玩家论坛's Archiver

Onepagebook 发表于 2007-5-5 16:28

空冷散熱新王者, Thermalright Ultra 120 extreme

Part I & II
We have seen so many "heavy duty" coolers since Intel's Core 2 Duo series processor hit the retail market. They are usually marked as "....can be applied up to some serious percentage after overclocking. However, what is the truth behind all those "heavy duty" CPU coolers? The most popular air cooling way of CPU coolers from the moment C2D appeared is so called the "heat pipe cooled" which focus on the low noise level but can reach high efficiency. Such kind of coolers are all over the market currently; from my personal point of view, there are some of them really can reach what they guaranteed from the advertisement and even give you beyond what you expect. Today, we are going to take a serious look on the newest product from Thermalright's improved version of Ultra-120 , "Ultra 120 Extreme".
[color=blue]第I、II部分
自从Intel Core 2 Duo系列处理器在零售市场上发布以后,我们看到了很多“重型”散热器也随之出现。对于这些散热器往往都会标注对超频有着直接的帮助等字样,如“超频幅度可以达到一定的百分数”等。但是,在这些“重型”散热器背后的真相到底是怎样的呢?当C2D处理器出现后,最流行的风冷散热器被称为“热管散热器”,这种散热器的关键在于工作时的低噪音以及提供高散热效能,现在这类散热器充斥整个市场,以我个人观点来看,在这些散热器中还是有部分产品可以达到广告中承诺的效果甚至超出你的期望值。今天,我们就仔细来看一看THERMALRIGHT的最新产品,ULTRA-120的升级版--ULTRA 120 EXTREME。[/color]
From the website press release news, we can easily know that this version of cooler is really the one you need if its performance can be just as well as its advertisement. From this point, I am so looking forward to checking out this heat dissipation monster under the Intel's C2D flagship, the quad core extreme qx 6700 processor.. In this short review, I will have have 3 parts  of analysis, they are:
Part I, The cooler and mainboards compatibility
Part II, The basic performance monitoring record and the serious heat load bench , (i.e., Orthos for multiplier CPU)
Part III:The efficiency of heat dissipation under 3d gaming. (full loaded)
[color=blue]我们通过从网上的获知的消息可以确定如果这款散热器的性能真象是广告中说到的,那么它就值得你拥有。冲着这点,我找到了这款热管的散热怪兽,用它来镇压C2D的旗舰QX6700处理器。在这篇短短的介绍中,有三个部分的分析,分别是:[/color]
[color=blue]第I部分:散热器与主板兼容性[/color]
[color=blue]第II部分:基准性能测试记录与严谨的发热负载测试(ORTHOS多核版本)[/color]
[color=blue]第III部分:在3D游戏下的散热效能测试(满负载)[/color]
*******************************************************************************
now , let's look at the Cooler and its accessories. Note the changes compare with ultra 120 are really necessary not only by the cooler physical functions but also the mounting tools and "the electronic short circuit prevention from the back plate." I believe that you can't see that type of back plate from the utral 120.
[color=blue]我们先来看一看散热器及其配套的附件。请注意与ULTRA 120相比的一些必要的变化,不仅仅体现在散热器的物理变化,在安装扣具方面也有不同,“背板采用防止短路的设计”我相信在ULTRA 120上是看不到这种样式的背板的。[/color]
[img]http://www.opbcorner.com/HWbench/tru120e/u120e/cooler/u120e.jpg[/img]
[img]http://www.opbcorner.com/HWbench/tru120e/u120e/cooler/plates.jpg[/img]
** I really like the screw with that type of spring without sticked to the screw, so when install the cooler, you can hold the spring and gentally use your hand screw that into the hold couple turns, and the whole cooler put on the CPU IHS will not out the position it should be.
well,now, we are looking is the bottom finish of U120 extreme. It has been a "must check" when you buy a new cooler. Even though it is not copper type of shinny  bottom. However, you won't be disappointed by it's smoothy  feeling and
cool looking.
[color=blue]**我真的很喜欢这种螺柱上有弹簧的螺丝,在安装散热器的时候,你可以拿着弹簧,轻轻地用手就可以上紧螺丝,整个散热器安装在CPU IHS上的时候就不会发生位置上的变化(不会引发安装时的位置偏差)。([/color][color=red]IHS:Integrated Heat Spreaders,CPU的导热保护外壳,又名集成式散热器)。[/color]
[color=blue]我们来看一看U120EXTREME的散热底座,这可是当你购买一款新散热器时都“必须检查”的部位,即便不是铜质的底部。当然,你不会对U120EXTREME的底座失望的,因为那里看上去非常光滑,非常的COOL。[/color]
[img]http://www.opbcorner.com/HWbench/tru120e/u120e/cooler/bottom.jpg[/img]

