THG CPU HSF round-up

[AmberClad]

Tom's Hardware Guide CPU HSF round-up

I found it a bit harsh and sensationalist. They "failed" some good performing coolers because of difficult mounting mechanisms...

The Silverstone I'm using got failed because of the fastening nuts, lol.

Mounting the cooler is impossible while the motherboard is installed in the case, since the cooler's backplate is attached directly to the cooler using Phillips head screws, without a spacer. Thus, the board also needs to be removed every time you upgrade your CPU in order to fasten the screws on the back of the motherboard, since there is no other way to install the cooler.

I'm not sure they actually read the instructions for the backplate installation...there are in fact spacers, which affix the backplate to the motherboard.
In my case, mounting the cooler is impossible while the motherboard is not installed in the case.

The Zalman 8700 and the Zerotherm Nirvana got failed too.

 

[DerwenArtos12]

Dee-de-DEE, retarded! I can't read at toms anymore, they fail at life. Last time I read something there I came back to AT and actually started a thread titled "I am now stupider for having [read] that, I award you no points and may god have mercy on your soul"

 

[evident]

toms is so lame now... try and read their super mario galaxies review to see how low they've got

 

[Zap]

Originally posted by: AmberClad
They "failed" some good performing coolers because of difficult mounting mechanisms...

I've got that page open... working through it... gonna ignore their "ratings" but it is always nice to have more pictures/descriptions of coolers. Also, while they're being silly for having difficulty installing some coolers, there's a lesson to be learned from their failings, and that is sooner or later someone else might have a similar issue and post a "help I can't install this HSF" thread around here.

 

[dampeal]

how do you base a coolers score on how hard it is or isn't to install overall...

sure that might figure into it in a small way but you don't fail it because it is hard to install.. but yet has good temps...

If the cooler works better than others then enthusiasts really couldn't care how hard it is to install, most coolers are a PITA to install anyway...

 

[bigsnyder]

I was getting ready to post a thread on the same subject! From what little I read, some coolers that
have received good cooling marks here were not rated high, like the Scythe Ninja Rev.B. I have the
mini-Ninja in my Antec Fusion and it rocks.

 

[AmberClad]

I kind of see another potential problem with this round-up. There's a huge variance in the fans used. Some of the coolers, like the Thermalrights, don't come with fans, so they arbitrarily selected one. Some of the fans used were running 760 RPM, others were running 1200 RPMs, some were PWM.

One other oddity I saw was that they used the Silverstone FN122 to test on the NT06. The actual fan that comes with the non-Lite version is the Silverstone FM122, which goes up to 2400RPM and moves 100+ CFM. Not that I'd ever use that jet turbine...it's shockingly loud unlike what they state in their review.

No mention of the odd tilt to the panel of fins that I and some other reviews have noticed (the thing isn't perfectly horizontal and slouches downward a bit at the end furthest away from the pipes).

I think they were trying to do too much at once with this round-up, and the quality of each of the individual mini-reviews suffered as a result. Shame they couldn't spend more time with each one to give them a proper look.

 

[Cookie Monster]

Yeah, it feels like they rushed this. And the positioning of the coolers was weird seeing as most of them were positioned so that the fans blew toward the bottom of the case.

Plus some of the coolers are intended for mid range cooling performance. Who in their right mind would use the Zalman 8700 to cool their QX9650? Its like using the vf900 to cool a 8800GTX.

But then again, its THG.

 

[aussiestilgar]

Horrible.

 

[Deathawk]

Really bad article.

I'm glad I'm not the only one who felt that way,

 

[bigsnyder]

I really don't know why I still read hardware reviews there. Guess old habits die hard.

 

[CurseTheSky]

"We tested all of the coolers in this roundup on an Intel quad-core Core 2 Extreme QX6850 CPU. When all four cores are under full load, this model has a power consumption of 113.87 watts. In order to achieve this load, we use a multi-core version of Prime95 (ver.25.5a). The CPU displayed this power consumption when combined with the best water cooler we tested, reaching 62°C." (2)

62C is the lowest they got on any of the tested cooling solutions, running a QX6850 at full load (Prime95). Something strikes me as majorly wrong; either they're pushing FAR too much voltage, or they can't get ANY of their coolers to sit flat. Maybe they had an extremely concave IHS. Maybe it was 95F in the room where the tests were conducted? In my opinion, they should have tested these on at least two processors each - maybe a dual core and a quad core. Speaking of which, did they even MENTION the ambient temperature (which almost definately fluctuated somewhat between tests), the thermal paste used, or how much time was dedicated to mounting?

Edit: I suppose I should mention that I don't have experience with the QX6850, so there is a possibility I'm criticizing them for nothing. I can say, though, that my friend's ~3.6GHz Q6600 doesn't go anywhere near those temperatures on Prime95 load and a Tuniq. 60C? Maybe on a super hot day. 70C? Not a chance.

