Test Report: Loop Order, does it make a difference?

[Vapor]

Preface
This test is somewhat of a detour from my normal waterblock testing, but one that was worthwhile--answering a long-debated question: does placement of the pump in relation to the CPU block matter? Specifically, does the increased inlet pressure from placing the pumps directly before the block increase performance as suggested by many.

The What
While my normal tests focus on how flowrate effects temperature performance of blocks (and how they compare to each other), this one is only comparing two scenarios: pump immediately before the CPU block versus pump then "everything else" then the CPU block. Theory says that the pressure drop of a block at a given flowrate is a non-variable quantity, so there should be no change in flowrate (though it is still measured) between the various options, so this test merely focuses on the position of the CPU block in relation to the pumps and what difference in temperatures are present.

Inlet pressure and pressure drop should not be confused...this test is isolating the change in inlet pressure (via changing the position of components in a loop) and how that compares to temperature. If I were to change the pressure of the pumping system, that would change flowrate, but I am not. If I were to change pressure drop, I would be changing flowrate, but I am not. I am only testing what happens when you increase inlet pressure.

If an increase in inlet pressure increases the thermal performance of a block, the pumps-before-block configuration will be superior. Gains are reported to be as high as 2C.

If an increase in inlet pressure has no effect on the thermal performance of a block, the rads-before-block configuration will be slightly superior simply because water leaving the pumps (in my case three DDC3.2s at at least 12V) will be slightly warmer due to the heatdump of the pumps. The difference should be insiginficant, a small fraction of a degree, but it should be there.

In this test I will be examining three waterblocks: 1) the Swiftech Apogee GTZ, a block with moderate restriction that I have confirmed to respond well to increases in flowrate (and therefore turbulence); 2) the Koolance CPU-350, graciously sent to be my NaeKuh for this testing (and future tests) because it's not only a flagship block, but it's a high restriction block and also a block has that has been reported to have an increase in performance based on pump positioning; 3) the Fuzion V2 + 3.5mm nozzle....simply put, it's the most restrictive block I know to exist and that's a quality that some theorize to have an impact in this scenario.

It's important to note that this is, in no way, a comparison of the three blocks. While I am using my normal block testing test procedure (and therefore the data is viable for cross comparison), that is not the objective. I'm running a non-stock mounting system on the KL-350 simply because the stock mounting system was not available to me. Also, three DDC3.2s in series in a CPU-only loop is a ludicrous amount of pumping power and does not paint the full picture for block performance. Those tests are partially completed and will be done when they're done.


The How

• The processor I'm using for this test is my C0/C1 i7 920. I'm running it at 21x196 (4120MHz) at 1.46V loaded on a Gigabyte EX58-Extreme. It is unlapped. I'm running 2GB of G.Skill DDR3 1600MHz at ~1570MHz. All heatsinks on the board are stock and I provide airflow over the mosfets to aid stability. The video card is a 4850 1GB with VF830 running in the top slot. The board is sitting on my desk alongside my Odin 1200W PSU and DVDRW and HDD drives.