Now it's the time to look more details about it. The main improvement from this monster is each side has extra 2 heat pipes compare with Ultra 120 without enlarge its physical size. Basically this change will speed up the heat dissipation  from the base
[color=blue]接着我们来看看这款散热器的其他部位的细节图片。与ULTRA 120相比,主要的变化就在于增加了两根热管而散热器的整体尺寸没有扩大。这种变动带来最直接的影响就是加速了散热器底座的热传导。[/color]
[img]http://www.opbcorner.com/HWbench/tru120e/u120e/cooler/12pipes.jpg[/img]
[img]http://www.opbcorner.com/HWbench/tru120e/u120e/cooler/fins.jpg[/img]
[img]http://www.opbcorner.com/HWbench/tru120e/u120e/cooler/sidepipes.jpg[/img]
[img]http://www.opbcorner.com/HWbench/tru120e/u120e/cooler/sides.jpg[/img]

Onepagebook 发表于 2007-5-5 16:29

Now, the next , we are about to get the U120 extreme on the board and see if there any issu  regarding installation and imcompatibilities.

Basically, this type of cooler has two orientations to choose from when you install. According to the Thermalright engineer told me, the better performace position of install is just as the picture I show below:
接着我们把U120 EXTREME进行了安装,检查一下散热器会不会有安装兼容性的问题。

[color=blue]基本上散热器的安装方向有两种,THERMALRIGHT的工程师告诉我,按照我下面照片的安装方向来安装可以取得最佳的散热效能。[/color]

[img]http://www.opbcorner.com/HWbench/tru120e/u120e/cooler/onboard01.jpg[/img]

However, here is a little problem when I try to install HR05 on the P965 chip; from the picture show, you might feel is still has room for that. however, that is before I hook the fan wire clip on. If I hook the clip on, then I have no way to install HR05 on balance of that I965 chip, that would postively damage the ic itself
[color=blue]但是,在我试图把HR05安装到P965芯片上的时候碰到了一个问题,如图所示,你可能觉得那里还有空间可以安装,然而,那是在我把风扇线夹安装上去之前了,如果我把夹子勾住,那就无法安装HR05,这将会损坏IC芯片。[/color]

[img]http://www.opbcorner.com/HWbench/tru120e/u120e/cooler/onboard02.jpg[/img]

see that space between them?
[color=blue]能看到这两个散热器之间的距离吧?[/color]
[img]http://www.opbcorner.com/HWbench/tru120e/u120e/cooler/probhr0501.jpg[/img]

So I end up will place the second choice, Enzotech's pure copper nb cooler instead of the great HR05;)
[color=blue]于是我只好用Enzotech的纯铜NB散热器来取代HR05。[/color]
[img]http://www.opbcorner.com/HWbench/tru120e/u120e/cooler/onboard03.jpg[/img]


**************************************************************************

Next, in order to be fair and accurate, I will show you where I plan to put the prob cable tip of DT200.
I figured there are 4 holes that tip can get in and won't block the socket T retention, meanwhile, the tip is also fairly closed to the place where we want to obtain the temperature.. so I decide it to put the prob tip there, it should be good enough to get temperature accurate.

