Cooler Comparison Testing
Cooler Testing Methodology
To best gage the quality of the system coolers under review, system CPU temperature and cooling system audio measurements were taken with the CPU idle and under load. To replicate CPU idle conditions, the system was rebooted and allowed to sit idle for 10 minutes. To replicate a stress system load on the Z77-based system, a combination of LinX and FurMark were run over a 30 minute period with LinX running for 500 loops with Memory set to All and FurMark running at 1280×1024 resolution and 2x MSAA in stress test mode. For the Z87-based system testing, Aida64 System Stability Test was used in conjunction with FurMark for 30 minutes per run. After each run, the system was shut down and allowed to rest for 10 minutes to cool down. Then the CPU cooler was removed, cleaned, and remounted to the CPU with fresh thermal paste applied. This procedure was repeated a total of nine times for each cooler – three times for the stock speed runs on the Z77 and Z87-based systems, and 3 times for the overclocked speed runs on the Z77-based system.
Temperature measurements were taken directly from the CPU thermistors using RealTemp v3.70. For the Z77-based systems, the highest recorded value for idle and load temperature were used for the run. Because of the volatile nature of the Haswell thermistor readings, the Z87-based system temperatures were measured in a different manner. For idle temperatures, the highest recorded value was used for the run. For load temperatures, a series of three values were recorded: the average (high and low) across all cores, the average (high and low) across the single highest core, and the high temperature. To measure these average values accurately, the RealTemp readings were reset 20 minutes into the run while the CPU remained at full load. This allowed the software to measure accurate load high and lows for all cores over the last 10 minutes of the run with the three measured values taken from these readings.
Note that the temperature values are reported as deltas rather than absolute temperatures with the delta value reported calculated as CPU temperature – ambient temperature. For all tests, room ambient temperature was maintained between 26-28C. Sound measurements of the system cooler where taken with the sound meter placed 3 feet away from the system with all other devices in the room silenced. The Sound Meter Pro applet on a Samsung Galaxy S3 mobile phone was used to measure decibel level.
Intel Z77-based Ivy Bridge System Testing
CPU Stock Speed Testing
The CPU stock speed testing was conducted with the BIOS defaults set for the CPU and Turbo Mode disabled, equating to a 3.4GHz CPU speed, 1600MHz memory speed, and 100MHz base clock. The Intel Speedstep functionality remained enabled for the duration of the testing to get realistic CPU idle performance conditions.
With an Ivy Bridge CPU at stock speeds, the XSPC Raystorm EX240 kit performs very well in comparison to the Corsair liquid cooler and the Thermalright air cooler. The XSPC kit shaved a whole 5c off of the Corsair's temps, while lagging the air cooler by 1C. At the ambient temperatures used for testing (26-28C), this performance is very good and track well to performance expectations. While the Thermalright cooler performed best overall, the XSPC kit's performance is at an advantage because of its lower operating noise.
CPU Overclocked Speed Testing
The CPU overclocked speed testing was conducted with known stable settings from a previous board review for the CPU with Turbo Mode disabled, equating to a 4.4GHz CPU speed, 1960MHz memory speed, and 105MHz base clock. The Intel Speedstep functionality remained enabled for the duration of the testing to get realistic CPU idle performance conditions.
Board voltage settings were configured as follows:
- CPU Core Voltage – 1.2750
- CPU I/O Voltage – 1.150
- DRAM Voltage – 1.6255
- System Agent Voltage(SA) – 1.0850
- CPU PLL Voltage – 1.7500
- PCH 1.05 – 1.0995
With the Ivy Bridge CPU overclocked, the XSPC kit performs even better in comparison to the Corsair Hydro 100i cooler, coming in a full 7C cooler. The Thermalright air cooler still maintains is lead, but that performance comes with the price of fan noise. The XSPC Raystorm EX240 still remains a better performer with all variables considered.
Intel Z87-based Haswell System Testing
CPU Stock Speed Testing
The CPU stock speed testing was conducted with the BIOS defaults set for the CPU and Turbo Mode disabled, equating to a 3.5GHz CPU speed, 3.5GHz Ring bus speed, 1600MHz memory speed, and 100MHz base clock. The Intel Speedstep functionality remained enabled for the duration of the testing to get realistic CPU idle performance conditions.
Using the Haswell test system at stock speeds, we see a much different performance picture with the XSPC kit. While the XSPC kit beats the Corsair cooler by 4C against the average temperature across all cores, the gulf widens to 7C when looking at the maximum core average and absolute temperatures. The XSPC kit remains tight with the Thermalright cooler as well, keeping within a respectable 1-2C across the board even with the CPU idling.
Sound Testing
With the exception of the Corsair Hydro 100i liquid cooler, the cooler's fan speed remained constant throughout the run. The Corsair unit's fans are dynamically controlled by the unit's internal logic board. The XSPC kit out-performed all the other coolers with its operation barely noticeable over the case fans from the Cooler Master HAF-XB test bench. From these results, you can easily see that the Thermalright's performance comes at a high price with respect to fan noise.
I love these bundled kits
I love these bundled kits from XSPC, but I’ve had terrible experience with these specific pumps. I had one die after about a year, the replacement was DOA and the replacement replacement makes inordinate amounts of humming noise when it first starts, but then quiets down.
