Stren's 2012 CPU Water Block Roundup

[st0ned]

Well that's true, also some of the ones with fancy new things such the one from phobya have a low/midle bow so it might get some boost from a flatter IHS such as the one from a 2011 CPU.

[musicfan]

Nice photos. Nice presentation. Cool video on your test setup too. Thanks for changing/fixing the camera part-way through. I look forward to your testing. Please take your time and keep it fun for yourself so you do this many years. Very subscribed.

[stren]

Finished the unboxing pics. It's going to be a busy month getting this all finished up!

[Martinm210]

Nice!
I look forward to your results!

BTW, Are you sure your King Flow meter indicator isn't inserted backwards?

Yours has a different floating cone of sorts, but I have three of them and the tapered portion always flares upwards narrow to wide with the direction of flow). I suppose you could verify accuracy by doing a 5 gallon bucket/stopwatch test as I did here, but that caught my attention. Now that I look at it, it might be ok and just a different float/indicator than my three.

[stren]

Nice!
I look forward to your results!

BTW, Are you sure your King Flow meter indicator isn't inserted backwards?

Yours has a different floating cone of sorts, but I have three of them and the tapered portion always flares upwards narrow to wide with the direction of flow). I suppose you could verify accuracy by doing a 5 gallon bucket/stopwatch test as I did here, but that caught my attention. Now that I look at it, it might be ok and just a different float/indicator than my three.

Interesting - I haven't calibrated it. I have got it hooked up right but I never did anything to change the float. I bought it off ebay supposedly new, and it looked brand new, so I'm not sure what to make of it. I'll have to do your 5 gallon test and confirm!

I found this image which looks the same as yours:



Here's a close up of mine:



They are different shapes so it's not just the same float but upside down. Plus mine is the 3.5gpm range not the 5gpm like the others I've seen - still it's weird!.

[Martinm210]

Ah...it's probably supposed to be that way because of the different range. 3.5 is a better range too.

[bds71]

looking forward to seeing the results - will they be listed as done? or all at once?

[stren]

looking forward to seeing the results - will they be listed as done? or all at once?

To be honest I'd like to release them all at once so I'm confident of the results. It also means people won't see half the blocks and run out and buy something when really there's more to come. I'll release the data based on IX first and then go back and run MX2.

[stephenswiftech]

To be honest I'd like to release them all at once so I'm confident of the results. It also means people won't see half the blocks and run out and buy something when really there's more to come. I'll release the data based on IX first and then go back and run MX2.

Makes sense =)

[bds71]

thanks - as i said before: i look forward to the results!!

[gabe]

Stren, we had users asking us why we didn't submit and Apogee Drive II in this roundup. I have no problem with that. Let me know. gabe@swiftech.com

[LeoAndrade]

Looking foward to see the results!

[Lanek]

My heart will make me think the supremacy on top,.... but .. many other brand have surprise me lately with their block ....

Huum, well a lot of work are waiting you... this make a nice numbers of block to test.

[Church]

gabe: it will be hard to put it as apple to apple comparison imho. Only if he'll switch off one of pumps @his dual top, to even out pumping power .. hmm .. but then won't switched off pump produce extra pressure drop? :/. Or if one intends to test impact of extra pump on ADII, flow increase/cooling impact from i's pump being 3rd will be too little (diminishing returns), and also three pumps per loop seems rather uncommon config.
sten: do you have single MCP35X pump alongside MCP35X2?

[gabe]

Sure, 3 pumps is a bit extreme. Yet, APD2 is a leading waterblock solution, and I can understand why users ask about it

Stephen, any suggestion ?

[stren]

Sure, 3 pumps is a bit extreme. Yet, APD2 is a leading waterblock solution, and I can understand why users ask about it

Stephen, any suggestion ?

Stephen originally suggested the APD2 because it hadn't been reviewed before, but I thought the HD would be easier for the apples-to-apples reasons. If we can come up with a fair way to slot it in, I'll run it also

Even with 3 pumps the test setup should handle it- the flow meter runs up to 3.5gpm and the DT 5Noz which is so far the least restrictive block was just hitting 3.5gpm with the Iwaki RD30 pump @ 29V. In comparison the MCP35X2 at max was hitting 2.6gpm (or ~1.8gpm with one pump powered down), so I think 3 pumps will still be less than the RD30. My 400mm cylinder reservoir does now have a crack though which I attribute to the massive power of the Iwaki

With another PWM splitter I should be able to control the flow well - so for the tests vs flow rate it would be fine, the real question is at what point to fix the flow for the fixed pump power tests.

I could turn it off or just adjust PWM so that the pump flow is the same as the HD? The HD and APD2 have the same insides though I thought, so then we'd just get data that's very close to the HD no?

Anyway I best get back to work on the testing!

[stephenswiftech]

Sure, 3 pumps is a bit extreme. Yet, APD2 is a leading waterblock solution, and I can understand why users ask about it

Stephen, any suggestion ?

other than testing with a single pump, I don't see any other solution. 3 pumps is neither realistic nor a good way to evaluate the product, because with 3 pumps you already are well into diminishing returns.

