @70hz) in the loop)

you dont *need* any more performance, unless its for bragging rights.

You're try'n to cool 4 differant blocks with a 1046 pump? Yet wish to claim it comes anywhere near being a performance rig. Or that water cooling no longer has a place vs high end air? :lol: :ROTF:

Not really funny though...... :sick:

OK, this requires correction......:devil:

You've the gall, after wasting money on crap like a flow sensor & NB chipset block so you end up with .2gpm of flow in your loop, to ask if we know what we're doing?

Real question is what makes you even think YOU DO?

Had you tested with a really high end US type rig. One with Iwaki MD20Z pump, TDX/RBX with #4 jet, Storm or Cascade CPU blocks, Fusion GPU block combined with a Weapon Shroud/monste core you'd have seen a H3LL of a lot more than 3C gain I promise you.

But that's no surprise, 7-8X the flow rate can do that for you.......10C over your set up would not surprise me at all.

You guys from the EU have a really good grip on design of really low/no noise rigs. To bad you keep try'n to convince your selves that they can also compete with a US style performance rig........they can't, or at least ones designed like yours can't. Good top shelf design for low noise is very, very differant than a top shelf design for Xtreme performance. They can be combined, to a degree, but only if you plan with great care. And such combined low noise/high performance rigs need exterior rad/fans/pump cases lined with accustic foams & with S shaped air entry/exit to reach/exit the rad/fans to trap and eliminate noise. OH, also a reobus for the fans for use in undervolting them when max air flow isn't wanted, but lowest noise is. Such a compromise rig gives up a bit of performance due to slightly reduced air flow through the rad caused by the S bend air flow pattern through the rad box. But with use of components like I listed above, such a rig can have very modest noise & be very close to the most Xtreme performance.

:shrug: I'm done now, that hits the higher points. :shrug:

EDIT:

I was so PO'd by his post I retyped my response like 4X. Now I see we've moved on to hot cars instead of his overly hot system........oh well, welcome to a hour ago.......

[MaxxxRacer]

Craig, very good points made in ur post. Hopefully he understands now why we gripe so much about euro blocks and why we have a very good reason to.

[EnJoY]

My friend Rob just got a 95 Mitsu Eclipse...all black, mint, GS-T...so turbo charged. He plans on giving it the works over the next year, with that engine, you can expect a lot. :)

I personally have been dying for a 91-95 Toyota MR2. Turbo, mid engine, rear drive...looks like a ferrari with 17's. Just exotic. :)

Ahhh...I love cars, tell me why I blow all my money on pc's?

[The Mofo]

My friend Rob just got a 95 Mitsu Eclipse...all black, mint, GS-T...so turbo charged. He plans on giving it the works over the next year, with that engine, you can expect a lot. :)

I personally have been dying for a 91-95 Toyota MR2. Turbo, mid engine, rear drive...looks like a ferrari with 17's. Just exotic. :)

Ahhh...I love cars, tell me why I blow all my money on pc's?

Dont feel bad bro. I blow money on my Lightning, 2 Cobras, and SVT focus. Add the computers, home theatre and other hobbies into the mix and im surprised i can eat sometimes. lol

[MaxxxRacer]

muahahah.. im soo poor i cant even affor insurance so cars dont matter much to me. I just get lots of exercise on my road racing bike.

[The Mofo]

Looks like your cat eats good :)

[MaxxxRacer]

yah... live with my parents.... it goes.. mom eats one bite, cats eat one bite. so all in all its a 3 way split as we have 2 cats.

as for money, its hard to get a job with my college schedule. no one wants to hire a kid who can only work firday sat and sunday...

[The Mofo]

Welp, you could either deal drugs or work in a night club. Hell, maybe you could be the next Dirk Diggler, thats how he got discovered! :ROTF:

Im 34 and got a decent job.

[saratoga]

I think it performs the same as my German Nexxos XP (supposedly it has less restriction than my XP, and the older AC Cuplex Pro was somewhat on par with the Nexxos XP). I've pretty much stuck with 3/8" tubing since 1/2 is unnecessary and ugly. My setup consists of a Black Ice Pro II radi, 2 Papst 120mm fans, a 3/8" Dangerden Acetal Block, and an MCP350 along with that Nexxos XP block from Alphacool. The FX55 at 1.68v was kept cool, never went below 48 degrees Load, at 3.0Ghz.