[color=blue]为了公平及准确,我会把DT200的探温线的摆放SHOW给大家看。
我测量了一下,在处理器装在底座后四边有四个小孔可以让探温头穿过,同时,那里离我们想测温的地方也非常的近,于是我决定把探温头放在那里,足够我取得准确的温度了。[/color]

[img]http://www.opbcorner.com/HWbench/tru120e/u120e/cooler/dt200prob01.jpg[/img]

Then next we need is plug in another k-type of prob cable to measure the ambinent temperature. and here shows about 26.9c for both T1 and T2.
[color=blue]接下来我们需要另一个探温头来获取测试环境温度,通过T1与T2的显示我们知道环境温度在26。9摄氏度。[/color]

[img]http://www.opbcorner.com/HWbench/tru120e/u120e/cooler/dt200prob02.jpg[/img]

now everything is set, my testbed looks like this,
[color=blue]万事俱备,我的测试平台如下图所示。[/color]

[img]http://www.opbcorner.com/HWbench/tru120e/u120e/cooler/testbed01.jpg[/img]

From that picture, you might feel weird , yes it is really the P5b deluxe "1.10g", and I have no idea why they did not use right pcb and just use P5b-e Plus pcb,, I did have vmod on this board for VNB, Vdimm, vdroop, as well as Vcore. here you see the the red cable is vNB, green is Vdimm, yellow is UEI DT200 prob cable.:)
[color=blue]看着上面的照片你是不是觉得有点怪,不错,这就是P5B DELUXE,我不明白为什么这款主板使用的是P5B-E PLUS的PCB,我在这块主板上面做了一些相应的改造,如VNB,VDIMM,VDROOP,VCORE等。你可以看到红线是VNB,绿线是VDIMM,黄线是UEI DT200的探温线。[/color]

**********************************************************************************************

** Before we get to the result page, here I also use another great and smart product from Thermalright, HR09, HR09 is for Mosfets cooler, I will have detrails introduction these couple days.
[color=blue]在我们去看测试页面前,我再向大家介绍一下THERMALRIGHT的另一个产品,MOSFET散热器--HR09,详细的介绍过几天再放上来。[/color]

[img]http://www.opbcorner.com/HWbench/tru120e/u120e/cooler/hr09su.jpg[/img]

installed:
[color=blue]安装:[/color]

[img]http://www.opbcorner.com/HWbench/tru120e/u120e/cooler/hr09onboard.jpg[/img]

********************************************************************************

Onepagebook 发表于 2007-5-5 16:29

Now for the performance and temperature monitor, I will do my best to get that under unbias and accurate. my testbed details is as followed:

[color=blue]我尽量做到准确详实的记录,下面是测试平台:[/color]
CPU: Intel Core 2 quad extreme 6700
mainboard: Asus P5b deluxe rev. 1.10g bios 1101
PSU: Silverstone  OP 1000W  single rail 80a 12v
Hard druve: Seagate Barracuda 7200.7 IDE 100gb
OS: Windows XP sp2
RAM: Team Group pc8500 CL 4
VGA: Geforce MX4000 PCI
CPU cooler: Thermalright Ultra 120 Extreme version with Evercool 120mm fan

************************************************************************************************

OK, let's begin the bench.

in order to the correctly records, we need to know what the exactly difference between bios and
UEI DT200 readings, so I cleared the CMOS and boot into bios , load the default setting and immediately save out to reboot into bios. I then check the pwm monitor  page,  the bios shows 28c for both CPU core temp and board temp.
[color=#0000ff]开始测试。[/color]
[color=blue]为了能够让数据更加准确,我们需要知道BIOS与UEI DT200对温度记录的不同到底在哪里。于是我清空CMOS进入BIOS,载入默认设置并立即重启进BIOS,打开PWM监视页面,BIOS里显示CPU的核心温度与主板温度都是28摄氏度。[/color]

[img]http://www.opbcorner.com/HWbench/tru120e/u120e/piorthos4cores/biostemp01.jpg[/img]

The reason I use quad core for bench is because it will give me the max heat load compare with some x6800 is kind of cool and inaccurate. Here is the picture that you can see what I use. And the qx6700 is still in 1.34~1.35v for that, and dt200 reading is up to 31.5 already.:)
[color=blue]我之所以使用四核心进行测试的原因就在于与一些X6800相比它可以提供最大的负载,请看下面的照片,当QX6700的电压设定是1.34-1.35V时,DT200的温度显示已经是31.5摄氏度了。[/color]

