Core i7 Roundup #2
Koolance CPU-345 and CPU-350 Installment
November 5, 2009



Intro

The Koolance CPU-350 is Koolance's current flagship block. It debuted roughly a year ago and has, in the minds of many, made Koolance a lot more relevant. The overall design has 4 main pieces: the mounting bracket, the injection top, a midplate, and the base. How it all comes together is a little different than other blocks (but apparently a Koolance norm)--the top and the base screw together. There are no assembly screws anywhere, it's pretty cool. As the flagship block, the base uses the finest machining they could muster--it's a nickel plated micropin design with an extremely thin and bowed base. It's gained a reputation as a very restrictive block and will (mostly) live up to that reputation in my tests. In the brief lifespan of the CPU-350, there have been two midplates included with it, and they're tough to differentiate, even in pictures. There are differences in performance that will be highlighted in these tests. The CPU-350 comes in both acrylic-topped and acetyl-topped varieties.

The Koolance CPU-345 is the lower cost sibling of the CPU-350. It's a bit of an unknown block--I haven't seen any tests of it or even anyone using it in their build on the forums I visit, but it's definitely a block that can hold its own. It's larger than the CPU-350 and uses an array of medium-sized pyramid-pins and doesn't use a midplate, just using the top to inject flow at the center. Originally, it was released with a mounting bracket that used rotating posts--one quintet of mounting holes and one set of mounting hardware for every socket--it was really cool in theory, but was a pain to use. Since then, they've updated the mounting system to be equivalent to that of the CPU-350, which is a big plus. The block only comes in the acetyl-topped variety, although the acrylic-topped variety was originally offered.

This test will focus on the performance of the blocks in general and over a large flowrate spectrum. Results from the installments of Roundup #2 will be compiled, as they're posted, into an Overall Comparison page.


Thermal Testing Methodology/Specification


Methodology

My waterblock testing methodology has evolved over the past few months and I think it's finally at a resting point where I can start piling up test results rather than tweak the methodology (and thus preventing cross-comparisons). I use Dallas One Wire DS18B20 temperature probes at various points through my watercooling loop and at the air intake to measure temperatures, I've isolated the radiators so that the flowrate through them never changes, I use six different pump settings for each block, and use good testing practice by performing 5 mounts. Where applicable, I will also test various modifications to the blocks. These include testing various orientations and removing/adding various midplates, nozzles, dividers, etc. In some cases I will also modify the mounting system and present results from increased mounting pressure. For my waterblock tests, I'll perform 5 mounts of each configuration for every waterblock. The best configuration will then go on to be tested through the full flowrate spectrum.

Specification


  • The processor I'm using for this test is my C0/C1 i7 920. I'm running it at 21x200 (4200MHz) at 1.52V loaded on a Gigabyte EX58-UD5. It is unlapped. I'm running 3GB of G.Skill DDR3 2000MHz. All heatsinks on the board are stock and I have fans blowing over the MOSFET area for added 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 a two MCR320s with three pairs of Yate Loon D12SH-12 fans in push/pull on each radiator. I use a D-Tek DB-1 pump on the radiator subloop.
    • For the block subloop, I use a Laing D5 and three Laing DDC3.2s for the pumps as well as Dwyer RMC-142 and RMC-144 flowmeters to monitor and track flowrates.
    • I use a shared Primochill 8-port reservoir between the two subloops.

  • I do a five mount test for each block configuration, each with their own TIM application and full cleaning between. I'm fond of semi-discarding the best and worst mount data--I present it to the reader, but my final analysis and numbers are all based on the median three mounts. As a reviewer, I feel it is my duty to present the reader with performance numbers of a product that represent what its typical performance is. Often times the best and worst mounts are somewhat anomalous; by performing five mounts and focusing on the middle three mounts (in terms of thermal performance), I feel I am best representing the expected performance of a product.

  • I have 28 temperature probes in use: 24 Dallas DS18B20 Digital one-wire sensors and 4 Intel DTS sensors in the processor.

  • For temperature logging, I use OCCT v3.1.0'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 air intake and various water temperatures 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 pump RPM.

  • For processor loading, I find OCCT v3.1.0 to be extremely competent. With the Small Data Set setting, 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 simultaneously, which helps. I immediately also start to log the Crystalfontz data via WinTest b1.9. I run a 1 hour and 40 minute program, the first minute is idle, then I have 95 minutes of load, and then 4 minutes of idle. The first 20 minutes of load data is considered warm-up and the last 75 are used for results.

  • I have found that simply using processor temperature minus ambient temperature is not adequate for Intel's 65nm Core 2 processors. However, I have found that ambient and core temps scale perfectly fine (1:1) with i7.