For testing purposes, I switched to a 1/2" setup with an MCP650, a 2x120mm Evercool fan setup with a bonneville 77 rad, same Acetal DangerDen CPu block, and a MCW6002 block. Difference in temps were 3 degrees, never went above 45 degrees Celisus. Change in overclock? Nothing.

I'm telling you, you guys worry too much about the watercooling differences between German and U.S. flow... where the real purpose in mind is overclocking.... and frankly, a 1/2" setup isn't worth the noise/fat tube ugliness with huge American rads.

This block looks nice, and you're already judging it as a "flow restriction" with comments that are so typical like "Can you say... flow restrictor?" Ohh please... watercooling now has diminishing returns, I'm reviewiing a Thermaltake Bigwater now that has the same overclocking results as a high-end 1/2" setup. If you're so worried about those small change in temps, get a Mach 1 for 350 used, or 400 brand new.

*edit*.. and I agree.. you don't even need watercoolign anymore with the heatsinks we have now... watercooling is geared more than ever now towards noise rather than performance. With the new 6xxx P4s that run cooler, and 90nm AMDs, high-end heatsinks are more than enough. The science of watercooling has already reached its limits in terms of performance.

I don't really understand your point. The 6000/6002 is really the only low flow waterblock ever made, so what does that have to do with high flow systems? That if you use a low flow optimized block in a high flow system its a waste? Swiftech could have told you that, hell Bill's always pushing 3/8 in stuff for that reason.

And 1/2 tube is noisy? Ignoreing that, the whole reason we use heatercores is because they're very low restriction which allows for very low noise fans.

And finally, what relevence is the overclock on your CPU to someone else? I sure as hell don't have your chip, so why should I care that a better block didn't make a difference to you?

[MaxxxRacer]

saratoga, the HC's are incredibly high air flow resistance. They are so good becuase they have incredibly low waterflow resistance compared to other solutions that dont perform as well.

[Butcher_]

I think it performs the same as my German Nexxos XP (supposedly it has less restriction than my XP, and the older AC Cuplex Pro was somewhat on par with the Nexxos XP). I've pretty much stuck with 3/8" tubing since 1/2 is unnecessary and ugly. My setup consists of a Black Ice Pro II radi, 2 Papst 120mm fans, a 3/8" Dangerden Acetal Block, and an MCP350 along with that Nexxos XP block from Alphacool. The FX55 at 1.68v was kept cool, never went below 48 degrees Load, at 3.0Ghz.
3/8" isn't really what we're talking about with "low-flow" systems, it's the 6mm (or smaller!) tube crowd with very restrictive blocks (like the nexxos xp for one) and multiple secondary blocks then they add in a woefully underpowered pump and call it the best thing ever.
Your system is much more typical of a US system really, so not a good basis for comparison. Personally I'd run 3/8" or 7/16" if I didn't already have an investment in 1/2" gear.


*edit*.. and I agree.. you don't even need watercoolign anymore with the heatsinks we have now... watercooling is geared more than ever now towards noise rather than performance. With the new 6xxx P4s that run cooler, and 90nm AMDs, high-end heatsinks are more than enough. The science of watercooling has already reached its limits in terms of performance.
I disagree that watercooling has reached it's limits, it's just that people don't want to pay the money to reach those limits (which is understandable).


Maxxx, HCs aren't too bad for air-flow, they beat the old tube+fin setups, but they aren't exactly low resistance either.

[stu_allen]

why even have a flow meter??? I think you answered your own question. So why do you watercool your puter???

Because its a hell of a lot quieter than air cooling. My system is set up to cool passively (ie completely silently) when just surfing, and as the water temp increases the fans start up and increase speed.

The flow meter is so that the controller (aquaero) can initiate a shutdown if the flow drops (ie incase something leaks).

[stu_allen]

Real question is what makes you even think YOU DO?

