ASUS Silent Knight, Triton 77, & Silent Square EVO

FrgMstr

Just Plain Mean
Staff member
Joined
May 18, 1997
Messages
55,532
ASUS Silent Knight, Triton 77, & Silent Square EVO - CPU heatsinks from ASUS? Sure, why not? They sent us three new heatsink and fan units to test and we give them a run for the money on a new Intel Quad Core processor specially fitted with a hardware temperature monitoring system.


The numbers speak for themselves. ASUS’ three new offerings into the enthusiast cooling realm are not going to win any awards based on performance. That said, the Silent Knight II was the standout among the three coolers in terms of looks and performance.
 
Great review. The new testing equipment and procedure sure takes the guess work out of it. I won't be buying one of those anytime soon but allittle more competition in any market can't hurt.
 
I enjoyed the review as well. Sometimes I wish Alpha would make a come back.
 
Another good review! I just got my Tuniq Tower last week, and it does a great job on my q6600. After reading this review I stand by my choice.
 
Love the way you guys are using [H]ardware testing :p.
Not liking the price point on the ASUS heatsinks however, The same performance can be obtained at a much lower cost.
 
Actually its more of a question. Is there any performance difference between a horizontal and vertical position of the heat pipe coolers? It is something I have wondered about for awhile. Seems that on most heat tube setups any reconstituted coolant wouldn't make it back to the heat sink if in a vertically positioned motherboard.:confused:
Any help welcome.
 
please start the graphs at 0 otherwise they are very misleading. it makes the performance delta seem much larger than it actually is.
 
Thanks for another great review Kyle.

One question I had though was why didn't you guys include the Thermalright 120 or Tuniq 120 in this review as they were the big boys last time.

Also, you guys have two cooling / heatsink related categories on the homepage. "Cooling" and "Heatsink and Fan Unit" which is kind of confusing and makes it harder to find older reviews.
 
please start the graphs at 0 otherwise they are very misleading. it makes the performance delta seem much larger than it actually is.

0 what, C or K? The graphs actually make a good representation of the performance delta, as 0 delta gives close to 0 bar length. However, the type of graphs you are asking for would give a better representation of the performance ratio.
 
please start the graphs at 0 otherwise they are very misleading. it makes the performance delta seem much larger than it actually is.

The graphs are started at the ambient temperature the units are tested in. I see little reason to start the graphs at freezing.
 
Hey, great review.

Just one thing that came to my mind: It may be better to use a thermal probe over the infrared thermometer for ambient temperature.

The reason is that the infrared measures radiated heat and not air temperature. It will pick up the infrared radation that comes of what ever you point it at. It's great to measure surface temperature of objects. Pointing it "at the air" isn't really working as air just isn't dense enough to emit meaningful amounts of IR. You could point it at a wall but given that every object has a certain thermal mass changes will be rather slow.

I think the temperature of the air getting to the heatsink is what is of interest. So putting a thermal probe towards the "air inlet" would be working better. Just make sure that the probe is not "thermaly coupled" towards any component. Even touching the wires anywhere near the sensor with your fingers can make the a meaningful diffrence as the wire will work like a "heatpipe".

So it's somewhat like "sunrays" vs. "hairdryer"

Yeah, that's nitpicking but I thought it might be something to keep in mind for the next time.

Have to say that I really like your new test setup with the drilled out thermal probe. Machining that heatspreader - That's the [H] way!
 
Hey, great review.

Just one thing that came to my mind: It may be better to use a thermal probe over the infrared thermometer for ambient temperature.

With the Sperry 4-way there it is easy to keep a thermister reading too.If you ever got to get one, the DT-506 can be picked up farily cheap. Marc, myself, and Paul use them. I have purchased all of ours from here. And yes, you are right about the infrared...
 
Hey, great review.

Just one thing that came to my mind: It may be better to use a thermal probe over the infrared thermometer for ambient temperature.