Thermal Test Results


Now finally some results! First up, the individual configurations testing.



Wow! Talk about a lack of differentation between orientations (three were tested: with the barbs horizontally, vertically, and at a 45 degree angle). Also interesting was that the new midplate ("New MP") performed identically to the old one. In my previous tests, I noted a .25-.5C difference between the two, but I did something different this time: I marked the outside of the block with Sharpie so that I would reassemble the block identically--last time I assembled by feel. The flowrate of the old midplate configuration was 1.24GPM, easily the most restrictive stock configuration I've tested so far. The flowrate with the new midplate went up to 1.40GPM while noticing no degradation in thermal performance. Still shocking just how little this block cares about orientation--something rare from a high performance block these days.

The CPU-345 is quietly sitting back behind the CPU-350, but the difference was surprisingly small, only a ~1.7C drop in performance. The CPU-345 is very low restriction (about the same as the Heatkiller 3.0 LC) and has the same solid mounting mechanism that the CPU-350 uses. This block is a sleeper.

Specific Pumping Power

Now that we've figured out what the best configuration is for each block, let's chart its performance over the entire flowrate spectrum.
  • Very High Pumping Power: All three MCP355 pumps and the D5 are on at full speed--this has a very similar PQ curve to a pair of RD-30s at 20V.
  • High Pumping Power: Two MCP355s with EK V2 tops are on at full speed. The other two pumps are off.
  • Medium High Pumping Power: A single MCP355 with XSPC V3 top is on at full speed. The other three pumps are off.
  • Medium Pumping Power: The stock D5 is on at full speed and setting 5. The other three pumps are off.
  • Low Pumping Power: A single MCP355 with XSPC V3 top is on at minimum speed (~7.7V, ~2450RPM). The other three pumps are off.
  • Very Low Pumping Power: The stock D5 is on at minimum speed--setting 1. The other three pumps are off.

Note: I do 5 mounts at "Medium High" then take the best config of a block and test the whole flow spectrum (after a TIM curing session) then realign that curve with average of the 3 median mounts to give you the "Adjusted" data.

Other Graphs

More graphs for your enjoyment...let's start with reusing the flow vs. temperature data, but including pump heatdump (i.e., CPU vs. air temps). I have two iterations of it: CPU temperatures vs. my air temperatures and a setup with my water-to-air delta included twice more. The latter is to mimic a setup with one third the radiator power of my setup (roughly a 120x3 radiator with 1600RPM fans).


Note: these results are derived from adding the water-to-air delta three times to my water temps. I add them three times to emulate the radiator power of a loop with 1/3rd the radiator power mine has. I use 2xMCR320s with push-pull 2200RPM Yate Loons and the data emulates the conditions of a loop with a single 120x3 radiator with ~1600RPM fans.

Here we can see something that, at face value, is somewhat of an odd tendency: the more restrictive block is actually less responsive to increases in pumping power. I don't have a conclusive explanation why this is, but considering the vast differences in the design of the base, it doesn't really faze me. Both blocks scale pretty well up until the dual DDCs, with the CPU-350 being more resilient to low flowrates.



Conclusion

This conclusion is, as usual, pretty cut and dry. The CPU-350 is a great performer, coming within a quarter of a degree of the vaunted HK3.0LT in my tests. Yeah, it's more restrictive, but the updated midplate alleviates some of the restriction. The mounting system is significantly better than the XSPC/EK/HK/DangerDen variety, but still a step behind the Swiftech/D-Tek offerings. One big improvement to the mounting system would be if the included threaded rods would stop at the right pressure (much like Swiftech's setup). The CPU-350, like other radial flow blocks (Fuzion V2, Sapphire Rev.A, Apogee XT), has a pretty good resilience to reduced flow as well. One big plus about this block is that any orientation will get you the best results--if for any reason you need to rotate it, you'll see no performance hit.

The CPU-345 is a really intriguing block. It's extremely low restriction, well priced, well made, and has surprisingly good performance. It beats the Apogee GTZ, the Supremes, and the Enzotechs. Surely more blocks will fall behind it as my testing continues. It has the same (and really good) mounting system from the CPU-350 now, and is a pretty understated in terms of looks. I only did a few test mounts to test its orientation preferences, but it seems to share that quality from its older sibling--it just doesn't care which way it's oriented.

Both of these blocks are great. The CPU-350 has flagship performance, being imperceptibly close to the HK3.0LT in thermals while the CPU-345 is a really competitive (and unknown) mid-tier block.

 
 
 


[  Return to Main  ]