• The watercooling loop I'm using is very untraditional, but allows me to test the way I want to test.
  • It consists of an MCR320 + MCR220Res sandwich with three Sanyo Denki "San Ace" 109R1212H1011 fans and 5 (3+2) 120x120x20mm Yate Loons cored out as shrouds. The sandwich allows for high-dissipation ability in a compact setup. The 'Res' part of the MCR220Res is used not as a res, but as a drain port.
  • For pumps, I use three MCP350s modded to MCP355s. One is attached to an XSPC Res Top and the other two are attached to the EK Dual Turbo Top--all three are in series. The MCP attached to the XSPC Res Top I can modulate the supply voltage freely between 7.65V and 12.65V. The two MCPs on the EK Dual Turbo Top always run at 12V. I have six pump settings I run with every mount: 1) All three on at full speed, 2) XSPC Res Top only (at 12.65V), 3) XSPC Res Top only (at 10V), 4) XSPC Res Top only (at 7.65V). The ability to consistently vary flow is a huge aspect of my testing.
  • I use a Koolance FM17 for my flowrate measurement. I recognize its lack of 'professionalism' (compared to a King Instruments flowmeter or something of that ilk) but still use it because it 1) covers the entire range I anticipate I'll be testing in (~.2GPM up to 3GPM), 2) outputs measured flowrate every second via RPM wire, which is logged for the entire test and then averaged and has thus far brought on extremely consistent results.
  • Loop order: CPU block -> MCR220Res -> Koolance FM17 -> MCR320 -> XSPC Res Top + MCP -> EK Dual Turbo Top + 2xMCP -> CPU block. Air flow order: in -> temp probe array -> MCR320 -> San Ace H1011 -> MCR220Res -> out

• I do a 5 mount test, each with their own TIM application. It takes a ton of extra time (each block takes 5x4x120min to test), but it's totally worth it. In the words of Martin "It's not uncommon at all to see mounting variations as high as 2 degrees or more, so with only one mount, that error is 2 degrees. When you mount 5 times and average those results, your standard deviation is significantly lowered and the overall testing confidence improved. In addition multiple mounts serve as a means to validate data, because each test is carried out again and again, chances are if some variable is affecting results, it will show."

• I have 10 temperature probes in use: 6 Dallas DS18B20 Digital one-wire sensors on the intake of my sandwich, 4 Intel DTS sensors in the processor.

• For temperature logging, I use OCCT v3.0.0.RC1's internal CPU polling that is performed every second on all four DTS sensors and is automatically output to .csv files. I also use OCCT for loading the CPU. For intake air temperatures, I use Crystalfontz 633 WinTest b1.9 to log the Dallas temp probe data on my Crystalfontz 633. I also use WinTest b1.9 to log fan RPM and Koolance FM17 flowrate output. Martin et al. have been over the many advantages and qualities of the Crystalfontz + Dallas temp probe combinations--it really is a wonderful setup and aids the testing process immensely.

• For processor loading, I find OCCT v3.0.0.RC1 to be extremely competent. It provides a constant 100% load (so long as WinTest b1.9's packet debugger is fully disabled) and is extraordinarily consistent. It allows me to, in one button push, start both the loading and the logging as well, which helps. I immediately also start to log the Crystalfontz data simultaneously. I run a 120 minute program, the first minute is idle, then I have 115 minutes of load, and then 4 minutes of idle. The first 26 minutes of load are thrown out as warmup and only the remaining 90 minutes of load are used for data compilation. During the last 4 minutes of idle, I adjust the pumps to be prepared to immediately begin the next 120 minute program.

• For TIM, I use MX-2. It's plentiful, representative of what a lot of people use, and has no break-in period. I use the dot in the center method and validate all my mounts to be at least "good" visually upon removing the waterblock.

• Like Martin, I have found that simply using processor temperature minus ambient temperature is not adequate for some processors. While my 65nm processors report an increase in CPU temps of ~1.22C for every 1C in ambient difference and I have to correct for it, my i7 processor does not have this deficiency and I've mapped out to be a perfect 1:1.

• My graphs....they may look a little different than what you've seen before, but I feel they're a great way to show all the individual data points from testing while also highlighting the averages of that data. I've termed them Planet/Moon graphs--each data point get its own moon and 3 moons get averaged into a planet. From there, the planets get a line drawn through them (not a trendline, just a regular line with the "smooth line" option checked). For something like flow vs. cooling, I've found Excel's trendlines to be totally incompetent. This applies to HSFs too. In fact, I have yet to see a situation where they do work involving flow vs. cooling.

• While I do 5 mounts, I discard the best and worst mounts and use the data of the middle three. I still show you the data from the worst and best, but it's not used in the 'big' graphs or the averages calculations. I take the middle three to hopefully get a fair representation of what to expect from the block in how it compares to other blocks.

