Beware Of Vague or Misleading Heatsink Comparisons

[BonzaiDuck]

Hello, everyone.

I've only posted a few times in the last week or so, and have been absent for some months. If you wish to know why, I should post something under "Social -> Political & News" -- let me put it this way though -- "Fired up! Ready to go!"

My purpose with this post is constrained to scientific objectivity -- untainted by partisan bias.

Last spring and summer, some pretty lively threads debated, proffered test results, argued and compared the virtues of many heatpipe coolers -- notably the spring '07 offerings from ThermalRight and the Sunbeam Tuniq. Anandtech offered two or three reviews showing cumulative comparisons of these and maybe 20 other coolers. The Anandtech comparisons were pretty darned accurate; I and a few others were able to replicate the results.

Since then, a new spate of coolers has been introduced into the marketplace. The latest innovation: "Direct Contact" coolers that do not separate the heatpipe from the IHS by imbedding the pipes in the traditional heatsink base.

Some people have posted references to reviews at FrostyTech and PC Perspective. There have also appeared reviews of the ThermalRight IFX-14 a little late in the game, since it was released last spring.

Earlier this year, there was much hoopla about the OCZ Vendetta 2. Now, I see people touting the SunbeamTech Core-Contact Freezer -- another Direct Contact entry.

I'm linking the FrostyTech comparison review here:

Frostytech heatsink comparison review result table

Let's examine some comparisons in this review. The Sunbeam Core-Contact is shown in top place. Two testing conditions control the results: a 150W over-clock thermal profile, and an 85W over-clock thermal profile. The Core-Contact shows a 16.1C rise over ambient temperature with 150W TDP and a 10C increase under an 85W regime.

Those of us who've had both the ThermalRight Ultra 120 and the TR U-120-Extreme can attest that the Extreme version of the heatsink provides a 5 to 8C improvement over the original version -- since we were able to test these coolers using the same motherboard, processor and CPU fan under measured or controlled room-ambient temperatures.

Yet -- the FrostyTech comparison review shows the Ultra 120 trumping the Extreme cooler -- by some 1.6C degrees. And we know that this is absolutely impossible . . . unless . . . .

Unless we examine the noise-level comparisons, which, in turn, very loosely suggest the use of different fans, at different speeds, with different airflow CFM through the coolers.

For instance, last year, we were able to extrapolate the U-120-Extreme's thermal resistance from two other reviews -- including an OverClockers.com review of the (original) U-120 by Joe Citarella. We estimated that the U-120-Extreme's thermal resistance was about 0.093 C/W. And we must admit that this estimate -- this extrapolation -- was based on airflow roughly equivalent to Citarella's U-120 test, or a 120mm fan speed of 2,500 rpm.

Of course, CFMs vary across fans of the same size with different fin designs, but the airflow doesn't vary all that much among fans running at the same rpm. The variation is a "flea on the back of a flea" in magnitude -- as compared to fans of different size, and fans of the same size running at different rpm.

CFM will vary with fan speed, and thermal resistance will vary with CFM. As CFM increases, so will thermal resistance to some limit.

The Sunbeamtech Core Contact shows a noise level that is deafening. 51.3 dBA approximates a 120x38mm DELTA Tri-blade fan running at its top-end speed of 3,700 rpm -- pushing approximately 142 CFM through the cooler fins.

What is the Core Contact's thermal resistance with this fan?

TR = (C2 - C1) / W where W is the thermal wattage.

I suggest that we can estimate the thermal resistance of the Core-Contact (or any of the other coolers) given the unknown about their fan-speeds and CFM as follows. We have two test results at different thermal wattages. We should be able to compare the temperature changes against the change in wattage and estimate the thermal resistance:

TR = (16.1C - 10C) / 65W = 0.094 C/W

This is -- I think -- what I saw in yet another review which used a more accurate means of measuring thermal resistance.

What about the ThermalRight Ultra 120 Extreme?

The same approach to estimating thermal resistance gives 0.108 C/W -- higher (and less effective) than the estimation we made last summer.

