| Antec Kühler H2O 920 Liquid Cooler | |
| Reviews - Featured Reviews: Cooling | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Written by David Ramsey | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Tuesday, 12 April 2011 | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Page 6 of 8
Testing and ResultsFor this test, I used the following heat sinks in addition to the Thermaltake Jing:
For heat sinks without a stock fan, I used a Thermalright TR-FDB-12-1600 fan, which puts out 63.7CFM at 28dBa according to Thermalright. This mid-range fan provides good air flow and reasonable noise levels. For "apples to apples" testing, where each heat sink is tested with the same fan, I used a Delta AFC1212D. This high-performance PWM fan is rated at 113CFM at a claimed 46.5dBa at full speed...which means that while it moves quite a bit of air, it's very loud. The Intel Core i7-950 I used in this test runs much hotter than the Core i7-920 I'd used previously. At 1.35 volts, with a BCLK of 175Mhz, the 4,025Mhz CPU pumped out enough heat to stress the very best heat sinks. AIDA64 would report throttling once any single core reached 100 degrees Celsius; any throttling resulted in cancelling the test and recording a "FAIL". Although this overclocked and overvolted Core i7-950 represents an extreme, these are expensive, high-end heat sinks. The chart below summarizes the results with the stock fans (hotter temperatures towards the top of the chart, and cooler temperatures towards the bottom). The twin-fan coolers (the Cooler Master V6 GT, Corsair H70, and the Kühler 920) have a real advantage here, since their dual fans move more air than the stock single fan of any of the other units. The Corsair H70's fans at their default 2,000RPM level move a lot of air together, but also generate a fair amount of noise. At the 1,600RPM level achieved with the in-line resistor cables, the noise level is much reduced, with a relatively minor performance hit. Stock Fan Tests
And we have a new champion! With its twin fans howling away at full speed, the Kühler 920 jumps to the lead, beating the Cooler Master V6 GT by less than a degree. Following closely behind is the Thermalright Silver Arrow and Venomous X, and the Corsair H70 with its fans on high. There's a 5.3-degree "break" between the Coolit Vantage A.L.C. and the Corsair H70, which neatly separates the coolers into "lower performance" and "higher performance" groups. Delta High Speed Fan Tests
With the Delta high-speed fan, our lineup changes. Showing what a difference a change of fan can make, the Coolit ECO A.L.C. moves from the bottom to the chart to mid-pack, improving by almost 11 degrees. The Antec Kühler 620 improves by 7.2 degrees, putting it— amazingly— between the mighty Prolimatech Super Mega and the Thermalright Venomous X— and in fact with this fan, it beats its big brother. The Kühler 920, though, drops far back in the pack. Although the 113CFM of the Delta high-speed fan is slightly more rated airflow than the 110CFM Antec rates the Kühler 920's stock fans at, temperatures actually rise slightly. Double-thick radiators simply don't do well with single-fan setups; you'll notice a similar performance drop with the Corsair H70 (as compared with its dual fans on high).
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||





Comments
A good case point is my current cooler, which is the old Cogage True Spirit 39 dollar wonder. I did put a Gentle Typhoon 1800rpm fan on it and also installed a backplate, bringing the grand total to about 65USD. I would have bought a Megalhelems probably, but back then I thought I was going to use the push pin mount system.
The thing is, with my current rig running the Gigabyte UD5 ver. 2 and my Bloomfield 920, I get a max temp running Prime 95 in heat mode of 75C at a 3.8Ghz OC. This is with a 21C ambient. And, more to the point, 75C is well within tolerance for a Bloomfield 920. You could run it 24-7 like that and never have wear from heat.
But more applicable is when I run Crysis 2 at maximum settings for two hours and never break 60C.
So the conclusion is why buy high when you can cool low for half the price? I love the V6, but just can;t justify replacing my old Cogage since it cools so well anyway.
Because of the ridiculous amounts of heat generated by modern chips, the coolers required have become seriously massive, most of them to the point they either need substantial backplates to prevent motherboards from warping, hanging over RAM slots which limits RAM use or simply just being too big to fit into a case.
