About #1 water block in the world

[EluSiOn]

what do you guys think about waterplanet.de's testing on water blocks?

here is the translation

http://translate.google.com/translate?u=http%3A%2F%2Fwww.watercoolplanet.de%2F index.php%3Fopen%3D4%26show%3D1&langpair=de%7Cen&hl=en&ie=UTF-8&oe=UTF-8&prev=%2Flanguage_tools

[ToolBox]

I think as BillA once said himself "minimalist, at least" Small tubing and we can't infer how much the pump is actually pumping. Unlike JoeCs whose is controled at 1GPM. There is a thread of this over at procooling.

[vonkaar]

And there have been several threads over here on it =p.

The testing platform is lame... they crippled the Cascade... non-standard flow rate...

The statement *could* be said that a 'true' top-block would perform better than any block, regardless of the platform. I would answer that with a, "well, don't use a Cascade unless you are putting it in the 'right' type of system, dumbass."

[EluSiOn]

that is what I figure too...

they use.. 1/8 tubing and also... eheim 1048 pump to test... awwww!

I feel that poor cascade is being raped by using a 3/8

[nikhsub1]

Originally posted by EluSiOn
that is what I figure too...

they use.. 1/8 tubing and also... eheim 1048 pump to test... awwww!

I feel that poor cascade is being raped by using a 3/8
I believe they actually use a 1046 to test... even worse.

[EluSiOn]

For the enterprise we use a Eheim 1046 with attached header tank. We decided for the 1046, since 300 litres are sufficient delivery/hour.

:eek:

[CrashOveride]

ya... notice the small block with small tubing and stuff all win over the larger blocks. It almost HAS to be the tubing and pump. From what I understand some people (esp in europe) watercool for silence over the extra cooling capacity. So I guess depending on the application those blocks might be better, but for what we want the cascased almost surley wins out.

[Cathar]

Take a good look at the radiator. Low flow restriction it is not.

[Lothar5150]

Cathar-Nice block looks very well made. I have a question for you...When you set about designing a new block do you use Reynolds numbers in modeling turbulence and flow? Or just experiment with different internal shapes.

[johan851]

WHY do people run setups like that? It's...it's against everything overly large and American! AHHHH! :D

[Soja]

Originally posted by johan851
WHY do people run setups like that? It's...it's against everything overly large and American! AHHHH! :D

Maybe because they're not Americans :p? Wow, 4 more blocks take the cascade since I last browsed watercoolplanet :eek:

[Korndog]

and they're like 30 euros cheaper then cascade, too good to be try right? lol

[Cathar]

Originally posted by Lothar5150
When you set about designing a new block do you use Reynolds numbers in modeling turbulence and flow? Or just experiment with different internal shapes.

Both.

[JFettig]

just because us americans do it doesnt mean we are right.

I am personally in favor of small(er) tubing.

Jon

[Cathar]

Originally posted by JFettig
just because us americans do it doesnt mean we are right.

I am personally in favor of small(er) tubing.

Jon

Smaller tubing with weak pumps present a hammer/nail scenario - as in when all you have is a hammer, everything looks like a nail.

In order to "best take advantage" of ultra-low flow rates, generally the base-plates need to be made thicker which raises the thermal conductivity cost through the copper, impingement-style cooling starts to lose efficacy, and the blocks need to heavily rely on very fine channels/pins in order to best make use of the low flows.

The problem with doing this though is self-defeating as one attempts to raise the flow rates. A very finely structured block offers very high flow resistance making the use of larger pumps almost futile. The thicker base-plates that are needed with the low flow rates also effectively "cap" the amount of gain that can be seen as the flow rates are raised as the metal conduction "costs" get in the way.

So if one designs a waterblock that naturally focuses on the assumption of 0.25-0.5gpm flow rates and optimises for that design point, then by token of doing such the ability to see substantially better results with higher flow rates gets nullified, and one is lead to the false assumption that higher flow rates yields little return. Thus the hammer/nail analogy.