But look at the difference in noise-level! 44.5 dBA is closer to a fan speed of my Delta Tri-Blade at something less than its top-end. It is therefore pushing less CFM through the fins of the Extreme than what appears likely for the Core-Contact. Therefore, these comparisons are borderline-bogus: they do not submit the coolers to precisely the same conditions!

Looking again at the (original) Ultra 120 results -- rated higher than the Extreme -- (which we know to be totally impossible) -- the dBA rating is louder than a DELTA Tri-Blade running full throttle.

Now, here is a comparison among some ThermalRight models, a Scythe, and a ThermalRight IFX-14, published by PC Perspective. At least, the review tests these fans at known fan-speed rpm, nevertheless measuring the dBA noise-level, and providing comparisons between different fans, fans volted differently, and fans running at different rpm for the same cooler:

PC Perspective on the ThermalRight Inferno IFX-14 Cooler

I suggest this: unless you want to spend $100 or so extra finding the "best" cooler by trial and error, you need to examine these reviews more closely, and find shortcomings in the testing approach if there are indeed any to be found.

Otherwise, you could just follow the review like a lemming, buy the Sunbeam Core-Contact (which may nevertheless be a "stand-up" cooler), and find yourself running a DELTA Tri-Blade fan full-bore with a noise level exceeding 51 dBA so you can capture the purported advantages stated in the review!

And may I also have been guilty of this sort of carelessness? Let's wait and see, after I test this darn OCZ Vendetta . . . . . with my Lockheed "Skunkworks" wind-tunnel of a DELTA Tri-Blade . . . .


I'll take your comments, and hope that I've raised your collective attention to these considerations as y'all shop around for yur heatpipe coolers . . . .

 

[toadeater]

HDT heatsinks vs TRUE using low and high RPM fans:

http://benchmarkreviews.com/in...=view&id=156&Itemid=62

 

[BonzaiDuck]

Yeah -- I saw that review. I also posted this review from the same site

Best CPU Cooler Performance Q1 2008

One would have to study the graphs more closely, to see if the actual numerical results are consistent. The bar-graphs in the other review have "stubbier bars."

Unless the 5C difference shown in this second review from Benchmark Reviews is really accurate or substantiated, the OCZ Vendetta 2 seems hardly worth the trouble for trying "something new."

 

[aigomorla]

im always a fan of bigger is better.

yes i like HDT sinks.

However nothing will pry me from the pure mass of my ifx-14.

:X

 

[Tullphan]

Originally posted by: aigomorla
im always a fan of bigger is better.

yes i like HDT sinks.

However nothing will pry me from the pure mass of my ifx-14.

:X

I was getting ready to pull the trigger on that until you suggested that due to my vertical mounting, it wouldn't perform as well.
I'm still using my Xigmatek HDT-S1283, but ain't real happy w/the temps.
I'm using the factory fan on the Xigmatek w/backplate provided by Xigmatek & TIM provided by AS Ceramique. It's installed on a DFI LP UT P35-T2R inside of a Stacker 832. I've got 3 Aerocool Turbine fans in the side panel & stock Coolermaster fans in the front & rear.
My temps @ idle (according to Everest Ultimate Edition) are 39c - CPU, 50c - CPU#1/Core#1, 45c - CPU#2/Core#2. This is an Intel E6300 @ stock settings.
I'm pondering swapping out the factory fan on the Xigmatek w/a Zalman ZM-F2 I have lying around. I haven't decided about swapping out the front or rear exhaust fans yet, but I do have extra 120mm fans lying around.
I'm happy w/the quietness of the case, but those CPU temps need to go down & I can't afford H2O right now! hehe

 

[Old Hippie]

I'm still using my Xigmatek HDT-S1283, but ain't real happy w/the temps.

How 'bout waiting for the new Achillies S1284 version 2 ?

Looks pretty interesting and it should start a whole new round of controversy/discussions!

 

[aigomorla]

Originally posted by: Tullphan

Originally posted by: aigomorla
im always a fan of bigger is better.

yes i like HDT sinks.

However nothing will pry me from the pure mass of my ifx-14.