 

[The Boston Dangler]

not only can i do better, i will do better with less.

 

[gorobei]

the testing methodology is a little sketchy, but not that bad.

the fan speed differences were separated into low and high speed comparisons at the end graphs.

the mounting orientations were a little suspect, not sure why you would point the fan down towards the video card. and why didn't they use multiple orientations for all the coolers: n/s e/w?

they were supposedly going to separate the coolers for C2D and C2Q according to manufacturer spec. Zalman does list the 8700 for the C2Q.

the temps are the biggest issue. they measured the temp of the cooler, not the cpu. they don't mention how they got the temp or what part of the cooler they got the temp from.

the failing of a cooler due to mounting issues is semi-valid. i've mounted a noctua outside the case with no real problem. mounting it while inside the case was nearly impossible to do with any confidence. if you are the paranoid obsessive type who re-applies TIM paste multiple times in search of those extra 2 deg, then a cooler that doesn't allow remounting while in the case is something to be avoided. [i enjoy the options of socket 775, but i still find myself marveling at the simplicity of old socket 754 and the single crossbar and tensioning cam lever.]

while I wouldn't rely on their results for anything other than one sample of all the cooler reviews, it should be a nice cross section view of all the coolers out there. [btw i just read a review that had the TR 120UE perform worse than the TR 120U. So everything with a grain of salt, etc.]

 

[Skott]

I tinkered with the idea of making a post but since Amber already has this one started I'll ask the question most on my mind about the THG article. How do my fellow enthusiasts feel about THG, or anyone else for that matter, to make hsf installation a part of the ratings score? Does it matter? Should it matter? To some degree I agree with THG that some of these coolers are a pia to install. I'm not sure how much it should affect the scoring though. Your thoughts?

 

[dampeal]

When I review colers I've always factored in how hard they would be to install.. but I wouldn't fail a cooler because it was hard to install..

I try and look at product and review it from a newbie standpoint and from a enthusiast point of view, that being would a person who has a stock cooler and goes out and buys this product have a difficult time installing it, i.e. are the directions easy to follow? do you have to take mobo out of case to install it? etc

I try and think would this be easy or hard for someone who has never done it before to install, and then if it si hard would it be hard for an enthusuaist to install as well? or not?

so it always factors into the review to some point, but most high end coolers require you to take the mobo out of the case, so really a newbie would have a problem with that, or would possibly be afraid to do it... I remember the first time I took apart my computer, I ws sweating and crossing my finger that it powered back on when I put it back together...

 

[AmberClad]

Originally posted by: Skott
I tinkered with the idea of making a post but since Amber already has this one started I'll ask the question most on my mind about the THG article. How do my fellow enthusiasts feel about THG, or anyone else for that matter, to make hsf installation a part of the ratings score? Does it matter? Should it matter? To some degree I agree with THG that some of these coolers are a pia to install. I'm not sure how much it should affect the scoring though. Your thoughts?

My feeling is that installation-related issues should be noted, but I don't necessarily think that a "difficult" installation process would deter most people here. It would be a minor consideration if the cooler performs adequately. Of course, that's barring show-stopping issues, like the cooler simply not being able to clear components in any orientation.

I also think that it can be a somewhat subjective measure, unlike temperatures. For example, with the NT06, even with my motherboard already installed in the case, it would still take me far less time to install that cooler than it would take me to do the stock Intel pushpin cooler, which I've always had difficulty with. I know there are others out there though, who could probably install a pushpin cooler blindfolded.

As far as the fan issue, it's fine to test with the stock fan that comes with the cooler, but I also think it'd be useful to test one particular aftermarket fan across all of the coolers. To get a sense for the comparative performance of the heatsink itself, since it's not uncommon to use a non-stock fan.

Edit: I just realized that they gave the Foxconn cooler, which is essentially a clone of the stock HSF, a passing grade . Erm...I know it's cheap, but there are far better coolers that are only a bit more.

 

[Midnight Rambler]

how do you base a coolers score on how hard it is or isn't to install overall...

Real easy, esp. if you have seen warped mobos, cracked mobos, etc. that were all the result of difficult/over zealous installations.

Recently worked on a friend's PC which was "dead". Sure was, the mobo was badly warped from the install of the stock Intel push-pin cooler. Worked for a while that way but then it "died". Backplates are your friend ...

That said, marking down a cooler due to having to have the mobo out of the case for install is pure stupidity.