Had you tested with a really high end US type rig. One with Iwaki MD20Z pump, TDX/RBX with #4 jet, Storm or Cascade CPU blocks, Fusion GPU block combined with a Weapon Shroud/monste core you'd have seen a H3LL of a lot more than 3C gain I promise you.

But that's no surprise, 7-8X the flow rate can do that for you.......10C over your set up would not surprise me at all.


Yawnorama.

A high end us rig would give me what exactly? (apart from more noise). One or two mhz more? Id be surprised if it gave more than that.

The point i was making, which most of you seem to have missed, is who cares if it could be 10 degrees colder. It aint going to do a fat lot.

[stu_allen]

I've looked at the numbers and done the math. Have you? It certainly doesn't sound like it.
Why not just use air cooling? Like wasting your money? Bragging rights indeed (particularly the flowsensor).

I really couldnt care less what you think of my setup, i bought components to achieve a particular goal (silent as possible, whilst not restricting overclocking with increased chipset, cpu and gpu voltage) and it acheieves it perfectly.
Again, why would i want more flow, when what i have does the job just fine?

[Butcher_]

A high end us rig would give me what exactly? (apart from more noise). One or two mhz more? Id be surprised if it gave more than that.

The point i was making, which most of you seem to have missed, is who cares if it could be 10 degrees colder. It aint going to do a fat lot.
Your requirements are not those of others, you'd do well to keep that in mind.

Extra cooling with a high end system can give anything from 0 to maybe 20% more speed, to some it's worth the effort and noise, to others it isn't, each must make their own decision about the trade-offs. Blanket statements such as yours that low flow is just as good as high-flow are very narrow-minded, especially given you have zero factual evidence to support your claims, just half-baked numbers and conjecture.

[MaxxxRacer]

Stu. With your setup, you are most surely not getting the most out of your components. when we all said that your setup is equivlanet to high end air cooling in cooling capacity, we wern't kidding. Its just about impossible to get a system with lower flow than yours, and i hate to break it to you, with only a trickle of water, you arnt gonna be able to cool anything really effectively.

And to say that you are getting the maximum out of ur components is a statement of ignorance. If you have over your componets and we put them in one of our water rigs i could pretty much garuntee you that we would get a higher overclock than you did in your setup.

But i guess this is all mentality to a degree. Your want a system to be as quiet as possible, but in doing so forget to a large degree there there is a performance factor. And some of us, do forget what noise is (ancient and your tornandos), but to be honest, the majority of us watercoolers run our fans at a modest 7volts and they are incredibly quiet, and perform incredibly well.

[idleuser]

heheh the tornado is the reason why I jumped on h2o :x

[MaxxxRacer]

lol tell ancient that. he is battling out heat with his single 120mm rad.. he is too stuborn to get a bigger case and go with a full on 302 HC. But he says that he has so many computers in his office that it doesnt matter. It would be a windtunnel with or without it...

[EnJoY]

Ancient is running three fing blocks on one 120mm radiator? Wow?!?

Time for a Lian-Li, Coolermaster, or Silverstone case. Something with space and elegance...ditch the pos cheiftec.

[stu_allen]

Maxxx - i never said i was getting the absolute maximum. But i do believe that in a room temperature setup (so ignoring chillers and phase change) you arent going to get an enormous lot more (and certainly not 20% lol!). Sure if you are competitively benching then every mhz counts, but in the real world i doubt that me running at 36deg under load would give me a fat lot more than running at 46deg, so just enough, equals enough. No need for any more.
If anyone wants to prove me wrong, i'll gladly eat my words. :)

[stu_allen]

Your requirements are not those of others, you'd do well to keep that in mind.

That was kind of the point I was trying to make in the first place ;)


. Blanket statements such as yours that low flow is just as good as high-flow are very narrow-minded, especially given you have zero factual evidence to support your claims, just half-baked numbers and conjecture.

mm. So are "high flow is better" statements. ;) I was saying there comes a point where flow is good enough in a system.
And as for zero factual evidence, you must have missed the part where i said that with an overclocked and overvolted fx55, overclocked and overvolted 6800's x2, overvolted chipset i hit a max of 46-48deg under load (gpu and cpu)?