:X

I was getting ready to pull the trigger on that until you suggested that due to my vertical mounting, it wouldn't perform as well.
I'm still using my Xigmatek HDT-S1283, but ain't real happy w/the temps.
I'm using the factory fan on the Xigmatek w/backplate provided by Xigmatek & TIM provided by AS Ceramique. It's installed on a DFI LP UT P35-T2R inside of a Stacker 832. I've got 3 Aerocool Turbine fans in the side panel & stock Coolermaster fans in the front & rear.
My temps @ idle (according to Everest Ultimate Edition) are 39c - CPU, 50c - CPU#1/Core#1, 45c - CPU#2/Core#2. This is an Intel E6300 @ stock settings.
I'm pondering swapping out the factory fan on the Xigmatek w/a Zalman ZM-F2 I have lying around. I haven't decided about swapping out the front or rear exhaust fans yet, but I do have extra 120mm fans lying around.
I'm happy w/the quietness of the case, but those CPU temps need to go down & I can't afford H2O right now! hehe

and you made a smart choice.

The SINK IS BIG HEAVY AND MASSIVE!

 

[Jessica69]

From what I've been able to glean from FrostyTech's heatsink testing, one should read what fan was used with testing the heatsink in question.

As for the Sunbeam and its fan, that you describe as having to have a Tri-blade Delta on it to have so loud a sound output, I find it ludicrous that you cannot believe that the included 120mm fan with the cooler, running at its full 2000rpm speed at 12V, blowing through the cooler's fins, won't produce that noise level. I know that no matter the fan chosen, making a fan blow through a restriction, like a densely finned air cooler or a radiator, will increase the noise generated, so why is the noise level they measured out of line?

As for the two Thermalright coolers....it may indeed come down to what fan was used with each cooler........if you'd managed to actually read the respective reviews instead of just looking at dBA numbers and flying off to a conclusion, right or wrong, you may have noticed that Frosty Tech did indeed use two different fans for the two coolers......a Martech D712025125Z2N fan on the Ultra and a Mechatronics G1225S12B fan on the Extreme. How much difference this made, I don't know. I'm not going to do a detailed analysis of the two fans, their max CFM output or static pressures, but I'd bet that's where the difference is. And since the two reviews of the two Thermalright coolers were done 7 months apart, I'm not surprised the fans are different.

Should Frosty Tech standardize on one fan to test with on coolers that don't supply a fan with the cooler? Probably...but as long as the information as to what fan was used in testing, I see no misleading or misinformation or whatever going on at all. (And the info is there....you just have to read.)

All in all, I think your rant is more a tempest in a teapot......

 

[BonzaiDuck]

Originally posted by: Jessica69
From what I've been able to glean from FrostyTech's heatsink testing, one should read what fan was used with testing the heatsink in question.

As for the Sunbeam and its fan, that you describe as having to have a Tri-blade Delta on it to have so loud a sound output, I find it ludicrous that you cannot believe that the included 120mm fan with the cooler, running at its full 2000rpm speed at 12V, blowing through the cooler's fins, won't produce that noise level. I know that no matter the fan chosen, making a fan blow through a restriction, like a densely finned air cooler or a radiator, will increase the noise generated, so why is the noise level they measured out of line?

As for the two Thermalright coolers....it may indeed come down to what fan was used with each cooler........if you'd managed to actually read the respective reviews instead of just looking at dBA numbers and flying off to a conclusion, right or wrong, you may have noticed that Frosty Tech did indeed use two different fans for the two coolers......a Martech D712025125Z2N fan on the Ultra and a Mechatronics G1225S12B fan on the Extreme. How much difference this made, I don't know. I'm not going to do a detailed analysis of the two fans, their max CFM output or static pressures, but I'd bet that's where the difference is. And since the two reviews of the two Thermalright coolers were done 7 months apart, I'm not surprised the fans are different.

Should Frosty Tech standardize on one fan to test with on coolers that don't supply a fan with the cooler? Probably...but as long as the information as to what fan was used in testing, I see no misleading or misinformation or whatever going on at all. (And the info is there....you just have to read.)

All in all, I think your rant is more a tempest in a teapot......

