how-to block ads
|
|
Share Topic
 |
 |
|
|
|
 koitsuPremium,MVM
join:2002-07-16
Mountain View, CA
kudos:14
1 edit | reply to DKS
Re: [My Review] Adventures in Cooling - Corsair H50 Sorry, I should explain myself. There's absolutely no bragging going on here -- I happen to be one of those people who focuses strongly on temperatures, noise, and power draw (in that order). I don't overclock anything I own/use/manage, so that's something to keep in mind as well.
I've never built a water-cooled system before because I'm one of those old dudes who believes putting water near electrical components is an accident waiting to happen[1] -- but, I still occasionally check water cooling systems out because it's very interesting to see the results (cooling-wise).
Water cooling is, generally speaking, expensive. Meaning, the cost of the blocks, tubes, pump, clamps, etc. usually comes to a few hundred USD. The main reason people go with water cooling is to decrease CPU (and GPU, if they have a waterblock on their card) temperatures and sometimes decrease noise (depending on how loud the pump is). The intention, usually, of decreasing CPU temperatures is to provide room/capability for overclocking.
So when I see someone trying out water cooling and sharing their experience -- with screenshots of their temperatures -- my first assumption is to compare what they have to an air-cooled system.
All I'm saying is that I'd expect to see absolutely amazing temperatures, but my Q9550 has lower temps than yours at idle and load, yet I have very few case fans (all running at about 1200rpm) and a single CPU fan (running at between 1200-1500rpm). I'm not flexing my nuts -- I'm simply concerned that your water block may be not be mounted flush against the CPU or something along those lines.
The link I provided above basically says that on an i7 860 -- which runs significantly hotter than your Q9650 and my Q9550 by a longshot -- their idle temperatures are lower than what you're seeing, while their load temperatures are about the same. There's a couple other reviews I skimmed through which show similar behaviour, and one which basically shows there's no temperature improvement but there is an improvement (decrease) in system noise compared to the Intel stock fan (which we all know is absurdly loud, and thus works pretty good as a universal reference point).
Please don't take what I've said to mean "I'm here to piss in your cornflakes". Please take what I said to mean "Was your Q9650, prior to the H50 installation, really idling at 55-62C and 67-71C under load w/out overclocking? If so, maybe we should try to figure out why that is, as that seems awfully high compared to my Q9550 system".
My system consists of the following:
- CPU: Intel Q9550
- MB: Asus P5Q SE
- GPU: BFG nVidia 9800GT ("Low Power" model; model number BFGE981024GTGE)
- Disk: Western Digital WD5001AALS (internal temp 40C according to SMART)
- Case: Antec P182
- CPU HSF: Nexus LXM-8200 with fan powered off 5V header on mainboard, with system BIOS set to adjust fan speed dynamically (approx. 900-1200rpm)
- Fans: Case is using stock fans running at medium speed setting (maybe 1500rpm, not sure)
- Fans: Nexus D12SL-12, located in middle bay, powered via 12V header
- Ambient room temperature is about 76F / 24.4C
Again -- not bragging. This is a very simple air-cooled workstation that's running 24x7x365 in my bedroom. I can take external temperatures using a cheap external IR thermometer if you'd like.
I'm just concerned that something may not be connected quite right or flush with a surface. Quite possibly the temperatures you have are spot on -- I don't know. They just seem high for a water-cooled system, you know? Do you know what your ambient room temperature is? Maybe it's quite high and that would explain the overall difference. Again, I don't know.
[1]: A roommate of mine, many years ago (would have been in 2001) lost literally everything due to a well-known problem with the tubes the water cooling manufacturer provided gradually cracking after ~6 months of use. I forget who the manufacturer was, but it was a very well-known problem. And when I say everything, I do mean everything -- even portions of his hard disk PCB were blown. Not good.
--
Making life hard for others since 1977.
I speak for myself and not my employer/affiliates of my employer. | |
 pnjunctionTeksavvy ExtremePremium
join:2008-01-24
Toronto, ON
kudos:1 | My i7 860 only gets to about 55c under load with my zalman cnps10x extreme. I will say though that it was only slightly cheaper than the h50 (at the stores here).
One advantage I can see of the h50 is that it doesn't take up alot of room. I can't fill my second RAM slots and probably couldn't use my top PCIe 1x slot with this huge cooler, but might be able to with the h50.
To me there are downsides though, I wasn't comfortable with the system in terms of reliability and durability. The fan on a heatsink/fan solution is the only moving part and can be replaced. When the pump in the h50 dies its finished. I think when I go watercooling I'll go all out with a loop that includes the GPU and a dual 120mm radiator. | | |
|
 pflogBueller? Bueller?Premium,MVM
join:2001-09-01
El Dorado Hills, CA
kudos:3 | reply to koitsu
Just for comparison, my Q9550, which is overclocked to 3.8 GHz with a Thermalright Ultra 120 Extreme.