[img]http://www.opbcorner.com/HWbench/tru120e/u120e/piorthos4cores/biostemp01.jpg[/img]

ok, after that, we make sure the difference is not that great like over 5 c or more, so I decide to add some voltage  to desire fist attemp overclocking and also checking in bios again to compare .
Now, you can see is under the bios monitor page, CPU core temp has been up for 35,and DT200 just for 33.9, it is definitely correct because if bios reading is from the sensor IC of CPU area, I know it's from winbond sensor ic.
[color=blue]我们基本上可以确定两者的温度差不大会达到或超过5摄氏度,于是我决定增加一点电压进行超频再来对比BIOS下的温度与DT200的不同。[/color]
[color=blue]在BIOS的监视页面内你可以看到CPU核心温度升高到了35摄氏度,而DT200显示才是33。9摄氏度,确定的说这个数值是准确的,如果BIOS读取CPU附近的IC数据,我知道那一定是从华邦IC中读取的数据。[/color]

[img]http://www.opbcorner.com/HWbench/tru120e/u120e/piorthos4cores/biostemp03.jpg[/img]

So now, we boot into windows. ALright, Samething every beginging of bench, is calling Super pi out. Now is not the time for play tweaking, so let's just be regularly bench.
here are couple pictures just totally surprising me at the moment when I took it.
[color=blue]我们启动进入WINDOWS,进行一些常规的测试,运行SUPER PI,当测试进行的时候,下面有些截图令我震惊。[/color]
Even that low 3.6g for dual super pi running at that low vcore usage, my qx6700 will have at least one error stop and need to reduce to around 8x445 so I can get pass for the 1m~8m. Honestly never dual 32m at that voltage , impossible. Certainly impossible., but look at that
dual pi 1m loaded from the asus prob II, 31c..tell me about that. and the moment immediately get the screen is 25c..wow. actually the more serious jaw dropping results are after pi running ...
here are all the prob II I catched from the moment super pi done for calculation.
[color=blue]在低VCORE主频仅有3。6G时运行双SUPER PI,我的QX6700至少会出现一个错误并停止运行,只好把频率降至8x445就可以顺利通过1M-8M的测试,老实说,在那个电压下双32M是不可能通过测试的,但在运行双PI 1M时我从ASUS PROBE II上得到的温度是25摄氏度,PI运行后温度的下降更加明显,这里有所有的运行PI时的PROBII的截图。
[/color]
[img]http://www.opbcorner.com/HWbench/tru120e/u120e/piorthos4cores/1mfidol.jpg[/img]

[img]http://www.opbcorner.com/HWbench/tru120e/u120e/piorthos4cores/1mloaded.jpg[/img]

[img]http://www.opbcorner.com/HWbench/tru120e/u120e/piorthos4cores/8mfidol.jpg[/img]

[img]http://www.opbcorner.com/HWbench/tru120e/u120e/piorthos4cores/dual32mfidol.jpg[/img]

well you might say , hey OPB, and where is the temperature you took from DT 200 during the run.?
Certainly, , Do I need to take the 1m , 8m? of course not.:D
how about the dual 32m , then you can understand that all. It just completely rulesall over.
[color=blue]这时你也许会问我,OPB,运行PI时DT200的温度在哪里呀?当然,我还需要再去运行1M或8M吗?当然不用,不如用双32M来跑跑。[/color]

I already put some notes in order not to forget what was the situation when I took the pictures
sorry about some pictures are not that clear, because I don't wanna use flash to ruin the pictures. and I will also need to took all the meters and monitors so I won't be doubt again say I am cheater:D, I also running the asus prob II for whole spi runnig.
[color=blue]当我处理图片的时候已经放了一些备注上去以免自己忘掉这是表述什么内容的图片。[/color]
[color=blue]很抱歉有一些图片不够清晰,呵呵,为了不会再被怀疑做B,我把所有的图片都放在整个测试中我也运行了ASUS PROBII。[/color]