I can go back and check for the fan specs on those, and true -- I'm basing my appraisal of the noise ratings against the rated specs of known fans. But with several fans I've used -- Panaflo H1 (controlled), Delta Tri-Blade (controlled) etc., the noise levels at 2,500 are just not capable of 51 dBA.

There are other reviews -- links posted here -- totally inconsistent with the Frosty-Tech results -- using a thoroughly consistent test-bed. And as I said, knowing full well that an Ultra 120 will never outperform a U-120-Extreme, I am very mistrustful of that comparison review. I'll be interested if Anandtech repeats last year's practice with an updated comparison review -- same test bed, same room-ambient -- showing results for these Direct-Touch coolers.

Originally posted by: aigomorla
im always a fan of bigger is better.

yes i like HDT sinks.

However nothing will pry me from the pure mass of my ifx-14.

:X

I'm still eyeing that cooler. It's really not that big. You'll remember the experiment I made last year of turning the TRU-120-E around 90-degrees, making a flared duct, and pushing air through the narrow side of the cooler. With the IFX-14, you could do the same thing with no need for "flared duct" -- possibly only a rectangle of black-foam-board between the two sets of fins to restrict airflow exclusively through those fins.

In Tullphan's remarks, he notes an earlier exchange you had that seems to have focused on "vertical" versus "horizontal" mounting. Does he mean that his cooler was vertically mounted (perpendicular to the ground) or that the motherboard was vertically mounted (as we commonly do with tower cases)?

Anyway, the performance on the IFX is impressive, even though the review link I posted seems to speculate that a chunk of the improvement hinges on the little IFX-10 Motherboard Backside Cooler The fact is, as far as I can see at Sidewinder and other places, the IFX-10 will work with a TRUE, Ultima 90 or the IFX-14. If the performance improvement of the "old-tech" IFX-14 (do direct touch heatpipes) is due mostly to the IFX-10, then I could do just as well keeping the little Ultima 90 I currently have and adding the IFX-10 to it.

But back to Jessica69's remark. There can be implicit payola between manufacturers and reviewers, if the reviewers also advertise the product that they review (advertising being a major revenue source). This is evident in Maximum PC Magazine's choice of products they review -- with the review just appearing before a full-page ad of the unit just rated with a Kickass 9 award. The "comparisons" often exclude products known to perform better than items included in the test, as shown by other review comparisons elsewhere.

Those of us looking for the "best of the best" don't mind comparing multiple comparison reviews -- not to seek "second opinions" would be like saying you read one book about the JFK assassination and you know who done it. But -- if we descend into a wallet-draining trial-and-error cycle . . . . it can be more costly than it needs to be.

 

[toadeater]

Originally posted by: BonzaiDuck
If the performance improvement of the "old-tech" IFX-14 (do direct touch heatpipes) is due mostly to the IFX-10, then I could do just as well keeping the little Ultima 90 I currently have and adding the IFX-10 to it.

I didn't know the IFX-10 worked with any heatsink. Interesting idea for creating a hybrid HDT-Thermalright monstrosity, if it really can improve temps by more than a couple degrees? It would cost about the same as a TRUE with a cheap fan. But installation looks like it might be a little troublesome. It doesn't look like it would fit in cases with the PSU on the bottom.

http://www.thermalright.com/ne...cpu_cooler_ifx-10.html

Tullphan - According to Everest, my E7200 @ 3.5GHz, 1.25v in a CM690 with 4 case fans gets 53c core, 40c CPU running Prime95. At idle, it gets 40c core, 27c CPU. The stock fan on mine runs at 925RPM idle and 1200RPM load, so the temps could be lower with a faster fan I guess. I'm not using the bolt-through kit, and this was with the thermal paste that came with the heatsink.

(Don't ask about the strange 3.5GHz... I'm still trying to work out the cold boot problems with my motherboard. Temps weren't much higher at 3.8GHz, 1.3v.)