All of their other products have been wonderful though! Waterblock, radiators, fittings, etc. I wouldn’t ever buy another pump from them (although the bay design is very handy!)
Seems better to just get a
Seems better to just get a H100i if all you need to do is cool the CPU. cheaper and more convenient.
With good sealed systems, I will only use a custom setup if I need to watercool multiple items and want to do a setup using multiple radiators (eg cooling the CPU, VRM, chipset, and GPU)
I wish they could make a compound radiator, eg a single 120mm radiator divided into 2 sections (essentially 2 separate radiators that are in a housing that can fit in a 120mm mount) this way I can do a neat setup of a 240mm radiator for the CPU, and 1 half of a single 120mm radiator for the VRM and the other half for the chipset, and finally another 120-140mm radiator for the GPU, and a good pump
One problem you run into with
One problem you run into with compound radiators is cooling ability. In the case of serially-stacked radiators, you get only about 1.5x the performance compared to a single radiator because of pressure drop – serialization doesn't work as well with radiators as it does with water blocks. With parallel radiators, you either need double the fans or fans with higher static pressure so that air can effectively pass through both radiators. In both cases, you are using air that has been heated up passing through the first radiator to get to the second.
If you have two indendent chambers sitting side by side, you would effectively half your cooling performance in either loop because you have half the amount of space for the water to flow through as well as half the amount of surface area for heat transfer from water to air.
The reason I would like a
The reason I would like a split radiator, is to cool lower heat components e.g., the chipset and the VRM’s.
Mainly trying to make use of 3 top 120-140mm exhaust fans, as well as a rear 120-140mm exhaust fan, with no stacking of radiators, to cool the CPU, and GPU.
Will a powerful pump be able to handle a setup like that?
Should be without issue. In
Should be without issue. In my main system, I have a Swiftech MCP-35X DDC pump running the following with good flow maintained:
XSPC RX360 radiator
Magicool 140 radiator
Koolance CPU-360 CPU water block
XSPC Razor GTX 680 full cover water block
Well, i can not recomment bay
Well, i can not recomment bay res’ at all. It’s alyways a pain in the … to fill them. and my xspc res is just bended from the pressure of my D5. If you want a save costum watercooled system stay with aqua computer, EK and bitspower fittings. I just had way too much trouble with ofer brands.
Well they could. entrga do
Well they could. entrga do here in Brazil … certainly buy one
I’m looking at graph on
I’m looking at graph on Koolance.com stating that at a temperature delta of 15°C their 2x120mm 18-FPI aluminum radiator dissipates anywhere between 400 and 1000 Watts ( depending on the fans and coolant flow rate).
The delta they’re referring to is the one between temperatures of the coolant entering the radiator and ambient air not the CPU and ambient air.
Considering that their radiator should perform roughly similarly to the one being reviewed and considering how high the deltas were for the reviewed processors ( neither of which consume anywhere near 400 Watts ) does that mean that the temperature of the coolant in no point in the loop rises to 5-10°C above that of ambient air ?
The radiator in question : http://koolance.com/radiator-2-fan-120mm-18-fpi-aluminum
The graph is under the ‘specification’ tab.
The koolance graphs measure
The koolance graphs measure coolant temperature only over time which can be used for theoretical radiator performance. It tells you how well the radiator will cool the coolant flowing through the loop, which equates to how effective the coolant can absorb heat (the cooler the coolant, the more heat it can potentially absorb from the waterblocks). However, this does not tell you how the entire system will perform.
The graphs in the review attempt to show you what type of performance you will get with the system against CPUs in various situations – a stock Ivy Bridge, an overclocked Ivy Bridge, and a stock Haswell. The delta is a measure of the actual temp – ambient temp, to give you an idea of what the temperatures would based on your rooms ambient temperature (you simply take the reported numbers in the graphs and add it to you ambient). Its more of an absolute working system temperature for the kit than the numbers that koolance reports.
Make sense?
I don’t see any strengths in
I don’t see any strengths in this xspc wc setup compared to the thermalrights silverarrow except for sound levels.
but does lower snd lvls for xspc equal a plus if the cons are possible:
1. leakage & dmg to components.
2. cost of pump going bad vs fans.
3. $150 xspc vs. $90 thermalright arrow
4. arrow few deg cooler
i’d only give it gold award if it was fools gold
you can use non conductive
you can use non conductive liquid still shit
Any chance of you guys doing
Any chance of you guys doing a review with the Arrow coupled with a fan you consider to be noiseless (or at least near the noise of the xspc) so we can see how good it compares to the others with less noise? It seems like you guys should throw in a reasonable air solution (fan I mean) to show how much difference there is between it and the one that drives you out of the room 🙂 Maybe I missed it if you’ve already done that. But it would be nice to see the great heatsink with a less noisy version in every review. One fan tested for the crazy people or people with headphones and the other less noisy version for the rest of us.
If you only lose a few degrees doing this water seems pointless for most (I own a koolance, so it’s not that I hate water, just curious how noise free fans work with the best heatsinks). I always see reviews with the worst fan, which I’d never buy 🙂 I’d buy the monster heatsink to avoid the need for a noisy fan but get decent results (maybe that’s just me).