[img]http://www.opbcorner.com/HWbench/tru120e/u120e/piorthos4cores/dual32mtemp01.jpg[/img]

[img]http://www.opbcorner.com/HWbench/tru120e/u120e/piorthos4cores/dual32mtemp02.jpg[/img]

[img]http://www.opbcorner.com/HWbench/tru120e/u120e/piorthos4cores/dual32mtemp03.jpg[/img]

[img]http://www.opbcorner.com/HWbench/tru120e/u120e/piorthos4cores/dual32mtemp04.jpg[/img]

[img]http://www.opbcorner.com/HWbench/tru120e/u120e/piorthos4cores/dual32mtemp05.jpg[/img]


so now what? anything else? How about the 4 threads Orthos for over 2 hours , from that run, I never seen over 52c  for 4 cores loaded 100, I did leave the task manager and resource chart.
Check again for idle temp
[color=blue]现在怎么样?四进程的ORTHOS运行了超过两个小时,自打开始运行,我就没见着四核百分百满载的时候超过52摄氏度,我打开资源管理器来查看处理器的负载。[/color]


[img]http://www.opbcorner.com/HWbench/tru120e/u120e/piorthos4cores/idletemp.jpg[/img]

[img]http://www.opbcorner.com/HWbench/tru120e/u120e/piorthos4cores/orthos10m.jpg[/img]

[img]http://www.opbcorner.com/HWbench/tru120e/u120e/piorthos4cores/orthos30m.jpg[/img]

[img]http://www.opbcorner.com/HWbench/tru120e/u120e/piorthos4cores/orthos01.jpg[/img]

[img]http://www.opbcorner.com/HWbench/tru120e/u120e/piorthos4cores/orthos40m.jpg[/img]

[img]http://www.opbcorner.com/HWbench/tru120e/u120e/piorthos4cores/orthos1hr.jpg[/img]



[img]http://www.opbcorner.com/HWbench/tru120e/u120e/piorthos4cores/2hr36morthos[/img]

alright, 2 hours .....sleepy..
[img]http://www.opbcorner.com/HWbench/tru120e/u120e/piorthos4cores/dual32mtemp05.jpg[/img]


So how's your feeling about this? not what I wanna seduce you guys, but this cooler just too good to give up.
Previously, I all think about Tunique Tower 120 is the one ever made that good. Well, compare with U120 extreme, Tunique tower 120's performance will not be even closed , I guaranteed you.

I will have follow up for 3d bench by using this and HR09 and 03 plus, please stay tune for further report.
[color=blue]你感觉如何?我不想诱惑你朋友,但这个散热器真的太厉害了。一开始我认为TUNIQ TOWER 120能够做到这一步,OK,与U120 EXTREME对比后,我保证,TUNIQ TOWER 120的性能无法与之靠近。[/color]

OPB

多米小熊 发表于 2007-5-5 20:45

太考验我的英文水平了啊~~~~~

大尾巴松鼠 发表于 2007-5-6 11:35

[s003] [s002] 考虑到能够方便广大朋友的阅读,我就试着翻一翻吧。希望OPB能够谅解我编辑你的贴子,翻的不好的还望指正。

Onepagebook 发表于 2007-5-6 13:50

對國內朋友..的熱情與認真實在令我動容
不得不來磕個頭...[s016]

VictorWang 发表于 2007-5-6 13:53

[s018]   狂爷总是出这种精彩review
赞!

陈言 发表于 2007-5-6 13:58

呵呵,松鼠很是热心呢,好久没有来了 [s023] [s008]

陈言 发表于 2007-5-6 14:01

刚想起来Ultra120把上支架安装伸入散热片和散热低座都很费劲,那还是4热管的,不知EXE怎么样的,其实安装并不复杂了。[s008]

大尾巴松鼠 发表于 2007-5-6 14:47

[quote]原帖由 [i]Onepagebook[/i] 于 2007-5-6 13:50 发表
對國內朋友..的熱情與認真實在令我動容
不得不來磕個頭... [/quote]

[s023] 对于狂少能够把如此精品引入到本论坛,回磕致谢~!翻的不好,还望海涵。

不过我记得好象LGA775在安装时遇有扣具的设计缺陷这个问题,是不是已经解决?何时会有HR09的美文送上呢?严重期待~!