Old Hippie - the Achilles hasn't done too well in web reviews so far, and the review at Newegg confirms the same results. I don't mean the one from the noob babbling about it not fitting in his case and push-pins, the second review about the temps.

http://www.newegg.com/Product/...x?Item=N82E16835233017

 

[Old Hippie]

Old Hippie - the Achilles hasn't done too well in web reviews so far, and the review at Newegg confirms the same results. I don't mean the one from the noob babbling about it not fitting in his case and push-pins, the second review about the temps.

That was the first version. Click my link and look at the difference between the bases between versions I and II.

It looks like the base problem with the first version has been fixed.

Only time and a few tests will tell if there's been an improvement.

 

[Tullphan]

When I stated "vertically mounted", I meant mounted "normally" on a motherboard mounted in a tower case (Stacker).

Originally posted by: Old Hippie

I'm still using my Xigmatek HDT-S1283, but ain't real happy w/the temps.

How 'bout waiting for the new Achillies S1284 version 2 ?

Looks pretty interesting and it should start a whole new round of controversy/discussions!

Looks like there was alot of controversy in that thread! hehehe

 

[BonzaiDuck]

Originally posted by: Tullphan
When I stated "vertically mounted", I meant mounted "normally" on a motherboard mounted in a tower case (Stacker).

Originally posted by: Old Hippie

I'm still using my Xigmatek HDT-S1283, but ain't real happy w/the temps.

How 'bout waiting for the new Achillies S1284 version 2 ?

Looks pretty interesting and it should start a whole new round of controversy/discussions!

Looks like there was alot of controversy in that thread! hehehe

Some of the results from the PC Perspective reviews don't suggest to me that the margin of improvement due to orientation is all that worthy of concern. Those results were obtained with a TRUE and OCZVEND2 -- and a Xigmatek, but the indications may be more generally applicable.

The TRUE's orientation wouldn't be that much different than the IFX-14, regarding the pipes to vertical or horizontal.

Now that I see it again (checking pictures at HeatsinkFactory), it sure uses a lot of mobo-real-estate. My original concern with it was height, but it is stubbier than a TRUE.

I think that a lot more clarity will shake out concerning these direct-touch coolers, but it will take some more time.

Now . . . about that IFX-10 back-side cooler. It's supposed to work fine with all TR coolers that use their common under-the-mobo bracket. And I've been having some thoughts about this.

For instance, with my OCZ Vendetta 2 -- still in the carton -- I can see that it uses push-pins and mounts in the same holes. The TR hardware -- springloaded screws and the backplate -- would work equally well with the Vendetta2. So I can imagine using TR hardware to mount some of these other coolers. If that is so -- then you might be able to use the IFX-10 with coolers of a different manufacture.

 

[aigomorla]

you guys want my opinion about today's coolers?

Okey, we saw this in waterblocks.

But i think were at the max potential we can take air sinks now.

HDT is an experiment to go in another direction, because air is pretty much capped.


Sinks cant get larger without risk of hardware damage.

And large sinks perform on the same level as small sinks with low heat load.

So that means for a sink to truely outperform another, you would need larger heat loads.


OK, problem with larger heat loads, you require more voltage. More voltage kills the 45nm chip.


If anything we need to shift to focus cooling. This is the methodolgy ive been looking at in water. In water we call this term turbulance. Accerated water is like sandblast the heat off copper base and strip it away.

Since foot print is smaller and heat output is increased, im guessing cooling designers would need to focus concept on somehow getting supercooling in a tiny little spot.

IE HDT coolers.



Duck the HR-10 is crap. If it works and you get improvement, it means your case flow is crap. :X

Why they have it i have no clue. Is it useful, hellz no. Do i use it? Yeah right, its usless on watercooled rigs and open benches.

 

[imported_Scoop]

I only trust SPCR reviews for heatsinks, fans and power supplys.

 

[Gillbot]

I don't trust all heatsink reviews for the main reason they rarely keep the fans consistent. It's a nice tool for comparison, especially if the heatsink in question has a pre-installed fan but otherwise, you are at the mercy of the reviewer and whatever fan they decide to use.

 

[9nines]

Is the Kingwin Revolution HDT 120mm CPU Cooler RVT-12025 pretty much the same unit as the Xigmatek HDT-S1283?

 

[BonzaiDuck]

Originally posted by: aigomorla
you guys want my opinion about today's coolers?

