AquaComputer Cuplex Kryos XT
Wednesday, 30 June 2010 11:31
Thermal Testing Methodology/Specification
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 five different pump settings for each block, and use good testing practice by performing 6 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 6 mounts of each configuration for every waterblock. The best configuration will then go on to be tested through the full flowrate spectrum.
- The processor I’m using for this test is my D0 i7 920. I’m running it at 20×200 (4000MHz) at 1.50V loaded on a eVGA E758 with HyperThreading enabled. It is unlapped. I’m running 2GB of Kingston DDR3 1600MHz. I watercool the MOSFETs with a Koolance MOSFET block that is in the radiator subloop. The IOH and Southbridge sink is left stock. The video card is a passive ATi HD4350 running in the middle slot. The board is sitting on my desk alongside my PCP&C 610W PSU, Zippy 700W PSU, DVDRW drive, and HDD drive.
- 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 at 5V. I use a Koolance PMP-300 on the radiator and MOSFET subloop.
- For the block subloop, I use a Laing D5, two Laing DDC3.2s, and a Laing DDC3.25 for the pumps as well as a trio of Koolance FM17 flowmeters.
- I use a shared Koolance reservoir between the two subloops.
- I do a six mount test for each block configuration, each with their own TIM application and full cleaning between. I discard the data from the worst mount–I present it to the reader, but my final analysis and numbers are all based on the five best 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 worst mounts are somewhat anomalous; by performing six mounts and focusing on the better five mounts (in terms of thermal performance), I feel I am best representing the expected performance of a product. Additionally, I also present the best mount from each block in the comparison since I feel that is a good representation of what to expect in the best case scenario of each block.
- 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 2 hour and 35 minute program, the first minute is idle, then I have two and a half hours of load, and then 4 minutes of idle. The first 30 minutes of load data is considered warm-up and the last 120 are used for results.
- I use OCZ Freeze as my TIM. This is because it is the only true non-cure TIM I have found, to date. MX-2, as I used in my previous tests, does have a noticeable cure.
- I have found that each CPU stepping (and maybe even each individual CPU) has a unique temperature probe response curve, relative to water temperature. This i7 920 D0 has temperature sensors that scale linearly with water temperature, where the reported CPU temperature increased 1C for every ~.87C increase in water temperature. I account for this before presenting the data. Accounting for this is critical, as even a 2C variation in water temperature would cause an error of over a quarter of a degree.