The more modern waterblocks coming out of Australia/USA make use of thin-bases and impingement. These blocks are generally designed to work well with 1-2GPM flow rates, which are achievable with moderate pumps (Eheim 1250, Hydor L30, Via Aqua 1300, etc). Run these modern blocks at 0.25GPM and their performance will be around 5-6C worse than at 1.5-2.0GPM. This is not to say that they suck at 0.25GPM, they still perform very, very admirably generally being right up there with the best of most any other block, but to run them at 0.25GPM is to deny these blocks of the sorts of flow rates that they were designed to run with. Give them a chance to stretch their legs and they won't disappoint. Where the super-fine low-flow optimised blocks will choke the more powerful pumps, the modern blocks thrive on it.

It's not a matter of who's right or wrong. It's a matter of understanding the reasons behind each design methodology. The Aussie/USA blocks will eat the low-flow-optimised blocks for breakfast given a setup that plays to their strengths. Throw those Aussie/USA blocks on a low-flow testbed, and you're only ever seeing a shadow of what they're capable of.

I think the following round-up chart from BillA best illustrates my point. Note that the White Water (impingement, thin-base) starts off as being about the same as the MCW5002 (pin fin, thicker base) at 0.25GPM. Hardly a strong showing for the WW. If all we did was test the WW at such a flow rate, we would conclude that is was a decent block, but nothing special. Now give it 1.5-2GPM and it's about 0.03C/W clear of anything else (~3C on a hot CPU), and all of a sudden it looks a lot more special. Note that the MCW5002 with the pin-fin design and thicker base-plate doesn't achieve as much of a benefit from using a stronger pump. I personally predict that this is giving a good insight into the reasons for why the watercoolplanet is not really indicative of anything much at all, except for performance at ultra low flow rates.

http://thermal-management-testing.com/summar3.gif

In fact, take a good look at the PolarFLO. There's about an 0.08C/W (~8C) difference between 0.25GPM and 2GPM. If all we did was test it at 0.25GPM, then we'd conclude that it was a pretty crap block, when in fact it does fairly well given the flow rates that it was designed for.

[johan851]

I feel a need to point out that this:
WHY do people run setups like that? It's...it's against everything overly large and American! AHHHH!
Was entirely sarcastic. In fact, it even makes fun of America by saying "overly large and American." Thought I'd point that out......

[#Rotor]

to me it looks like asking any German which is better.... a Porsche or a corvette..... not that a Chevy is Australian... but close enough....

as far as comparing blocks for blocks... the only way that can be fair to all, is if the "black-box" method is uses...

[EluSiOn]

i lived in us and I have to say porche kicks corvette in every way.....

sorry... off the topic...

but #rotor is right... everyone has a preference.

[gusgizmo]

though this has been covered, id still like to say it myself, that using 1/8 tubing is sickening. i dont see how using smaller tubing makes the system very much quieter. just submerge the pump and be done with it. maybe even run the radiator outside.

[JFettig]

I think we arent talking about 1/8" tubing, they are talking about 3/8" od tubing I think What I am talking about is 1/2OD tubing,
I agree, larger tubing often if not always leads to better temps, but then something you have to consider is: is it worth it? all that hassle for 1 degree?
Thats the main reason I am in favor.

Jon

[Freeloader]

By Cathar: In order to "best take advantage" of ultra-low flow rates, generally the base-plates need to be made thicker which raises the thermal conductivity cost through the copper, impingement-style cooling starts to lose efficacy, and the blocks need to heavily rely on very fine channels/pins in order to best make use of the low flows.

It appears to me that the blocks in the top spots are using thin bases since they are only using a small area for heat transfer. I'm not certain about the blocks in 1st and 3rd spot as there is no internal pictures. It looks like all the thick base blocks are doing quite poorly.

By Cathar: It's not a matter of who's right or wrong. It's a matter of understanding the reasons behind each design methodology. The Aussie/USA blocks will eat the low-flow-optimised blocks for breakfast given a setup that plays to their strengths.