Okey, we saw this in waterblocks.
[1] But i think were at the max potential we can take air sinks now.
HDT is an experiment to go in another direction, because air is pretty much capped.
[2] Sinks cant get larger without risk of hardware damage.
And large sinks perform on the same level as small sinks with low heat load.
So that means for a sink to truely outperform another, you would need larger heat loads.
OK, problem with larger heat loads, you require more voltage. More voltage kills the 45nm chip.
[3] If anything we need to shift to focus cooling. This is the methodolgy ive been looking at in water. In water we call this term turbulance. Accerated water is like sandblast the heat off copper base and strip it away.

Since foot print is smaller and heat output is increased, im guessing cooling designers would need to focus concept on somehow getting supercooling in a tiny little spot.

IE HDT coolers.



[4] Duck the HR-10 is crap. If it works and you get improvement, it means your case flow is crap. :X

Why they have it i have no clue. Is it useful, hellz no. Do i use it? Yeah right, its usless on watercooled rigs and open benches.

[1] You may be right, unless someone comes up with a better innovation than just putting the heatpipe right on the IHS.
[2] That's also true, but most manufacturers have kept their product below 1,000 grams in weight. The IFX-14 is 790 Grams. I haven't checked the latest-gen mobo-specs recently, but there was a time when Intel recommended something less than 450 grams. Yet -- we've seen that keeping the weight to within double that doesn't have an appreciable risk of damage. Anyway, the damage would result from torque; if most of the weight is flush against the board, I don't see how it would stress it too much.

[3] with the smaller die-sizes, -- yup -- that's less real-estate to dissipate heat.

[4] Looking at this and looking at the entire thread, I have to note several things:
-- Yes -- if under-the-mobo-cooling is inadequate, the advantages of this IFX-10 would be greater.

-- Also, it adds complexity. Another part. Also -- I'm basing my speculation on a SINGLE REVIEW -- which I used to start this thread. THEY SPECULATE that the added performance is due to the IFX-10 -- but they don't prove it. Maybe it's worth a couple C degrees. There's only one way to find out -- which is another . . . . P-I-T-A.

Answering Gilbot: These gadgets cost around $50. At current wage-rates, figure the price of the cooler translates into an hour or two worth of time. I'd guess that you could spend three, four . . . or more . . . . times that just going through the review of reviews. Throwing out the economic logic though, you'd want at least two or more comparison reviews, each listing results with as many sinks as possible.

Like Aigomorla said, the declining marginal returns to heatpipe/sink technology means that we're splitting hairs -- possibly over a couple degrees difference in performance. That FrostyTech review may exaggerate things -- with the different fans and test conditions -- or simply confuse them. The only thing I'm sure of now is that the SunbeamTech cooler has four direct-contact pipes, while other models seem to have three. I think the Achilles picture showed four.

 

[aigomorla]

Originally posted by: BonzaiDuck

Like Aigomorla said, the declining marginal returns to heatpipe/sink technology means that we're splitting hairs -- possibly over a couple degrees difference in performance. That FrostyTech review may exaggerate things -- with the different fans and test conditions

You wanna prove them wrong?

Lemme see if theres a way to send you my ifx-10 for you to test.

You'll be glad i sent you one and you didnt buy one.

EXPECIALLY since you duct.

 

[BonzaiDuck]

I could take you up on that, but I'm getting so old with so much time on my hands that I'll spend days deciding whether or not to do something, trying to second-guess whether it's worth doing.

I was thinking about your earlier remarks, and went back to Sidewinder's ad:

IFX-10 at Sidewinder Computers

Scrolling to the bar-graph, we see that they used an Intel stock CPU cooler, and that the IFX-10 reduces load temperatures by 4C degrees.

Now we've been this route before. For instance, with my ducting-mods. The ducting is great for motherboard components, because less attention had been paid to them as we went through several generations of boards. Belatedly, they've beefed up chipset cooling, eliminated capacitors for solid-state components, added copper heatpipe necklaces -- geez! -- even providing a built-in water-block for some chipset-sinks in addition to the heatpipe.