[Butcher_]

I saw that, but I assumed you use a mobo sensor reading, which is far from factual.
As for how much you can gain, how can you say you won't get 20%, without trying it?

[stu_allen]

I'll stick one of my temp sensors on the copper waterblock base for you then with some thermal epoxy.

I tell you what. I havent got a high flow setup to test, but you can do it for me by dropping flow, using some sort of obstruction. Show me that a 10degree rise in temperatures hurts your clocks by 20%.

[Ancient_1]

10° would make a difference of about 150mhz on my 3.0c. That was basicly the difference I had between the sp94 with a tornado and the 1st setup of my water cooling.

Not near 20% but is enough to justify my watercooling to me. It was probably a little more with my 2.8c and 2.8e but had my setup configed different and had closer to 14° difference in temps. But with your flow I really wonder if your system could even match the sp94 and tornado performance wise.

[Craig]

I'll stick one of my temp sensors on the copper waterblock base for you then with some thermal epoxy.

And this won't tell you any more than the MB sensor so far as real CPU temps go.




I tell you what. I havent got a high flow setup to test, but you can do it for me by dropping flow, using some sort of obstruction. Show me that a 10degree rise in temperatures hurts your clocks by 20%.

So you want someone to drop a flow rate of 1.3-2gpm down to .18gpm to prove this to you? Very few have the gear built into a system to allow accurate flow rate readings. And not sure if you'll find anyone willing to drop flow rates so low.

Sigh, I looked all over, nobody tests below .5gpm which is almost 3X the flow of your system Stu. Such higher flows make accurate comparison impossible at this moment. It's too bad to, I would have loved to have seen such a comparsion.

[Ancient_1]

PH has started testing at either .25 or .3gpm for a few months now. I think he started with the MCW6002 review. It shows that it really starts getting bad when you drop below .5gpm and a lot of blocks really start falling off at .75gpm, just like the gains get much smaller over the 1.5 ~ 2gpm for most all block also.

Here is a shot of his from the G5 review that shows the difference starting at .3gpm. Now you can pretty much guess how it goes if you had the flow between that and 0.

http://webpages.charter.net/chew_toy/g5flow.jpg

[Craig]

10° would make a difference of about 150mhz on my 3.0c. That was basicly the difference I had between the sp94 with a tornado and the 1st setup of my water cooling.

Not near 20% but is enough to justify my watercooling to me. It was probably a little more with my 2.8c and 2.8e but had my setup configed different and had closer to 14° difference in temps. But with your flow I really wonder if your system could even match the sp94 and tornado performance wise.


Thanks for that info, and IMO your Tornado & SP94 would cool better, though would also be a great deal louder doing it.

If Stu is after no noise, then he's got what he wants. Nothing wrong with that at all, but that doesn't make it any sort of performance rig.

One thing you've also overlooked Stu is that the block you wish to defend here is designed for use with a pump that can drive those inlet jets for good impingement. Your 1046 can't do that, it's flow & pressure are greatly inadaquate. Without the pressure to dive good water velocity through those jets this new block will sux for cooling.

On the other hand, I think all here should reconsider the possible value of this new GPU block.

Yes, it IS restrictive, but so are the Stom G4 & 5, Cascade & SS Cascade, TDX/RBX with #4 or 5 jets. Yet these blocks are amoung the very best for low temp cooling of a CPU.

Also consider this, where are the greatest gains being made these days on systems for really HIGH 3D marks found? Are they at the CPU or GPU? We all know the GPU is becoming ever more important than the CPU in system performance these days.

So why not accept the idea of a GPU block designed with jets for max cooling by use of impingement jets? I'll stick my neck out and say that this is the sort of direction that GPU blocks need to move in. Further, perhaps we need to consider a more moderate flow resistence designed CPU blocks use to allow more flow for the GPU.

Consider this,
A Xtreme pump like a Iwaki MD20Z can drive approx .5gpm more flow through a TDX than it can through a Storm G5.

While I don't question that the G5 cools better will 2C be enough to compare to what .5 gpm of increased flow through the Fusion could do for GPU clocks?