多米小熊 发表于 2007-5-6 16:19

松鼠英语肯定是专8了~

lujie1806 发表于 2007-5-7 18:12

我喜欢啊[s008] 不知道那里有买,几米[s013]

Onepagebook 发表于 2007-5-7 18:52

我打算玩大一點....

有人拔過qx6700鐵蓋嗎?

[img]http://www.opbcorner.com/HWbench/tru120e/u120e/qxrihs.jpg[/img]

[img]http://www.opbcorner.com/HWbench/tru120e/u120e/qxrihs01.jpg[/img]

alom 发表于 2007-5-7 18:55

[s003] 难道已经拔下了不成?

Onepagebook 发表于 2007-5-7 18:57

[quote]原帖由 [i]大尾巴松鼠[/i] 于 2007-5-6 14:47 发表


对于狂少能够把如此精品引入到本论坛,回磕致谢~!翻的不好,还望海涵。

不过我记得好象LGA775在安装时遇有扣具的设计缺陷这个问题,是不是已经解决?何时会有HR09的美文送上呢?严重期待~! [/quote]
不會耶...現在很好裝

順便回復陳兄
類似此類螺絲,您可以扶住彈簧, screw driver從頂去鎖--->方法一,先鎖2圈
或是用手先鎖--->方法2,

但一定要先鎖好對角線.....才能達成balance.

Onepagebook 发表于 2007-5-7 19:00

[quote]原帖由 [i]alom[/i] 于 2007-5-7 18:55 发表
难道已经拔下了不成? [/quote]


當然啊
不然圖從何來?:D

正確之法是要用兩把"razor blade各深入8mm,個走半圈不要拔起

將CPU鐵蓋朝下,用torch烤至350c直到鐵蓋掉下
因為intel的core是用焊錫與鐵蓋溶在一起的..
硬拔只是提早讓CPU死
等溶掉後在處理die上的殘留物..如圖


刀工不佳者請勿嘗試
[img]http://www.opbcorner.com/HWbench/tru120e/u120e/qx6700rihs03.jpg[/img]

補充說明

[img]http://www.opbcorner.com/HWbench/tru120e/u120e/qx6700rihs04.jpg[/img]

[[i] 本帖最后由 Onepagebook 于 2007-5-7 19:10 编辑 [/i]]

VictorWang 发表于 2007-5-7 20:10

......你真是[s004] [s003]

逆水 发表于 2007-5-7 21:13

看不懂英文[s009]

萌萌 发表于 2007-5-7 21:40

OPB的文怎么可以不顶,虽然帖子不多,不过个个都是精华[s023]

alom 发表于 2007-5-7 21:46

[quote]原帖由 [i]Onepagebook[/i] 于 2007-5-7 19:00 发表



當然啊
不然圖從何來?:D

正確之法是要用兩把"razor blade各深入8mm,個走半圈不要拔起

將CPU鐵蓋朝下,用torch烤至350c直到鐵蓋掉下
因為intel的core是用焊錫與鐵蓋溶在一起的..
硬拔只是提早 ... [/quote]

暈死,我看的時候照片沒出來呀,狂少,你。。。。。可憐的U呀,這樣子處理後,溫度下降幾多?

萌萌 发表于 2007-5-7 22:10

QX67的壳子要是开不好,找个盆接眼泪吧。狂大真敢冒险,不过前提是那么出色的手艺[s002] 支持一下,不过请汇报究竟温度降了多少啊?

Onepagebook 发表于 2007-5-7 23:30

5~7 c左右

artkobe 发表于 2007-5-12 03:30

狂少,請問讓Ultra-120 Extreme發揮較好的效能時,用多少轉風扇比較合適?
ASUS的主板用的是HR09U吧?

gmx168 发表于 2007-9-29 20:19

这种好文不能沉了。

醉生梦死 发表于 2007-9-30 14:46

狂少的文章超强
顶一下

页: [1]

Powered by Discuz! Archiver 6.1.0  © 2001-2007 Comsenz Inc.