But with the improvement in CPU coolers, the ducting provides marginal benefits per the CPU temperature alone.

Now going back to this IFX-10. Suppose you replaced the Intel stock cooler in their test with a TRUE? My bet: you wouldn't see a 4C reduction in temperatures. It would be less -- how much less -- I don't know. So you wonder -- really -- why they included it with the IFX-14 anyway. I could be right or wrong here, but at LEAST it would serve to improve memory cooling if you mounted a fan on it! But I have that problem licked without any fan . . . . [like the movie: "DucT, you Sucker!"]

 

[aigomorla]

Originally posted by: BonzaiDuck

Now going back to this IFX-10. Suppose you replaced the Intel stock cooler in their test with a TRUE? My bet: you wouldn't see a 4C reduction in temperatures. It would be less -- how much less -- I don't know. So you wonder -- really -- why they included it with the IFX-14 anyway. I could be right or wrong here, but at LEAST it would serve to improve memory cooling if you mounted a fan on it! But I have that problem licked without any fan . . . . [like the movie: "DucT, you Sucker!"]

which is why i said if you see improvement with it, you have problems with your case temps.

Duck i see you live about 1 hour from me.

maybe we can meet up for coffee and you can just pick it up. Test it, after your done playing with it, drive it back to me.

 

[Tullphan]

Originally posted by: 9nines
Is the Kingwin Revolution HDT 120mm CPU Cooler RVT-12025 pretty much the same unit as the Xigmatek HDT-S1283?

yes

 

[Gillbot]

Originally posted by: BonzaiDuck
That's also true, but most manufacturers have kept their product below 1,000 grams in weight. The IFX-14 is 790 Grams. I haven't checked the latest-gen mobo-specs recently, but there was a time when Intel recommended something less than 450 grams. Yet -- we've seen that keeping the weight to within double that doesn't have an appreciable risk of damage. Anyway, the damage would result from torque; if most of the weight is flush against the board, I don't see how it would stress it too much.

Most bow the board now and not to mention even if most of the weight is against the board, there is still some hanging out away which leads to a torsional effect on the board and socket. Compound this with the ever increasing weight and SIZE of todays sinks and I do think it's an issue.

 

[BonzaiDuck]

Well, Gilbot -- the thing I've observed over a few years: These heatpipe coolers, if the fins are aluminum, have most of the weight at the sink. Actually, I'm stunned that the IFX cooler is 790 grams, but that's less than the Zalman "copper-flower" coolers.

But here's the rub. It's the fans. Maybe some are happy with 120x25mm fans -- that's OK. I'm inclined toward 120x38mm fans. Some of these beefier fans weigh as much as 250 grams -- and that's definitely "torque" -- hanging as it does off the heatpipes and fins.

In the last machine I built, the TRUE was serviced by a 120x38mm exhaust fan bolted to the case, and fed by a 140x25mm intake fan mounted on the drive-cage-rear. So basically, there was no extra weight on the TRUE. Now, just looking at specs again, I now see that the TRUE is 790 grams and equal in weight to the IFX-14.

 

[Gillbot]

Originally posted by: BonzaiDuck
Well, Gilbot -- the thing I've observed over a few years: These heatpipe coolers, if the fins are aluminum, have most of the weight at the sink. Actually, I'm stunned that the IFX cooler is 790 grams, but that's less than the Zalman "copper-flower" coolers.

But here's the rub. It's the fans. Maybe some are happy with 120x25mm fans -- that's OK. I'm inclined toward 120x38mm fans. Some of these beefier fans weigh as much as 250 grams -- and that's definitely "torque" -- hanging as it does off the heatpipes and fins.

In the last machine I built, the TRUE was serviced by a 120x38mm exhaust fan bolted to the case, and fed by a 140x25mm intake fan mounted on the drive-cage-rear. So basically, there was no extra weight on the TRUE. Now, just looking at specs again, I now see that the TRUE is 790 grams and equal in weight to the IFX-14.

But most users strap on 2 fans to the true and have at it. The weight of the true + fans plus the rotational torque from stop/start of the fans has to take it's toll over time IMHO.