Let's hear it guys what do you think regarding this?

[Craig]

More good info Ancient_1.

And yeah, if you continue those curves to the point of .18gpm which he reports getting, the temps will be much higher.

What do you think regarding my ideas above?

[Ancient_1]

I honestly am not sure that the GPU blocks will/should go that way. I believe more people will be using pelts on the vid card and those types of blocks are not suited to the TEC's. With more than one med to high restriction blocks in a system flow can become a real problem (considering sli will add a second gpu to the circuit) and the pumping power required will add its heat to compound the problem.

If that were to be the case I feel they would have to try and get the resistance matched to the CPU block and really think about running parallel (the only way I really see running more than on med to high restriction blocks in a system).

The main thing would be the most efficient way to get the flow through the blocks in the 1.5~2gpm range and with 2 or 3 restrictive blocks I believe that parallel may be the better choice along with things like the single pass rad.

[Craig]

Really if you look at that graph, anything below 1gpm shows a rapid decrease in performance with any of those high performance blocks.

And if the G5 were installed in Stu's rig it would further lower his system flow rate. Differance would not be great, but would be there. And any further reduction would only make things even worse.

Ancient_1, I tried going over to Pro Cooling to make use of the block comparison graph, but it loads crazy. Could you give it a try & let me know the results, I want to be sure if it's me or the site. Thanks

[Ancient_1]

It is the site. I would guess it has something to do with the fact Joe is moving the site to a different host this week.

[Ancient_1]

Really if you look at that graph, anything below 1gpm shows a rapid decrease in performance with any of those high performance blocks.

Yes I agree but in real temps they would still have acceptable performance for most people as long as it was over .5gpm. Its below .5gpm/1lpm that things really just get plain rediculous and lose any benefit of watercooling other than maybe noise. I feel many people would be satified with the performance of a lot of the German style setups if they just used a decent pump (the Laing DDC looks like a great match) and atleast 8mm id and prefferably 10mm id tubing. Also dump a few of the block like WC PS's, HD coolers and ram coolers.

[Craig]

It is the site. I would guess it has something to do with the fact Joe is moving the site to a different host this week.

Thanks much! :toast:

In regards to TEC's I agree those blocks shouldn't be designed to be very restictive.

I was thinking in terms of a loop without TEC's in it.

I also agree with you regarding a set up with 2-3 restictive blocks. I touched on that in my earlier post in comparing the TDX #4 to the G5. I really would love to be able to test that, a TDX #4 jet vs G5, though to be fair, the TDX should also be silver or both should be copper blocks.

In regard to parallel vs series for a rig with SLI. Why not send all flow through the CPU block then Y for splitting the flow to both GPUs. Even with the most recent vid cards GPU heat is still a good deal lower than that of a CPU & the devided flow from a strong pump could do the job.

Thanks again for checking that for me. :up:

[Craig]

Yes I agree but in real temps they would still have acceptable performance for most people as long as it was over .5gpm. Its below .5gpm/1lpm that things really just get plain rediculous and lose any benefit of watercooling other than maybe noise. I feel many people would be satified with the performance of a lot of the German style setups if they just used a decent pump (the Laing DDC looks like a great match) and atleast 8mm id and prefferably 10mm id tubing. Also dump a few of the block like WC PS's, HD coolers and ram coolers.


Agree 100%!

And the DDC would still allow for a very silent rig while also giving 3X + the flow of that 1046.

[Ancient_1]

In regard to parallel vs series for a rig with SLI. Why not send all flow through the CPU block then Y for splitting the flow to both GPUs. Even with the most recent vid cards GPU heat is still a good deal lower than that of a CPU & the devided flow from a strong pump could do the job.
Atually I would do that only without any Y's. Use a RBX or WW and use the outlets instead of the 1st Y and use the dual inlets on the res for the 2nd.
When I get my G5s I plan on putting the third barb on it to replace my RBX to keep the other two blocks in parallel which is exactly what I would do if I had an SLI setup also. Having a Y between the CPU block and the vidcards would make routing almost impossible with 1/2" tubing also, I have some idea on that from experience with my RBX when I used the Y and a T-line.