At load it gets to maybe 52-54 or so, I'm quite happy with this cooler.
--
The poster formerly known as deblin... | |
 DKSDamn Kidney StonesPremium,ExMod 2002
join:2001-03-22
Owen Sound, ON
kudos:2
 ·Bell Sympatico
| reply to koitsu
said by koitsu:Sorry, I should explain myself. There's absolutely no bragging going on here -- I happen to be one of those people who focuses strongly on temperatures, noise, and power draw (in that order). I don't overclock anything I own/use/manage, so that's something to keep in mind as well. Thank you for the explanation. I do not share those priorities.
I've never built a water-cooled system before because I'm one of those old dudes who believes putting water near electrical components is an accident waiting to happen[1] -- but, I still occasionally check water cooling systems out because it's very interesting to see the results (cooling-wise). This is the first time I have done something like this.
Water cooling is, generally speaking, expensive. Meaning, the cost of the blocks, tubes, pump, clamps, etc. usually comes to a few hundred USD. The main reason people go with water cooling is to decrease CPU (and GPU, if they have a waterblock on their card) temperatures and sometimes decrease noise (depending on how loud the pump is). The intention, usually, of decreasing CPU temperatures is to provide room/capability for overclocking. It can be expensive. It all depends.
So when I see someone trying out water cooling and sharing their experience -- with screenshots of their temperatures -- my first assumption is to compare what they have to an air-cooled system. That is like comparing apples and oranges. Especially considering the variables involved.
All I'm saying is that I'd expect to see absolutely amazing temperatures, but my Q9550 has lower temps than yours at idle and load, yet I have very few case fans (all running at about 1200rpm) and a single CPU fan (running at between 1200-1500rpm). I'm not flexing my nuts -- I'm simply concerned that your water block may be not be mounted flush against the CPU or something along those lines.
The link I provided above basically says that on an i7 860 -- which runs significantly hotter than your Q9650 and my Q9550 by a long shot -- their idle temperatures are lower than what you're seeing, while their load temperatures are about the same. There's a couple other reviews I skimmed through which show similar behaviour, and one which basically shows there's no temperature improvement but there is an improvement (decrease) in system noise compared to the Intel stock fan (which we all know is absurdly loud, and thus works pretty good as a universal reference point). Again, there are variables here which make the comparison somewhat invalid.
Please don't take what I've said to mean "I'm here to piss in your cornflakes". Please take what I said to mean "Was your Q9650, prior to the H50 installation, really idling at 55-62C and 67-71C under load w/out overclocking? If so, maybe we should try to figure out why that is, as that seems awfully high compared to my Q9550 system". On air, yes. The former cooler was an Arctic 7 Pro, not stock Intel. Much quieter than stock Intel, too.
My system consists of the following: - CPU: Intel Q9550- MB: Asus P5Q SE- GPU: BFG nVidia 9800GT ("Low Power" model; model number BFGE981024GTGE) - Disk: Western Digital WD5001AALS (internal temp 40C according to SMART) - Case: Antec P182- CPU HSF: Nexus LXM-8200 with fan powered off 5V header on mainboard, with system BIOS set to adjust fan speed dynamically (approx. 900-1200rpm) - Fans: Case is using stock fans running at medium speed setting (maybe 1500rpm, not sure) - Fans: Nexus D12SL-12, located in middle bay, powered via 12V header - Ambient room temperature is about 76F / 24.4C
QuadCore Intel Core 2 Quad Q9650, 3200 MHz (9 x 356)* New (replaced Q6600)
Asus P5Q Pro
2048 MB (Kingston HyperX KHX8500D2/512 x 4)
NVIDIA GeForce GTX 470 (1280 MB)
Samsung SyncMaster 940BF/MagicSyncMaster CX903B/CX916B (Digital) [19" LCD] Monitor
Samsung SyncMaster T220HD/T220MD (Digital) [22" LCD] (HVGS100165)
Auzentech X-Fi Prelude 7.1 Sound Card
Hitachi HDP725050GLA360 ATA Device (500 GB, 7200 RPM, SATA-II)
Hitachi HDP725050GLA360 ATA Device (500 GB, 7200 RPM, SATA-II)
OCZ AGILITY ATA Device (29 GB, IDE)
Optical Drive HL-DT-ST DVDRAM GH22LS50 ATA Device
SONY BD-ROM BDU-X10S ATA Device
TSSTcorp CD/DVDW SH-S183L ATA Device
Mouse Microsoft USB Trackball Explorer (IntelliPoint)
Network Adapter Intel(R) Gigabit CT Desktop Adapter
Wrapped in a new Thermaltake V3 case and the CPU cooled by a Corsair H50 CPU cooler.
