What happend to the Nexxos XP review?

[lolito_fr]

as those seeking very low noise don't O/C. And addition of a GPU block wouldn't matter as you point out.

Sure, but not quite what I was saying.
It seems you're a little blinkered into thinking that flow is all
Why wouldn't a system with two high performance Nexxos blocks in series perform on par with say 2 MCW6000's? It doesnt seem obvious to me...(without plotting PQ of pump and system)

[Etacovda]

Because with even one nexxos Xp in the loop, flow rates start to suffer

Add two, and double the restriction?

In saying that though, even if flowrates dropped to 0.6gpm, the nexxos would still be 'winning'.

On a (more on topic) note, anyone have internal images of the design of this block? I found some a while ago, but sadly can't seem to dig them up again.

[Blackeagle]

Originally Posted by freeloadingbum

Considering the large die size and generally low wattage (50 watts max) of gpu's, you may have a tough time trying to measure any difference and an even harder time actually noticing that difference in any game

With the G5's .2 gpm you are right, measuring the slightly lower temps wouldn't be easy. Same with the increase if any of the O/C of the GPU with the G5.

But with the .45 or .82gpm flow increse of the other two Little River blocks you'll have less problem measuing it. And while the size of the O/C increase might be hard to see in day to day gaming it wouldn't be hard to see if you bench the cards. Hell that's true of many differant improvements of the past. They are often not really that large taken by themselves, but they add up as further improvements combine with them.

50 watts max, for a highly overclocked card? ? ? Source?

[Blackeagle]

Originally Posted by lolito_fr

It seems you're a little blinkered into thinking that flow is all

Not at all. But flow rates are far from meaningless as you seem to want to suggest.

My point, and this is the last I'll try to help you understand it, is this.

Nexxos performance is very good, but at the cost of a very high pressure drop.

Little River SS Cascade, G4 & G5 ALL offer lower pressure drop & at the same time giving equal (SS Cascade) or better temps (G4 & G5)! !

A further example.

A Nexxos installed into a system of pump + rad + Nexxos had a flow rate for the loop of 1.4 gpm. Swap out the Nexxos for the SS Cascade & flow increases to 2.2gpm due to lower resistence of the SS Cascade. With the SS Cascade you have a 50%+ increase in system flow vs the Nexxos, yet both give the same CPU temps. Not something you can ignore.

A Fusion GPU block is added into the same loop with the SS Cascade & the GPU card is max overclocked to XXX speed. Now exchange the SS Cascade for the Nexxos, still leaving the Fusion in the loop. But with the large reduction in flow due to the higher resistence of the Nexxos you now must reduce the GPU cards overclock by XX% to avoid the card artifacting, lock ups ect.

Pressure drop & flow rate interest me in this case as all 4 of these CPU blocks are within less than 1C in performance. So the point of interest in the Nexxos, along with it's low temps, is in the pressure drop. These are all clear facts of the testing data shown up to this point. With such close temps, pressure drop is another way of comparing the Nexxos to other blocks.

[Blackeagle]

More good work pH! Thanks

[psychofunk]

Kind of seems like an apples to oranges comparrison to me. I mean who is gonna run a nexxos with an iwaki MD20RZ (to acheive the dt of ~9) and likewise who is gonna run a little river block with a Eheim 1048 (to acheive a dt of ~11? They are two nice blocks with two diffent putposes. I'm outta here before this turns into a tubo inline4 vs V8 arguement.....








oh yeah and Little River Owns all!!!!

[Blackeagle]

Originally Posted by psychofunk

I'm outta here before this turns into a tubo inline4 vs V8 arguement.....

OMG! forgot to have 10 characters.

[Cathar]

Originally Posted by psychofunk

Kind of seems like an apples to oranges comparrison to me. I mean who is gonna run a nexxos with an iwaki MD20RZ (to acheive the dt of ~9) and likewise who is gonna run a little river block with a Eheim 1048 (to acheive a dt of ~11? They are two nice blocks with two diffent putposes. I'm outta here before this turns into a tubo inline4 vs V8 arguement.....

Hmmm

Storm/G4, Eheim 1048, HE120.2, 1/2" = 0.9gpm (see graph)

XP, Eheim 1048, HE120.2, 1/2" = 0.75gpm (see graph)

From Phaestus's chart we're looking at ~9.5CdT for the G4, and ~9.8CdT for the XP.

Okay, so the E1048 is too powerful for a "low flow" scenario? Let's try the Eheim 1046.

Storm/G4, Eheim 1046, HE120.2, 1/2" = 0.75gpm
XP, Eheim 1046, HE120.2, 1/2" = 0.6gpm

G4 @ 0.75gpm = ~10.2CdT
XP @ 0.6gpm = ~10.5CdT

Hmmm, so what was the argument over high flow vs low flow again? Or is some low-flow proponent going to say that we should restrict the XP with 8mm ID tubing and a looping coiled 8mm ID tubed radiator and will that make it perform better?