But what you were saying about the more restrictive gpu blocks would lose much of the advantage of the impingment design but cutting the flow that much by runnung after the cpu before splitting. It really would come down to the pump choice on which would be the best way to get good flow through all the blocks. For me the GPU is WCd only for reliability which includes lower temps and eliminating the small fans in my system (the only fans I have had problems with failing on me over the years has been the NB and GPU cooler). I have only had 1 80mm or larger fan fail on me in the last ten years and atleast 5 NB and GPU fans bite it.

[Craig]

Good points Ancient_1

I overlooked your point in regard to runnin off of the outlets of the WW or RBX as I've avoided such blocks. But yours is a point well taken.

I think I'm going to go with a DFI NF4 board with a single card though as costs to much to go dual vid cards IMO. And a high end A64 + vid card with both water cooled and overclocked should do me fine for a good long time.

I seem to recall saying something like that before however.......LOL! And a year can seem like a VERY long time. ;)

[MaxxxRacer]

wow craig and ancient.. you guys really went on a rant.. not sure where to start. I guess the only thing that I really want to comment on is craigs statement about the gpu blocks moving to impigment design. I hve thought about this long and hard, and here is the reason that I think its not a great idea... First off is the flow killing habbits of impigment blocks. With two of them you would need iwaki sized power to get good flow rates.. it would just be sick how bad the flow rates woudl be in a 1/2inch system with a g5 (for example) on the cpu and gpu.. you would be getting .7gpm with a MD20Z!!

With that said, I have noticed that gpus tend do not care so much about temps and cpus do. Call me crazy, but gpus can oc ok even at 70C!!. So until you get into negatives your gains of 5C from maze4 design to impigment wont be that great. Your overclocking gains that is i mean. and it might, i stress might even be less gains due to the reduced flow rate of having a second impigment design.

Oh, one more rational. You really want the best cooling on the cpu as it is the highest heat output device (although my x800xt has a higher thermal output than my winchester (RIP)), which is generally the cpu. this is very true in intel systems. the gpu is a secondary device.

but with all that said, what i would love is to use two monster heatercores, two RD30's and have a G5 on my cpu and have a gpu cooler based off of the G5 design on a seperate loop for the gpu... i would absolutely love to have fun with that.. it would be nutty expensive but it would be sweet.



I have come to like ancients idea on the RBX split to two blocks more and more since he first told me about it. It would really work best for SLI setups. having the two blocks in parallel would really help out with the flow rates, thus increasing the rads effeciency, and improving your cpu temps. Im not sure how much it will affect your GPU temps, but atleast this way your gpu temps will be the same.

[Butcher_]

Another thing to consider is that GPU dies are much larger so heat density is lower which makes them easier to cool with a maze design than a cpu.

[bimmer01]

damn nice waterblock, where u can u get these?

[caLume]

damn nice waterblock, where u can u get these?

by time.

http://www.aquacomputer.de/

hf :)

[idleuser]

preorder will be taken soon at www.snt-systems.com

[tinker77]

They look nice thats for sure but.....

[Craig]

One thing is very clear. The onset of the new SLI systems is going to require, no demand, some new testing in how to get the best System performance. A slightly less high performance CPU block that does not restrict flow nearly so much as a G5, like a RBX with #4 jet, could better provide good flow to dual vid cards in a set up plumbed lke Ancient_1 suggested, though it would still need a Iwaki to drive it all. Still, I know where I can get one of those.....;)

This whole SLI or the ATI equal of it, is going ot cause some fun in working out the best arrangement of blocks in a loop.

And we havn't even begun to consider the rad requirements when all 3 chips in such a set up are really driven hard. A heavy O/C'd CPU + dual overvolted GPUs is going to be dumping a good deal of heat 200+ watts into the water, even before you add in a extra 30w of pump heat. When making a all out 3D marks run it's very possible we'll be seeing a total heat load of 250 watts, which is a bit high compared to the recent past when no TEC's are in use.