I am glad you brought this up. When testing the ambient temperature I use both to confirm.

I use the infrared thermometer to test and then cross reference one of the extra thermal probes to confirm the temperature. I don't believe I mentioned this in the article but rest assured I make sure the ambient temps are accurate or else none of my scores would be worth a damn.
 
Great review. I The testing methodology is very well thought out and germane to circumvent the hyperbole that often follows this type of product. Most of my research of late has been dedicated to finding a great air cooler with a low noise signature. I look forward to more reviews in the future.
 
Great review. I just picked up the Silent Knight II for my new rig. It received positive reviews at other sites, but a [H] review is worth its weight in copper heat pipes.
 
It would be great if you could have compared the Asus heat sinks to other popular/well performing heat sinks (TRUE, Xigmatek, Tuniq Tower). I don' think I've seen anyone here recommend a Big Typhoon.
 
Nice job Marc, when I saw the intro part about the different temp readings via different methods I must say I wasn't that surprised. Software readings, bleh.
 
I hope the one thing every takes away from this article is that the [H] respects your opinions and we are willing to do whatever it takes to make sure we get the job done right.

I mean drilling a channel in a $1,000 CPU. :eek:

Only at the [H] :D
 
So when do we see the guide on how to make our own hardware monitoring systems and how to drill into the cpu :)

would be curious to know how you avoid any small pieces from going into the casing causeing a potential short.
 
I'm confused, why do people want the graphs to start at zero? :confused:

I don't see what benefit this would provide, please explain.
 
Where was this testing done? 38dba ambient noise is ridiculous, how do you guys work :p.

/ADD

Very nice review, I liked the performance/price comparison at the end as well.
 
I'm confused, why do people want the graphs to start at zero? :confused:

I don't see what benefit this would provide, please explain.

Well it would just provide less visual difference between the coolers as 50-55 would look much smaller if the graph was 40-60 (20 pixels, 20-5 = 25% visual difference) than if it was 0-60 (60 pixels, 60-5 = ~9% visual difference)..

Eg:
Code:
0              60
............x
.............x

vs

40            60
.......x
............x

The difference in the second graph looks bigger than the first even though they are the same.

As long as you look at the numbers on the graph and not just the bars it doesn't matter either way.
 
Where was this testing done? 38dba ambient noise is ridiculous, how do you guys work :p.

Be glad you aren't in my office. SM-8800 is 50dBA+, 2x IBM x330 1u servers, Intel 510T switch :eek: .....plus all the PC's.
 
Maybe they could add screened text that is obviously an alternate scale?


view.php
start the graphs at zero guys comon.
 
Interesting about the temp differences, for a good while now people have been using coretemp as the be all end all and now it looks like it could be around 10c out.

Does anyone have alink to the original hardocp article they had about bringing in this type of testing, been looking and can't find it. :mad:
 
Awesome review :)

+1 for starting the graphs at ambient. That makes it very easy for me to visualize the relative thermal resistance of these coolers, since they cool using ambient air.

+1 for including the MOSFET temps. That's a big concern when comparing any cooler that changes the air flow vs. the Intel stock cooler. You guys nailed it in true [H] style!

A few requests:
- on heatpipe coolers, pls tell us the heat pipe diameter
- pls indicate the wattage of the OC CPU - that will help those of us who are considering a different CPU make an assessment if the results are applicable to our intended CPU overclock choice, since heatpipes work great up to a saturation point

An outlandish idea:
- A synthetic load that exceeds common CPUs (say, 150W) would really help separate the wheat from the chaff wrt heatpipe capacity. Before I get flamed, I already know this is unpopular, less relevant than the work already performed, and yet another permutation. In spite of this, I feel the additional data point would be helpful to get an impression of the headroom a heatsink provides for overclocking.