I disagree. Assuming that the test results are accurate, the top block has a 2.6C advantage over the WW with an unknown flowrate. Let's assume .25 to .5 gpm for the WW. I would think that in a normal decent setup with 1/2 inch tubing, big Arse rad and Eheim 1250 pump, that the WW would get about 1.5gpm flow. With that flowrate, the WW should gain 3 to 4C depending on the prior flowrate. If we apply the same tubes, pump and rad to the 1st place block, at least a 1C gain shouldn't be too far off. This means that either block might win by a small margin.

Of course this is based on alot of assumptions, but until there is more tests, I don't think it's clear who will be having who for breakfast.

[johan851]

The difference we would get by going with a more powerful setup is a little different though. If we assume that the system is regulated to 2gpm, the blocks aren't going to have the same flowrate. The Cascade might still get 1.5gpm through, while the very small blocks might only get 1gpm or less through. The problem with going for high flowrates with such small blocks is that the smaller blocks are just too...well...small to achieve much higher flowrates. That's why (I think) the Cascade would perform so much better in a higher flowrate system.

[Freeloader]

johan851: The Cascade did poorly, however, and its base is so thin that excessive lapping isn't recommended. I'm not sure if that theory holds true for all the blocks in the test.

Poorly?, I though it did quite well. Just because it didn't perform the best doesn't mean it performed poorly. The results are close enough that it's questionable that it even lost.

[johan851]

Poorly for the Cascade. It came in about midway at 15.9k, while the best was 14.6 and the worst was 17.04. Usually it takes the lead by a good margin in American/Aussie type systems.

[Freeloader]

You missed the other four pages. The worst was 32C

[johan851]

Hmmm...you're right. I still think that the Cascade and other large-barbed blocks would benefit much more from a flowrate increase than the more restrictive blocks, however.

I will edit out the word "poorly."

[Freeloader]

johan851: Hmmm...you're right. I still think that the Cascade and other large-barbed blocks would benefit much more from a flowrate increase than the more restrictive blocks, however.

So do I, but is it enough to overcome their deficit? (excluding the Cascade).

[Cathar]

Impingement blocks benefit greatly from an increase in flow rates. Hard to know what the flow rate actually is. Doesn't help that the Cascade was user-modified. Doesn't help that WCP only ever do a single mount of each block.

We can sit here and debate the reasons for what WCP shows until the cows come home, but all of it is pure conjecture until the blocks are put onto a testbed that gives a proper analysis of the flow/performance curve, and attempt to remove the mounting variable.

I'd say we're looking at more like 5-6C differences for a 100W heat-load for blocks like the Cascade/WW/RBX between ~0.25gpm and 1.5-2.0gpm. 0.25gpm really is well below the performance knee of these blocks, and reaches the point of rapidly increasing temperatures as flow rates go down.

[Cathar]

Another important aspect that is being missed is the flow-restriction of the blocks. Given target flow rates of 0.25-0.5gpm one can design a very highly restrictive block (and indeed many of those blocks look very highly restrictive) and have it perform well. Problem is that these blocks may have a great deal of trouble getting much above 1GPM flow rates even with fairly powerful pumps, whereas many of the Aussie/USA blocks are designed to be free-flowing enough to allow at least 1.5-2.0GPM flow rates with moderate pumps (Eheim 1250/Danner Mag3/Swiftech MCP600).

There will be the issue of trying to predict apples to oranges of the performance of some of those blocks at >1GPM simply because the pumping power required to push that sort of flow rate is prohibitive.

Again, these are all elements that the WCP test does not show us.

[AngryAlpaca]

The Cascade tested there has two barbs, not three... The fact that it uses a different WB makes it inaccurate on its own.

[nikhsub1]

Originally posted by AngryAlpaca
The Cascade tested there has two barbs, not three... The fact that it uses a different WB makes it inaccurate on its own.
All the copper Cascades have only 2 barbs. Just some of the SS editions have 3. There is no real temp difference between 2 or 3 barbs.

[CrashOveride]

Originally posted by nikhsub1

All the copper Cascades have only 2 barbs. Just some of the SS editions have 3. There is no real temp difference between 2 or 3 barbs.

Weren the SS Cascades with 3 barbs custome for certain people (or just one for a person?) who wanted it more symetrical?