Thermalright T-rad2 GPU cooler reviewed!

[Helmore]

4850 core temps under load

So you have the exact same system under the exact same circumstances, the only difference being your cooler and under those conditions your cooler fares better. Then we are not even considering the inaccuracy of the on board temperature sensor and inconsistencies between different cores. I don't think you can draw any conclusion on this cooler yet, not without someone doing the test under better and more consistent circumstances.

[cegras]

I'm so sick of guys with theories based on the color of the sky,when I talk out of experience. I already have a HD 4850 which is properly cooled. Well the GPU is properly cooled, cause the VRM can't take an OC with those RAM Sinks.

Even guys with Watercooling have similar problems, cause that WC doesn't apply on VRM area, only the GPU gets cooled. Here's others with similar problems:

http://forums.amd.com/game/messagevi...threadid=98153

http://forums.techpowerup.com/showthread.php?t=66763

http://www.rage3d.com/BOARD/showthread.php?t=33928000

http://www.rage3d.net/board/showthread.php?t=33928548

===========

Oh, yeah I see you own a BFG 9600GT OC, yet you contradict with my opinions as a HD 4850 user. .

So, uh, wait, your card has died because the vrm's overheated?

Oh.

I see.

*nods*

I thought so.

It hasn't.

[XSAlliN]

Please, STOP Trolling! I'm talking about serious issues with after cooling of a product that I personally posses and fought of an idea to solve those problems. You have no knowledge of this issues (since you own a different product from a rival company), yet you come here and trash my opinions with ideas based on the alignment of the stars or color of the sky.

[initialised]

Anyone know if any vendors are planning to prefit these for silent edition cards?

[trans am]

I like XSAlliN's idea. he' thinking in the right direction. vrm area is a pain to cool. It can make or break a successful vga overclock. 3rd party cooling companies need to focus on vrm cooling. making dinky heatsinks with thermaltape for vrm area is not a solution.

[Helmore]

Anyone know if any vendors are planning to prefit these for silent edition cards?

I'd say the chances of that happening are low, but that's only because I've not yet seen an AIB with cards that feature Thermalright coolers. I've no idea what the future plans of Thermalright are, so that might change.....

[cegras]

Please, STOP Trolling! I'm talking about serious issues with after cooling of a product that I personally posses and fought of an idea to solve those problems. You have no knowledge of this issues (since you own a different product from a rival company), yet you come here and trash my opinions with ideas based on the alignment of the stars or color of the sky.

I'm not trolling. Rather, my first post was quite neutral and then you spat all over it, so I responded as such.

And you know what, has your product failed? No? No.

I'll put it bluntly for you:

heat != temperature. If a card is at a certain state and it's dissipating heat at some, say, 50 watts, it's going to dissipate 50 watts no matter how cool it gets.

So unless you get an exhaust-out-the-case solution, you're dumping the same amount of heat into your system.

My previous card, for your information, was a 3850 256. So yeah, I did fight with the same 'problems' as you did. Same 80 degree loads under games.

VF700 + Enzotechs = problem solved.

[oohms]

I like XSAlliN's idea. he' thinking in the right direction. vrm area is a pain to cool. It can make or break a successful vga overclock. 3rd party cooling companies need to focus on vrm cooling. making dinky heatsinks with thermaltape for vrm area is not a solution.

Thats going to be hard... every design has a totally different layout for the VRM's, so the most universal solution is to provide little crappy sinks with thermal tape that you stick onto each one.

Personally id be delighted if a company were to make something like the D-TEK unisink, but only for the vrm area. They would be cheap and not that hard to manufacture. (There would just be heaps of different models)

[XSAlliN]

It's not that difficult for a company like Thermalright, or even others with similar equipment. Actually is quite easy, I've seen other designs from their cooling solution which or more difficult to make (even the one from this topic). They showed us a lot of times that they are capable of making excellent matrices, for producing quality products like the one above.

heat != temperature. If a card is at a certain state and it's dissipating heat at some, say, 50 watts, it's going to dissipate 50 watts no matter how cool it gets.

So unless you get an exhaust-out-the-case solution, you're dumping the same amount of heat into your system.

My previous card, for your information, was a 3850 256. So yeah, I did fight with the same 'problems' as you did. Same 80 degree loads under games.

VF700 + Enzotechs = problem solved.

Yes, Heat = Temperature ... what do you think I was talking about, power consumption?

It's true that is also related to power consumption, since more power equals more heat. You go for a little OC and the card can take it and no the it's not the GPU whom fails it's the VRM, since the GPU is properly cooled and the VRM is not. About your last card the HD 3850, it's irrelevant since my last card was an HD 3870 and I didn't had problems with the VRM on that card.The problems come with current generations HD 4xxx where the VRM gets really HOT and we have no solution to properly cool it if we aim for an aftermarket cooling solution. The net is full complains about HD 4xxx crashing in games with intensive graphics and most blame it on driver, since some users have water cooling for their VGA which cools their cards down to 36*C in load so those guys don't expect to be a heat problem. Little to they know, that their GPU runs at 36*C in Load wile their VRM goes beyond 100*C. HD 4870 has a digital VRM and you can see his temperatures in LOAD and it fails beyond 120*C:

[KoHaN69]

Can this be mounted both ways like the hr-03?

xsallin- wat about thermalrights mosefet sinks, looked at those?