Charts
I have no charts for this test....just one simple table

Tables
All applicable data from the three blocks I tested in one table
Temps are adjusted for 21C....yes my 920 runs really cool, I hate it even though it clocks well.



Conclusion
There's no change in flowrate and basically no change in CPU temps. The rad-before-block config performs slightly better, but the difference is insignificant (even with a lot of heatdump from the pumps). I think this myth is busted. Thanks for reading

[faster3200]

Thank you so much for this. We finally have actual data to point newbies and debaters to. Excellent job and write up.

[Expat GriZ]

Well Done Vapor.

I'm sure there will be "yah but" & "but what if" ....etc. However, thank you for taking the time & putting in the work to dispell the various theories & out right misconceptions.

[Fragger]

I agree with your result, having just changed my own loop order I see zero diff in temps.

[Martinm210]

Thank youthank youthank you

That's a ton of work, but so nice to have with all the past debates.

Don't mind my drooling over your mounting consistency, doesn't get any better than that.

Thanks for putting in all that work, much appreciated

I see the KL likes lots of pumping company as well..

[Vapor]

Thanks for the kudos guys

Glad to put the data up....I had questions myself and just wanted to see it first hand. Taking a little break before GOOC 2009 but I'll be back testing soon enough, might need yet another 30g tube of MX-2 sooner rather than later though

The Swiftech and D-Tek mounting systems are great for consistency--just keep screwing till it stops

The KL-350 I measured each screw-to-plate distance with calipers....kept it as close to .695in as possible (totally arbitrary and used a homemade mounting system borrowing parts for Scythe, Swiftech, Home Depot, K|ngp|n, and the Koolance mounting plate).

[twwen2]

Awesome stuff Vapor, myth busted!

We appreciate the time and effort.

[RubberDuck]

Thx Vapor for the Excellent hard work you have done

[shazza]

Excellent! Thanks for this ... I know it was a ton of work, but it will be an oft-quoted post in the future.

[santiagodraco]

Wow, interesting. And that's with 3 pumps, so you'd expect them to add enough heat to make a difference. But I guess they don't pump out as much heat as one would expect.

I think I need to add another pump into my loop. I'm getting about half your flow on my KL350, but I'm only running 1 Koolance 455 (D5) and I have two of the VL3 quick disconnects, and the 487 block (oh and the FM17 flowmeter).

[MpG]

Very interesting, and a test that was long overdue for someone to try. Thanks!

All that remains now is figure out wth is going on with Naekuh and Honda's cases.

[Waterlogged]

I guess I'm gonna be the only person here that thinks that testing this with 3 pumps is overkill and in a sense...fixing the results.

[Zehnsucht]

This must be inserted in the stickie.
Not that anyone reads them though...

[santiagodraco]

I guess I'm gonna be the only person here that thinks that testing this with 3 pumps is overkill and in a sense...fixing the results.

I think he was trying to show that even with three pumps the extra heat dump doesn't effect temps enough to be meaningful. I guess he could have hidden the actual block temp comparisons and show only the variances (but I have to admit to a certain fiendish delight at the results).

But you do have a good point, certainly the results should not necessarily be taken as demonstrating an absolute comparison of the blocks themselves.

[alacheesu]

Thanks for your hard work.

[Hondacity]

Very interesting, and a test that was long overdue for someone to try. Thanks!

All that remains now is figure out wth is going on with Naekuh and Honda's cases.

Vapor did the testing... I say his results are legit. My notes were very limited.

Vapor

Thanks for the testing, it pretty much confirms alot of things.

[NaeKuh]

for the record i stand corrected. :\

I am guessing inconsistant mount, or TIM or some type of application.

When i tested the mount i only did 2 mounts, not a full 5 like vapor did. I then asked honda to try 1 mount, not 5 and off that 1 mount he saw improvement.