Nobody ever said these all-in water units were as good as a proper custom loop, but they're not supposed to be. They're intended to compete with top-end air coolers without having the ridiculous mass to contend with, and in that regard I think they've been very successful.
I never used one on the Bloomfield builds I did because either I was pushing for some serious clocks (and therefore needed cooling performance beyond what these units can do), or was on a tight budget, so I fully appreciate the price point for these units always leaves a big "do I or not" question.
I do agree with you that the Cogage True Spirit is a little gem!
The first is that in the case like the 620 the cooling is one with one less fan, there by less noise. What I mean by this is the fan used by these types of coolers replaces the existing exhaust fan on the case. The typcial cooler system introduced another fan to the system and often two to achieve the same level of cooling, thus more noise.
The second advanatge is the reduction of overall internal PC temperature. Normal air cooling solutions take the CPU heat and then blow it back into the case. These liquid cooling solutions take the heat and move it outside the case, reducing overall internal temperature. It might only be 2c or 3 c but every little bit helps.
While there are some great standard air coolers I have been sold on the self contained liquid coolers from day one of the H50 and Coolit ECO. Sure you can cool for less money but then again you can drive a Festiva for less than any other car model.
Have you considered including the coolermaster 212 plus with your benchmarks? The 2600k has a mockery of a stock cooler so many people consider the $30 212 the minimum solution. It would be interesting to see how much more cooling you get for spending more money.
this is just my 2 (euro) cents... ;)
I have actually replaced the fans on my 920 with the same Delta fans as in this review. While the fans run the software does not display any data for them. Since it is not communicating with the fans, it can not follow the settings I have for the rpm changes. Frankly, the only way I know they are even running is if I have it on the Extreme setting and can hear the air movement. I have tried uninstalling and re-installing all software related to the 920 and it is the same. Any ideas concerning the issue? Anything would be appreciated.
Travis
Larger diameter hose will lessen the load on the pump, resulting in higher flow.
Every millimetre of hose length does impose *some* flow restriction, no matter what diameter, so increasing the diameter *will* to some degree reduce the restriction and thus increase the flow.
"Plus, you'll actually get slower movement with larger hose..."
So? The flow rate (volume/time) is what matters, not the flow speed through the hose.
On the contrary I'd argue that since larger diameter, same length, hose will hold a larger volume that factor alone is a bonus in terms of increased thermal inertia.
It's interesting, because I can't imagine replacing a (pair of) 2400rpm fan(s) with something else and expecting increased airflow while getting much less noise, and yet it seems like Antec would have had to unintentionally design their 620 better than they expected for it to perform better than the flagship with only a simple fan swap. That could be the exact situation here though, is what the testing shows.
Olin is correct: the larger diameter hose will permit more fluid flow, but that is entirely up to the pump. I guess the secrecy act needs a full frontal assault by a U.N. committee to let us know what we're dealing with here.
A little hint on mounting of the radiator. I ran a thin metal pin (a straigtened paper clip actually) through the case mounting hole, fan mounting hole and into the radiator screw hole to align, making it much easier to line up and secure a screw through the case, fan and to the radiator. Saves a lot of headaches even when almost blindly mounting anything. (There are tapered pinning tools used in construction designed specifically for this same scenario, just bigger joins.)
Determining the inner hose diameter and pump flow rate isn't hard, but it's destructive to the coolers...
1. Find out at what temperature the CPU throttles. One way is to run the cooler at ?Silent? and use a third prty program to read the CPU temperature when the CPU throttles (Tt).
2. Set the cooler to ?Extreme?, and calculate the difference in temperature between the liquid and the CPU (deltaT) when the CPU is working att full load without throttling.
3. Close the program keeping the CPU busy. Enter the cooler Settings panel. For the Full fan speed temp I chose
Tt ? deltaT- 5 degrees
to keep some margin.
For the Ramp start temp I chose 25 degrees, to keep the fan working in the lower register longer without having to actually go to full speed.