[Helmore]

Thats going to be hard... every design has a totally different layout for the VRM's, so the most universal solution is to provide little crappy sinks with thermal tape that you stick onto each one.

Personally id be delighted if a company were to make something like the D-TEK unisink, but only for the vrm area. They would be cheap and not that hard to manufacture. (There would just be heaps of different models)

Something like this:
http://zalman.co.kr/ENG/product/Prod...ad.asp?idx=189
but then for many more different GPU models and maybe a bit better quality.

Can this be mounted both ways like the hr-03?

Have you even looked at the cooler? The answer, no.

[XSAlliN]

Can this be mounted both ways like the hr-03?

xsallin- wat about thermalrights mosefet sinks, looked at those?

Yes, but they only go with small coolers. There's no way to pare up this with Acccelreo S1 or even the one in this topic.

Oh yeah, got an answer:

Dear Sergiu,



Thank for your mail and suggestion.

To develope a new product or coming up a new product combination have to take many aspects into consideration.

Although your idea might work, but it also bring new compavility issues.

In the limited spce in the chassise, size is the issue.



Furthermore, our HR-03GT and T-Rad2 is really good for modification.

As long as you have follow the on-line mannual to clean the chip thoroughly, the ramsinks will work great.



Please read the FAQ2 to check the standard procedure.

http://0rz.tw/a64HK


Best regards
Armand Cheng
Lea-Min Tech., Co., Ltd.

(Thermalright, Taiwan)

Ramsink they say, already have that and works just as it should - not to good.

PS.Guess they didn't checked there spam filters cause my IP got banned wile trying to get my message to them.

[Helmore]

Ramsink they say, already have that and works just as it should - not to good.

I remember someone 'gluing' his RAM sinks to the PCB with some nail polish and then he simply used normal TIM for the contact area with the chip. It helped quite a bit with lowering temperatures according to him. Although that was quite a while ago and I have no idea where I have to look up his posts.

[XSAlliN]

The gluing is not problem, at least not in my case - the performance is. With every generation the GPU got hooter and hooter, so they redesign the cooling solution. But now even the memory gets hooter, luckily it has big resistance but few take the VRM heating in consideration. I've seen some models with a more potent VRM when it comes to high temperatures like the one on Palit HD 4850 Sonic, but reference model has a less potent VRM.

[deathman20]

Thats cool. NOW RELEASE IT Been waiting for it and really need it to complete my setup.

[Zaskar]

Individual sinks aren't the optimal soloution, but I don't know why anyone uses those horrible aluminum ones when they can get the forged copper Enzotech ones, which also use some of the highest quality 3M thermal tape you can get.

[XSAlliN]

The answer is quite simple, case they're not available in every country, yet I see they're full CU and might do better then the aluminum models. If they atached some piepes to that and cover the size of the VRM might do a great job.

[mr2sw21]

Thats cool. NOW RELEASE IT Been waiting for it and really need it to complete my setup.

me 2. I been waiting for it from june or may.

[KoHaN69]

The Thermalright T-rad2 is coming hitting stores in September
there will be 2 revisions of it - for new nvidia cards 260/280 and another revison for normal cards (8800GT/38x0/48x0/X1x00/7900)
It's a single-slot solution and can only be mounted on one side of the card (gpu side) not both siides like the hr-03

-from thermalright

I personally dont like it more than the hr03 since it has a higher fin density, meaning it require higher cfm, which means more noise

[Helmore]

I personally dont like it more than the hr03 since it has a higher fin density, meaning it require higher cfm, which means more noise

That's not completely true, the higher fin density is offset by a decrease in thickness which probably makes it have a lower resistance than the HR-03. Although I don't think there is that much of a difference and Thermalright is probably the only one whom knows these kind of things for sure.
BTW fin density of the HR-03 is 0,425 fins/mm, while this is 0,5 fins/mm on the T-Rad. So there is not a whole lot of difference between the two.

[deathman20]

Main reason for the T-rad2 is the thin-ness. Making it suitable for higher performance fans on CF and SLI. Im totally down for it, TR hasn't let me down yet and as said dying to get 2 of these bad boys for my CF setup. Might not be for pure cooling might, but the fact that I know these can and will be much quieter and cooler than the stock fans at higher speeds which is thumbs up for me.

[sf rich]

going to start a new thread on related subject ... if anyone gets their hands on a T-RAD2 please post your results! We're all dying to see how the thing actually works.

[LuckyNV]

apparently due in 7th Oct at Scan for UKers

[nullface]

Arent there more reviews T-Rad vs HR03?

[XS2K]

Truth be told 4850 is a fine example on how a cooler shouldn't be built,the card gets ridiculosly hot,mine hadn't had problems with stability @ stock clocks but when OC'ed it stopped @ ~670Mhz/1050 which was really low even after V-mod I couldn't hit more than 700/1080 but with the addition of a 120mm fan @ 1100RPM things got stable @ 780-1140,cooling the VRM is a tricky thing even with the fan the temp is hovering around 93 degrees
Also the fan get's really loud @ over 35% when it touches 80% the thing sounds like a vacuum cleaner.