Add in 2 TEC's and you better consider Weapon's monster core as a starting point for getting rid of the heat. :eek:

With the increasing importance of the vid cards performance, I suspect even a rig with a single card might show higher System performance with a max performing GPU block, even if it reduces the CPU blocks performance a bit. Time will tell. ;)

EDIT:
I'd have to hit the lotto to be able to afford to do all the testing I'd love to have a try at. I'll just have to hope someone with the needed testing gear feels interested in the same things....... :( .

[MaxxxRacer]

talk to pH, or bill adams. they both have alot of gear. bill has alot more... just snoop through his ebay endeavours. he has more test gear than nasa..

[RaptorRaider]

Al+Cu = corrosion, you'd think people would have learned by now. :(

I've heard people complain a lot about this reaction; a way to prevent this would be Zerex I think.

Anyway, the reason I'm posting this is because I'm wondering WHY?
My guess would be that there would be some kind of redox reaction but I don't understand what copper would have to do with it. :confused:
And isn't it true that aluminium in water should already corrode a little bit?

[Butcher_]

Google galvanic corrosion.
Zerex will reduce it, but not prevent it entirely, especially if it's untreated aluminium.

[Craig]

talk to pH, or bill adams. they both have alot of gear. bill has alot more... just snoop through his ebay endeavours. he has more test gear than nasa..

Maxxx, it's not that I don't know what gear to buy, or where to find it.

My next build is going to cost big time, no funding for such testing gear as I'd need for good accuracy & without high accuracy......why bother.

I did make a offer of help in that thread for GPU block testing at Pro Cooling. I'd want to see pH have a Fusion in the test, even if I have to pay for it. But that's the limit for some time.

[Ancient_1]

The way I suggested using a 3 barbed block was actually more for ease of routing than performance. Overall performance would still probably be series with using GPU blocks like the Maze4 which have rather low restriction and decent performance (according to JoeCs testing the Maze4 GPU cools better than the CPU maze4). But going from card to card could be a real problem and depending how it was done cause as much or more restriction than the blocks themselves.

For me it was an easy choice since the only time I ever OC my vid gard is to run a few benchmarks and I do that very seldom. But if a person used faily restrictive blocks for the gpus I do believe running them in parallel would give the best system performance overall.

[MaxxxRacer]

Craig, i mean to ask ph or bill to help you out with testing. Thanks for your offer on the thread i made way back when.. pH is getting around too it.. I IM him every once anda while and he has been bizy i suppose.. lol... he hasnt done anything with it...

[Craig]

No problem Maxxx, to be honest I didn't notice it was a old thread. LOL

And it's something that should have been done before now. It would have given some valueable info as to which CPU & GPU combinations would be best, as well as which GPU block is the best single performer in it's own right.

Hope pH gets to it soon. Let me know if he does, I'm here much more than there.

[MaxxxRacer]

i just talked to him. he just got a 6600gt to test on.. he has a

mc50
mc60T
maze4
maze4TEC
silverprp fusion HL (he still has to buy that)
old polarflow

he is going to do the temp readigns from the on die temp gauge through software. he said if it isnt accurate he will think of something. as usuall he will do the delta T's wiwth respect to W.

[EnJoY]

MCW60? New swifty gpu block eh?

[Craig]

I could care less regarding TEC blocks, but will be highly interested in all the data on the other blocks.

[MaxxxRacer]

well when i asked DD and bill to send me the blocks, both of them went and sent pH the TEC blocks on thier own accord... either way it will be fun to see the temps on those TEC blocks.

enjoy, im glad you noticed that. its swifty's new top secret block that pH will be testing. it uses a 500watt TEC and is made of pure silver. its been said it can get a 6800ultra down to -20C.

[Ancient_1]

Here (http://www.systemcooling.com/mcs_wc-08.html) is another place to get decent results on block testing.

[MaxxxRacer]

I like his testing mothodolgy as well. he is very good.

he gets some interesting results there... somewhat contradictary to what pH is getting... it seems that the rbx stays blow the temp curve of hte mcw6002 the entire time..

[Ancient_1]

It was discussed on ProCooling and best explanation was mounting pressure. Robo uses staock mounting hdw.