As far as I'm concerned, Phaestus's tests of the XP finally breaks apart the whole high flow vs low flow myth, just as I knew it would.

[lolito_fr]

awww crap.
(deleted)

[freeloadingbum]

Originally Posted by Blackeagle

50 watts max, for a highly overclocked card? ? ? Source?

http://www.xbitlabs.com/articles/vid...v-power_9.html

If you deduct for voltage regulators, ram and everything else, 50 watts should be a safe high estimate at the gpu. Also, in a system with a mcp600 and double rad, the flow rate difference should be alot less than your estimates.

My personal experience with a mcw50 showed no loss going from 1.5 to .75 gpm using atitool to overclock with on a 9800

[Cathar]

While we're also discussing pressure drops and their impact on flow rates (/me cheers as this is really what all this stuff is about), we also need to consider that radiators are not static with respect to varying flow rates too.

Throw in a super restrictive block and peg the flow rates back by 20-30%, and if you're already operating just below the 1gpm mark, this will make a significant difference in the performance of heater-core style radiators:

http://thermal-management-testing.com/Thermo12.gif

Ideally we don't want to be falling much below 0.70gpm/2.5LPM if we can help it, even for a weak pump like an Eheim 1046.

[lolito_fr]

...enter the looping coiled 8mm ID tubed radiator :shrug:

[BillA]

Originally Posted by lolito_fr

...enter the looping coiled 8mm ID tubed radiator :shrug:

for less performance

[psychofunk]

Originally Posted by Cathar

Storm/G4, Eheim 1048

This just seems wrong to me. I would never put a tiny winy pump with your block. I would just get a Swiftech 6002A. Just seems like putting $13K of engine mods into a civic (the 1048 being the civic and your block being the $13K). And please don't misunderstand me I am not knocking your blocks or their price I am only saying that to spend the cash on a Storm and stick it with a whimpy pump is not what I would do.

[Cathar]

What the XP test shows is that so-called "low flow" blocks need just as strong, if not stronger, pumps than the much lamented "US-style high flow" blocks, in order to achieve their best performance.

The shape of the Swiftech MCW6000's performance vs flow curve would appear to be a better overall solution for the truly "low-flow" scenarios (<0.3gpm) than any of the so-called low-flow optimised blocks. The MCW6000 is also significantly less restrictive, so in super-restricted setups (i.e. Eheim 1046's with 6-8mm ID tubing) it will offer better performance from the radiator too when matched with any given centrifugal pump.

Where the XP's very marginal performance benefit lies is in the rather narrow range of 0.3-0.6gpm, and even then I would hazard a guess that the MCW6000 would be a better solution on the basis of radiator performance. Really need to get Phaestus to show some 0.25gpm results for the MCW6000. I reckon it would be in front of the XP in the 0.25-0.5gpm range as well.

So where does that leave the "low flow" argument? On one hand we have the MCW6000 offering a better solution for the super-restricted setups that the low-flow proponents argue for. On the other hand, in order to get above what the MCW6000 can offer you start to tread into the realm of requiring the powerful and higher pressure head pumps, where you then hit an increasing number of competitive, and even superior, "high flow" style designs such as Storm, Cascade, White Water, RBX with #5 nozzle, and still the MCW6000 hangs in for a fair while too once we factor in their lower pressure drop, not to mention the better radiator performance as a result.

The "low flow" blocks are no different to most of the "high flow" blocks. They still require big and powerful V8's to get the most out of them. One could argue that about the only difference is the width of the tyres, with the "high flow" blocks allowing more contact patch to get the power down to the ground.

[nigelyuen]

Originally Posted by Etacovda

On a (more on topic) note, anyone have internal images of the design of this block? I found some a while ago, but sadly can't seem to dig them up again.

the one on the left is the xp, the other one is 1asl2

[JWFokker]

I don't think anyone will argue that for a mass produced block, you can't get better performance. The NexXxos XP will be available for a lot longer than the Storm. Despite the much larger quantities the Storm blocks are available in, they are still a limited run. When they're done, they're done....until they're taken apart and copied by another manufacturer, but until then, this is the best mass produced block you can get, and it is a significant improvement over the previous best. Not quite Storm performance, but close enough. The Storm is a much better product - better mounting and construction, not to mention the fittings are more suitable for 'American' style watercooling than the ones on the XP. But when all the Storm blocks are gone, this is going to be the block to get.

[pHaestus]

http://phaestus.procooling.com/temp/...exxos-flow.jpg

some more pics

[Cathar]

Originally Posted by JWFokker

I don't think anyone will argue that for a mass produced block, you can't get better performance. The NexXxos XP will be available for a lot longer than the Storm. Despite the much larger quantities the Storm blocks are available in, they are still a limited run. When they're done, they're done....until they're taken apart and copied by another manufacturer, but until then, this is the best mass produced block you can get, and it is a significant improvement over the previous best. Not quite Storm performance, but close enough. The Storm is a much better product - better mounting and construction, not to mention the fittings are more suitable for 'American' style watercooling than the ones on the XP. But when all the Storm blocks are gone, this is going to be the block to get.