We then assumed it was ligit. Thats the keyword ASSUME.. sorry honda i made an ASS out of U and ME.

But whats funny is a third confirmed the result too, but him also im guessing 1 mount.

[Waterlogged]

I think he was trying to show that even with three pumps the extra heat dump doesn't effect temps enough to be meaningful. I guess he could have hidden the actual block temp comparisons and show only the variances (but I have to admit to a certain fiendish delight at the results).

But you do have a good point, certainly the results should not necessarily be taken as demonstrating an absolute comparison of the blocks themselves.

Well the thing is, with 3 pumps, I doubt there has been a loop made that could pose restrictive enough to affect the temps at all. A test like this needs a single DDC 3.2 or D5 to show any kind of usable data. The test was valuable in proving that the 350 does indeed love lots of pressure.

[Vapor]

I guess I'm gonna be the only person here that thinks that testing this with 3 pumps is overkill and in a sense...fixing the results.

The reason why I went with three pumps was because NaeKuh's loop used dual DDCs and that's where one of the reported gains was. I really wanted to increase the odds that I'd see a gain and I did it by mimicking the test setup that anecdotally saw a gain.

It was also my gut feeling that if any gains were to be seen, they'd be at ludicrous pressure values. The rest of my testloop isn't particularly low restriction either: two MCRs, a KM17 flowmeter, a 90 degree bend, ~7' of 3/8" tubing; so the difference in inlet pressures should be pretty tangible. (the very, very informal "squeeze the tube going into the block" test agreed too )

Maybe after GOOC I'll test a block at very low pumping power, but I really don't think it's necessary--if there was any gain to be seen, I think it would have reared its head in this test at least a little bit.

[Hondacity]

No problemo dude, we do test our hardware and share info, others just blabber. Vapor did the community good.

for the record i stand corrected. :\

I am guessing inconsistant mount, or TIM or some type of application.

When i tested the mount i only did 2 mounts, not a full 5 like vapor did. I then asked honda to try 1 mount, not 5 and off that 1 mount he saw improvement.

We then assumed it was ligit. Thats the keyword ASSUME.. sorry honda i made an ASS out of U and ME.

But whats funny is a third confirmed the result too, but him also im guessing 1 mount.

[LOUISSSSS]

thanks for the confirmation! you're the man! this test will definitely help me in re-ordering my next loop.

can we get this stickied for a while or at least have someone add it to the sticky about WC tests and reviews?

[Waterlogged]

Vapor, Loop pressure is an interesting subject to me. I did a test a while ago with a loop as long and almost as restrictive as I could make it. I didn't really have the right equipment at the time for the test but it revealed something that has maintained my interest in solving my questions. Here's a pic I took of that test and setup with the results. The pressure gauge is an old 0-15 PSI fuel pressure gauge that didn't appear to be sensitive enough so I'm saving for some (2) that will cost ~$110 ea so I can redo the test. While it's completely different from what your doing, it did point out to me just how powerful 2 DDC's (let alone 3) really are.

[santiagodraco]

Well the thing is, with 3 pumps, I doubt there has been a loop made that could pose restrictive enough to affect the temps at all. A test like this needs a single DDC 3.2 or D5 to show any kind of usable data. The test was valuable in proving that the 350 does indeed love lots of pressure.

Nod. It would be interesting to see it with 2, then 1 pump. Does it continue to scale significantly or not? Me I'm just looking for a good old single pump at full power test. But there's been a few of those, and more coming I believe.

[LOUISSSSS]

Well Done Vapor.

I'm sure there will be "yah but" & "but what if" ....etc. However, thank you for taking the time & putting in the work to dispell the various theories & out right misconceptions.

but would the same results apply to a single ddc3.2? a.k.a. there would be no difference in loop order even with a single pump?

[Jah]

<3 thanks a million times!
Nice to get some data to back up what physics say should happen, you never really know until you actually tries it.