 

[dampeal]

warped mobos, cracked etc are not really part of the installation process it's part of the finished installation.. we're talking how hard it is to install or the process of it, not what happened after it was installed...

and yeah I've seen warped mobos... and I've seen coolers with backplates warp mobos...

and I'm not talking about marking it down for having to take the motherboard out, but making a note of it in the review



-I keep forgetting everything needs to be spelled out ot some people....

 

[CurseTheSky]

To me, terrible mounting would be something that merits mentioning in the review, but not part of the rating itself (unless the rating was extremely complex). For example, if you had a solid 10 / 10 (rating) HSF but it took six hours to mount, I would be very tempted to bump it down at least a point. On the other hand, if the HSF was a 3/10 at best but still had the same six hour mounting drawback, I'd be much more tempted to take away two points or more.

Furthermore, mounting isn't relative to so many other factors - the user's experience with HSF mounting, the user's setup (ATX full tower, micro ATX), the size of the user's hands, etc. A HSF may be difficult for one person to mount, and easy for the next.

Above all, my biggest gripe is the fact that they didn't mention the ambient temperature. I would like to believe that a hardware site at least has the ability to mount a HSF properly, so I'm going to give them the benefit of the doubt and guess the room temperature is pretty high. Give me the same processor and that Scythe Kama Cross, which failed for hitting 100C, and I would be willing to bet I can get it to cool just fine. It's somewhere in the low 60s / high 50s in this room (feels very cold).

 

[lopri]

I closed the firefox tab when I saw that test setup. Does anyone run a computer that way? Screwed in two pieces of wooden cardboard?

 

[Cookie Monster]

Originally posted by: lopri
I closed the firefox tab when I saw that test setup. Does anyone run a computer that way? Screwed in two pieces of wooden cardboard?

Ive seen worse. Parts of the PC (no case) like the mobo hanging off a wall, attached by strings and not to mention it working perfectly normal as well.

edit - but these are rare examples. They should have atleast put in the effort of using a high end case with the QX6850. Sure its going to consume alot of time and resource to review all those coolers, but its better than half a$$ing the review and further bringing their reputation down as a hardware site.

 

[aussiestilgar]

Do yourself a favor and skip the review of each individual cooler, and just check the graphs at the end. Their "ratings" are just stupid.

Case in point, they fail the Nirvana because of install issues but it is clearly one of the top performing coolers in the roundup. Pros: Top performance Cons: Can be loud at high speed. I have this cooler and the installation is actually really easy. I'm very happy with it and it only cost me $32AR.

 

[AmberClad]

The Nirvana scores were indeed pretty surprising to me, because of the reasons you just mentioned. I don't know of much anecdotal evidence as far as the Nirvana retention mechanism causing shorts and killing rigs. Personally, I wouldn't let the Nirvana's "failure" score deter me from using it on any of my own rigs, because it's quite a nice cooler.

The Ninja B...I don't know how they managed to install it in such a way as to get those unusually high temperatures. The results fly in the face of the other reviews I've seen for the Ninja. I'm not discounting the possibility of quality control issues with the base, but we don't really know those details, since they didn't measure the levelness/smoothness like Frostytech would.

The NT06, well I already mentioned the issue with using a 760 RPM, 120mm x 25 mm fan on it, since it appears from the chart you just mentioned that all of the rest of the coolers were using 1200-2400 RPM fans. The real stock fan is 2400 RPM, 120mm x 32mm, double ball bearing, so there's going to be a major discrepancy with the amount of air being moved / static pressure. I'm willing to give THG the benefit of the doubt here though, since Silverstone is more to blame for sending them the wrong fan.

The Zalman - I tend to consider that type of cooler more common to HTPCs and low profile cases, which they're well suited for. I don't know if many people would attempt to use that on a quad, despite what Zalman says about it's quad core compatibility. After all, isn't that what the 9500/9700 are for?

The Kama Cross - I haven't seen any review where it's fared as abysmally as it did here. Granted, the unconventional design leads to temps that aren't the best for a Scythe, but its performance still shouldn't be so bad as to cause an overheat situation. Something went wrong here...whether it's simply the cooler itself, or incorrect installation remains to be seen.

They make some negative comments as far as the size of the Thermalright and the fact that it must be installed outside of the size, but they still gave it a pass, which seems awfully inconsistent to me.

The Noctua seemed pretty unremarkable, maybe a touch higher than expected as far as temperatures.

I don't really have an opinion as far as the other coolers though, as I haven't seen those before.

 

[JackMDS]

LOL, when you look at the front page at the pictures of all the coolers, it suddenly becomes apparent that the Major variable is the Visual design.

The HSF manufacturers are late bloomers, they suddenly discovered, that the even enthusiasts get hooked on Fashion and clever visual Design rather than technological substance.

I do not really ?cherish? TH, but it Not their fault that Fashionable Gizmos fail to do what they suppose to do technology wise.