If the scope of this test is to compare WB in typical LC loops (i.e. most common loops), I'd say a single pump is probably THE best way to do it. In that case, testing the APD2 shouldn't be a problem. Sure you have 2 pumps but it's still a water block which brings added pumping capacity and redundancy. It will not be comparable per se in a Pressure/Flow curve but performance wise it would still be very comparable as long as you don't have a super restrictive loop (which tends to give the APD2 a clear advantage)... But for simple loops: rad / pump / WB the difference from the additional pump isn't going to be substantial.

In my (designer) eyes, the APD2 fits 2 different markets pretty well: 1. people would want a compact system with superior performance. 2. high end users would want a top of the line block and the added safety of a fully redundant system thanks to the second pump.

Right now, 1. is pretty clear for most people but so far 2. isn't. If it's because reviewers are concerned about the fairness of the testing then just know that designing a test setup suitable to compare the APD2 to any other block is completely possible and very easy (like I said: avoid super restrictive test bench and run a single pump loop).

[stren]

other than testing with a single pump, I don't see any other solution. 3 pumps is neither realistic nor a good way to evaluate the product, because with 3 pumps you already are well into diminishing returns.

If the scope of this test is to compare WB in typical LC loops (i.e. most common loops), I'd say a single pump is probably THE best way to do it. In that case, testing the APD2 shouldn't be a problem. Sure you have 2 pumps but it's still a water block which brings added pumping capacity and redundancy. It will not be comparable per se in a Pressure/Flow curve but performance wise it would still be very comparable as long as you don't have a super restrictive loop (which tends to give the APD2 a clear advantage)... But for simple loops: rad / pump / WB the difference from the additional pump isn't going to be substantial.

In my (designer) eyes, the APD2 fits 2 different markets pretty well: 1. people would want a compact system with superior performance. 2. high end users would want a top of the line block and the added safety of a fully redundant system thanks to the second pump.

Right now, 1. is pretty clear for most people but so far 2. isn't. If it's because reviewers are concerned about the fairness of the testing then just know that designing a test setup suitable to compare the APD2 to any other block is completely possible and very easy (like I said: avoid super restrictive test bench and run a single pump loop).

Maybe the answer is to add it in when it comes to retesting with MX2, then drop the mcp35x2 out of the test rig and move to the mcp35x for all other blocks, and remove it for the APD2. For the MX2 I was going to run the mcp35x2 at an equivalent single pump power setting anyway. The test rig is pretty high flow so performance should still be similar to a regular loop.

[Church]

stren: btw, there might be rpm/performance variations even between pumps of same type, so if you switch out double top to single one, i suggest to remount pump itself. Though maybe i'm taking that a bit too far

[Martinm210]

If you want apples to apples on the APD2 why not simply match up the number of running 35x motors?

35x plus APD2 vs 35x2 plus standard blocks

Although you should also show the APD2 without the extra pump and other blocks with just one.?

I would guess that while we all like toying with two pumps in series, the majority >90% of users run a single pump solution so that is most important. Heck, I don't even like to run one at full speed so a 35x at 50% or so PWM is probably what I would be most interested in. I know people like thier pissingposts, but testing with two or three pumps is realistically meeting a condition probably less than 1% of the users out there will operate in daily.

Typical use would be a single pump motor, MX-2, and probably cpu+gpu type restriction levels.

Of coarse you will never satisfy the routine "WE WANT MOAR FREE WORK" mob, so you just have to make a plan YOU want and stick to it or you will never be done. Just know people never stop asking for MOAR..if you hadn't noticed already.

Oh by the way can you do the MX-2 round first?

Sorry..couldn't resist..

[RoadconeTuning]

if i didnt already have a pump with a top i would have gone with the ApogeeDrive2 but i went with the EK Supremecy to match my EK 5850 block

[tiborrr]

Good work Simon, I hope you have solved your IX TIM reflow issues. I would also recommend testing the blocks using a single pump as this is the most common application.

[bds71]

the real question is at what point to fix the flow for the fixed pump power tests.

is it possible to "fix" the flow rate for your loop (regardless of number of pumps)? i'm thinking some kind of ball restricter that artificially restricts the total loop flow to that of the least flowing CPU. note: you would have to adjust each loop manually, but it shouldn't take that long, and would be pssible to "fix" any loop at a specified flow rate...that is about as apples to apples as it gets, and would allow you to test the ADP2.

[rge]

is it possible to "fix" the flow rate for your loop (regardless of number of pumps)? i'm thinking some kind of ball restricter that artificially restricts the total loop flow to that of the least flowing CPU. note: you would have to adjust each loop manually, but it shouldn't take that long, and would be pssible to "fix" any loop at a specified flow rate...that is about as apples to apples as it gets, and would allow you to test the ADP2.

Actually that is not apples to apples imo, that is removing one of the important variables. Also it gives a poor representation of what the end user can expect. Part of how a block functions is 1) contact (bow size, flatness, etc) 2) internals/surface area/jet plate/flow restricton. Why are you trying to eliminate one of the important variables as to how a block performs? An end user isnt going to adjust his pump speed down on a good flowing block so it performs like a more restrictive one. To me that is like saying why not change all the internals of a block so they have the same surface area. To me you use the same pump speed as an end user does, and the restrictiveness of the block is part of its performance.

[musicfan]

Agree with rge. Thermal and flow performance (restriction) are both important test parameters. Fixing flow rate can favor more restrictive blocks and is not what most users do.