Really? Once radiator performance is taken into account, and ignoring the blocks that I make, I would be more inclined to argue that for anything up to an Eheim 1250 sort of pump that the MCW6000 would be a match for the XP, with the MCW pulling a clear lead as the pumps get weaker and the rest of the setup is made more restrictive. Above the Eheim 1250, pumps like the MCP600 and better would tend to imply that either the White Water or the RBX w/ #5 nozzle would start to take over once radiator performance is factored in.

The XP is an extremely impressive block, no doubt about it, but its level of flow restriction coupled with the impact that this has on the overall picture can't be ignored - it is extremely important to consider what the radiator is doing as well, because unlike artificial testbeds people aren't using chillers/heaters to hold water to a fixed temperature.

Depending upon what pump you choose there already exists a few options on the US market today that match or exceed what the XP would do for you once the whole picture is taken into account.

[BalefireX]

What ever came of groth's experimental graphing system that showed the performance of a block when combined with a certain pump? I think that it would be very helpful here to see the comparative performance (given identical tubing/radiators/fans) of the Storm G4, Nexxos XP, MCW6000, etc using a couple of pumps; say a 1046, 1048, MCP600/650, and an Iwaki.

[Cathar]

Originally Posted by BalefireX

What ever came of groth's experimental graphing system that showed the performance of a block when combined with a certain pump? I think that it would be very helpful here to see the comparative performance (given identical tubing/radiators/fans) of the Storm G4, Nexxos XP, MCW6000, etc using a couple of pumps; say a 1046, 1048, MCP600/650, and an Iwaki.

Not sure about his work, but I've got one that's very close to full completion that I'm presently using before making any statements. What I'm missing though is the system vs pump flow prediction system. Presently I just plot the two, find the intersection point by eye, and then type that into the simulator.

[Blackeagle]

Originally Posted by freeloadingbum

http://www.xbitlabs.com/articles/vid...v-power_9.html

If you deduct for voltage regulators, ram and everything else, 50 watts should be a safe high estimate at the gpu. Also, in a system with a mcp600 and double rad, the flow rate difference should be alot less than your estimates.

My personal experience with a mcw50 showed no loss going from 1.5 to .75 gpm using atitool to overclock with on a 9800

Well first I really did enjoy that article, thanks.

Please note that the O/Cs of the cards are on stock air, good after market copper coolers can do a bit better & not at all what I'd call high for a water cooled card with good copper sinks on the memory and a strong side panel blower blowing cool air over the card. There's a guy at Xtreme with a non Ultra 6800 at 492/1290 on water......I suspect there's some Ultras on water well over 500/1300. And those would be some high wattage cards.

Not sure what your referance to the MCP600 and double rads is about. A Iwaki MD20 could also be said to offer higher flows & larger differances between blocks than pH's numbers, but so what, (his set up has much higher resistences due to measurments taken.) I didn't estimate those differances in flow, they are from ph's testing data of the differant blocks as I posted. Check out the Pro Cooling interactive graph. And pH's numbers are accurate, which is a major help in comparisons.

I've not used the MCW50 block. But a open chambered and thick based block will do the most to dimenish the impact of flow changes. A block like the Fusion with a thinner base & micro fin internal structure will show up flow changes quicker.

On average I expect higher flows to result in higher clocks of a vid card. Every card is differant, but over a number of cards added cooling will show increases.

[Blackeagle]

Originally Posted by Cathar

Not sure about his work, but I've got one that's very close to full completion that I'm presently using before making any statements. What I'm missing though is the system vs pump flow prediction system. Presently I just plot the two, find the intersection point by eye, and then type that into the simulator.

Such a interactive set up that could include pump & blocks would be better than blocks alone of course. Has to be one hell of a lot of work to plot the numerous combinations.....

The one intended to include rads as well would be a even bigger nightmare when you factor in heater cores, tranny coolers and......and ......

[Cathar]

Originally Posted by Blackeagle

Such a interactive set up that could include pump & blocks would be better than blocks alone of course. Has to be one hell of a lot of work to plot the numerous combinations.....

The one intended to include rads as well would be a even bigger nightmare when you factor in heater cores, tranny coolers and......and ......

The simulator I have allows for varied CPU heat, GPU blocks and varied GPU heat, the impact of pump heat, and radiators. The only thing it doesn't really do is pre-determine the system resistance and measure that against a pump PQ curve. That bit is still manual.

[Etacovda]

Originally Posted by nigelyuen

the one on the left is the xp, the other one is 1asl2

Dont happen to have an image of the top plate? i know it has jets in it, just curious to the layout etc