Actually its more of a question. Is there any performance difference between a horizontal and vertical position of the heat pipe coolers? It is something I have wondered about for awhile. Seems that on most heat tube setups any reconstituted coolant wouldn't make it back to the heat sink if in a vertically positioned motherboard.:confused:Any help welcome.
The inside surface of a heatpipe is coated with a wicking material, usually some kind of stainless or copper wool. The wick pulls the condensed liquid back to the source of the heat (where the wick is dry), via capillary action. The effect is strong enough to make gravity (and therefore mounting orientation) negligible. Also, the wick increases the surface area available for evaporation at the hot part of the pipe. That's how heatpipes work in orbit ...and against gravity.
 
I'm confused. Are you saying that is really loud or really quiet? :confused:

Loud. My workshop ranges between 31dba and 34dba. Lower when my computers are off... but not by much. I took my ratings at 4am, as that's when I do all my work anyways. Naturally, it's quieter than during the day by a good 10dba. I also took all the ratings of my systems/components in a soundproof DJ booth (back when I was doing that) with an ambient noise level below the range of the SPL meter (so, under 24dba).

Be glad you aren't in my office. SM-8800 is 50dBA+, 2x IBM x330 1u servers, Intel 510T switch :eek: .....plus all the PC's.

Yeah, that would drive me insane. If I'm doing work I can deal with it, but for just general computing, I'd go nuts.
 
good stuff from asus, even though pricey those are still some pretty sharp looking units with not too shabby temps. the knight looks similar to the zalman style fins with nice noise/temp ratio, too bad you didn't throw a 95/9700 in there.

I don't see how scaling the bars from zero would make any more sense than ambient either, the results are always going to be a few degrees apart besides the stock so there's no point in making the graph with less contrast. we want exactly the opposite, highlighting the deltas as much as possible to compare, so a zero scale would kind of defeat the purpose wouldn't it.
 
Zero degrees is an arbitrary number anyway. Freezing point of water? That just happens to be convenient. Absolute zero would be completely rediculous. Using ambient temperature as the baseline makes the most sense because it's actually relevant to the current test. Only someone trying to make a poor performing heatsink look not as bad would use zero as the baseline.
 
Zero degrees is an arbitrary number anyway. Freezing point of water? That just happens to be convenient. Absolute zero would be completely rediculous. Using ambient temperature as the baseline makes the most sense because it's actually relevant to the current test. Only someone trying to make a poor performing heatsink look not as bad would use zero as the baseline.


QFT

We will not be using 0c as a baseline for our graphs as there is simply no logic in it at all. Ambient makes good sense to us and that is what we will continue to use.
 
As a side note, I observed an interesting phenomenon where only the vertical lines of the paper are reflected in the base. Rotating the unit 90 degrees causes only the horizontal lines to be reflected. It has absolutely no bearing on our testing but interesting none the less.
Hahaha...that made me chuckle. Come on Kyle, you know about anisotropic filtering, you should have known this too! Thin lines get thinner and blurry at an angle...until they eventually fade. :p

Anyway I liked the review, I just think it's a shame that at 60-70 bucks, these coolers just don't do so well. The Ultra-120 + a fan usually comes up to that much, or less in some cases.
 
Hahaha...that made me chuckle. Come on Kyle, you know about anisotropic filtering, you should have known this too! Thin lines get thinner and blurry at an angle...until they eventually fade. :p

Anyway I liked the review, I just think it's a shame that at 60-70 bucks, these coolers just don't do so well. The Ultra-120 + a fan usually comes up to that much, or less in some cases.

Actually this has to do with the direction the abrasive surface was used. If you go back and look at the previous round up, you will not see this issue. But you are also correct in that it has a lot to do with the angle of the surface as well. Quite frankly, I did not want to get into the discussion in middle of the article. :)
 
An high end heatsink review without including the accepted "best" Thermaltake TRUE 120? :(
 
Would lapping these heat sinks provide a worth while decrease in temperatures? I'd hate to ruin a $50 cooler.
 
Back
Top