[MaxxxRacer]

stock*

I know pH uses the same mounting springs for every test rig.. I guess you were right when you griped abut the mounting springs on the mcw6002. I need to get some new springs for my TDX for when it goes back on the a64.. using the axp springs isnt good enough.. I have to tighten them down all the way for good pressure.

[EnJoY]

enjoy, im glad you noticed that. its swifty's new top secret block that pH will be testing. it uses a 500watt TEC and is made of pure silver. its been said it can get a 6800ultra down to -20C.


Will this be available for mass production? A 500w tec? That makes the Dominator Pro CPU block look like a kids toy. Why such a huge tec when a 172w would be the suitable step up from their current 80w MCW50? Will the block also be available in non-tec form? Cool info.

[MaxxxRacer]

enjoy your a very guilable person. I was just kidding man. I guess I made it sound too serious. I though the -20 c and the silver nad the 500watt tec would have given it awy, but hey.. lol

btw im selling the goldgengate bridge for 200 bucks if ur interested :p:

lol the dominator is a joke.. a phase change chiller will use less power and cool your entire loop. and it costs LESS if you know how to do it.

[Butcher_]

The dominator is just silly. 600W power draw for just the TEC, that's just crazy.

[EnJoY]

Um...so was the entire MCW60 thing a joke or just a typo? :confused:

You're making me look very stupid.

[MaxxxRacer]

its ok enjoy. implied humor doesnt work well in text. At first it was a typo but when you asked if it was some new top secret block i had to turn it into a joke..

buterch the dominator isnt 600 watts is it?? if so, there is not a watercooling system in existence that can coool that effeciently.. lol.. thats just nuts...

[EnJoY]

The new Koolance PC3 and EXOS 2 can cool 700watts Maxxx! 700watts!!! :rolleyes: :D :banana4:

[MaxxxRacer]

ROFL!! if that thing can cool 700watts my setup can cool 808080808808080504585946 watts....

with 700watts put into the exos the tubing would melt and the solder job on the rad would start to melt... lol...

[EnJoY]

Implied humor Maxxxy! :P

But yea, it's amazing what companies will put in their descriptions. They must have had that unit outside in Alaska to get numbers like that.

[MaxxxRacer]

lol.. did they really say that???.. i think they mean that it could be used as a space heater with 700watts were applied...My guess is that they were using an air conditioning unit to cool the rad... lol

[Ancient_1]

My guess is they are using a delta of probably 40° between water and air temps. I believe thats what HW Labs use when they rate the BI series.

[MaxxxRacer]

ROFL!!! that is nuts.. Delta T of 40C.. you only get that when your system is overheating like mad and your within 1000miles of either the north or south pole....

[Ancient_1]

When you hear a claim like the 700watts of heat dissipation for a setup like that you know they arent talking 5~10° delta, which is pretty comon for many watercooled setups. Given a big enough delta between air and water you could claim a BIM could dissipate 700watts. Its all just marketing.

[MaxxxRacer]

lol... yah if you have boiling water in it and you have super chilled LN2 gas blowing through the radiator i bet you could get 1000 watts of disipation with the BIM. of course this is getting air blown through it with the nasa wind tuneel funneled into a 6 inch diamteter tube... lol... talk about static pressure.

[Ancient_1]

Actually with a Delta EHE you would need a delta of about 70° to dissipate 700watts on the BIM.

[MaxxxRacer]

really... lol.. thats not too bad.. lol.. that is only ner boiling water and slighlyt cooler than normal air to do that...

[Ancient_1]

The whole point I was trying to make tho was without knowing what the delta in temps they were using a 700watt rating means nothing.

[MaxxxRacer]

i know.. im just having some fun laughing at the way that numbers can be twisted around... i should sell a product like that... i could claim 9000watt disipation if i dipped the rad in a container of LN2 and pumped a chemical with a higher bolinng point than water but less than that of hte solder and coppers melting point..

a delta T of around 250C.....

[Butcher_]

buterch the dominator isnt 600 watts is it?? if so, there is not a watercooling system in existence that can coool that effeciently.. lol.. thats just nuts...
Spec is 24V, 25A - that's 600W by my book. Plus the CPU of course...