Current fans are the Corsair cooler, blowing into the case, as per installation spec, and 1 x 120mm blowing out of the case on top. There is also additional venting on the top, allowing for increased air flow. The Corsair NX850 power supply is mounted on the bottom and inverted, so it takes cooler air from the bottom of the case and kicks it out through the bottom. The rummer feet are double thickness to allow more air to be evacuated. The more than adequate venting on the top plus the 120mm fan do allow a very positive air flow from the CPU cooler.
The case itself is a mid-tower, which is a more confined space than a full tower. One would expect temps to be higher in that situation. In addition, it is surrounded on four of six sides by the computer desk, with three inches of clearance on top, four on the left side and one on the bottom. This also serves to increase case temps, as opposed to an open, freestanding case.
Again -- not bragging. This is a very simple air-cooled workstation that's running 24x7x365 in my bedroom. I can take external temperatures using a cheap external IR thermometer if you'd like.
I'm just concerned that something may not be connected quite right or flush with a surface. Quite possibly the temperatures you have are spot on -- I don't know. They just seem high for a water-cooled system, you know? Do you know what your ambient room temperature is? Maybe it's quite high and that would explain the overall difference. Again, I don't know. This is a basement room in an air-conditioned house. The temps are -+20C. The case, however, is surrounded by the desk, as will be shown in the next post.
--
Need-based health care not greed-based health care. | |
DKSDamn Kidney StonesPremium,ExMod 2002
join:2001-03-22
Owen Sound, ON
kudos:2 Reviews:
·Bell Sympatico
| reply to koitsu
Case Environment | 
Case Interior |
| |
 nerdburgPremium
join:2009-08-20
Schuylkill Haven, PA
kudos:1 | Oh man, get that bad boy out of the cabinet or at least take the back off in that section.
--
-meh | |
DKSDamn Kidney StonesPremium,ExMod 2002
join:2001-03-22
Owen Sound, ON
kudos:2 Reviews:
·Bell Sympatico
1 edit | said by nerdburg:Oh man, get that bad boy out of the cabinet or at least take the back off in that section. There is no "back". That's actually a painted wall that is about 10" behind the desk.
--
Need-based health care not greed-based health care. | |
 Ray422Premium,MVM
join:2002-03-04
Lost
kudos:11
 ·RoadRunner Cable
| reply to DKS
said by DKS 
The case, however, is surrounded by the desk, as will be shown in the next post. [/BQUOTE :
We all wondered why your idle temps were high. A picture is worth a thousand words. --
Things I like - Team Discovery - Quads©3.3 to 4.1ghz - Strawberry Ice cream
How's That Hope & Change Working For Ya?
| |
DKSDamn Kidney StonesPremium,ExMod 2002
join:2001-03-22
Owen Sound, ON
kudos:2 Reviews:
·Bell Sympatico
| said by Ray422:said by DKS
The case, however, is surrounded by the desk, as will be shown in the next post. [/BQUOTE :
We all wondered why your idle temps were high. A picture is worth a thousand words. "We all"? What is with this "idle temp" stuff? A picture of idle temps. The ones to looks are the minimums. They are quite comparable with the 9550 temps posted above. --
Need-based health care not greed-based health care. | |
pflogBueller? Bueller?Premium,MVM
join:2001-09-01
El Dorado Hills, CA
kudos:3 | said by DKS:They are quite comparable with the 9550 temps posted above. They are about the same as mine, true, but my ambient is 2C hotter and my chip is running at 3.8 GHz not 3 GHz.
When you OC'd it, did you change the Vcore? If it's set to auto or set too high, it can definitely cause higher temps.
I know that on this CPU for example when I use "auto" for the Vcore my temps are 6-7 C hotter.
--
The poster formerly known as deblin... | |
DKSDamn Kidney StonesPremium,ExMod 2002
join:2001-03-22
Owen Sound, ON
kudos:2 Reviews:
·Bell Sympatico
| said by pflog:When you OC'd it, did you change the Vcore? If it's set to auto or set too high, it can definitely cause higher temps. I know that on this CPU for example when I use "auto" for the Vcore my temps are 6-7 C hotter. No. I didn't touch the VCore.
--
Need-based health care not greed-based health care. | |
pflogBueller? Bueller?Premium,MVM
join:2001-09-01
El Dorado Hills, CA
kudos:3 | said by DKS:No. I didn't touch the VCore. I definitely think you can get it down quite a bit if you lower the Vcore.  Start with a level recommended online (can't recall for the Q9550) then if it's not stable, increase it one interval and rinse repeat until it's stable.
--
The poster formerly known as deblin... | |
pflogBueller? Bueller?Premium,MVM
join:2001-09-01
El Dorado Hills, CA
kudos:3 | reply to DKS
Did you get a chance to try changing the Vcore to lower your temps? I guess if it's still within spec even under full load and you're not getting TOO hot, it's ok to leave the Vcore auto...But I think you're better off getting that Vcore lowered!
--
The poster formerly known as deblin... | |
|
