Coolit or Corsair? The $100 question.....
I just built my first overclock system. ASUS P6t WS Pro, 920 2.66GHz bumped to 4GHz, GSkill 6G triple channel, GeForce8400GS (not a gamer...) Antec 650 PSU and 4 80mm fans (2 in, 2 out). I used a Titan cooler. The sink is a good design, but the fan is a POS! I can move more air flapping my arms! Anywhoo, I took an hour or so and lapped and polished the sink to a shine you could see your face in and used Arctic Silver. My temps are running 50C-55C at idle! I've tried core voltages from 1.3 to 1.6 with no reduction in temp. I know I need a beeter fan on the sink, but I'm going to go liquid cooled. I want opinions from people who have installed a Coolit or Corsair. they're both $80 anfd I don't want to spend much more than that. Thanks! Ski
The reason I'm leaning towards LC is that I don't want some mongloid air cooler stressing my MOB. I do have some TEC material I picked up just because I might use it for a project. ( had it about 5 or 6 years now lol) Maybe I'll see if I can integrate it into my Titan with a better fan. Still would like to hear from anyone using either system and their temps and speed. Ski
<hangs head in shame> No sir, it's not..... It's an old case. I did however pick out a new one the other day and was planning on picking it up since the Coolit won't fit my current case without mods. I don't remember if it's a 90 or 120mm case. Also, I should give some more info on the temp issue. The onboard ASUS monitor says the CPU temp is 32C and the MOB is 39C. Realtemp shows the core temps at 49C and that was after I clocked it down to 3.8GHz, 1.3v on the core and 1.6v on the DRAM. I know there is a difference between CPU temp and core temp, but that much difference? Something seems skewed there.... Oh, and I can put my finger on the heat sink and hold it there with no problem. I do know that the sink is making good contact with the CPU as it had a good paste pattern when I pulled it off to lap and polish it. Maybe the whole cooler is a POS and not just the fan.
skishop69 said:Afterthought... I would also like some info on your sym spped, temps and type of air cooler you're using. Maybe I'll change my mind......
My I7-860 is clocked at 3.8 and idles around 6c above ambient ( 21c ) under an H50 with push/pull fans, loaded the highest core hits 66c max.
Here is a direct comparison between the H50 and NH-D14:
If you plan on running at 4.0 Ghz or above your have two choices:
1: High end air (NH-D14, Megahelams, etc.)
2: Custom water cooling
Ok, let me make some thing clear:
1. The CoolIt, H50, and other like it are NOT TRUE WATERCOOLING! There is a HUGE difference between custom/kit water cooling an a all in one design like the H50.
2. Both are out performed by a good air cooler.
3. Case airflow makes a HUGE difference.
4.Quote:I've tried core voltages from 1.3 to 1.6 with no reduction in temp.
You should NOT put more than 1.45v for the 45nm CPUs, even less for the 32nm CPUs.
1. Have you tried better fans (ie Scythe Slipstream high CFM (80+))?
2. What case? I think you may be better off spending on a good case like a HAF 922/932.
3. Your vCore is probably WAY too high if it's anywhere above 1.35v. Is HT on or Off? How about Turbo? What VID?
4. GOOD water cooling will cost you at least $220-250 just for the CPU.
I've read the links, and I have to say, I have issues with one persons test without all of their system information, no baseline, specific thermalpaste or other such info. I'm not saying I don't believe it, but what I've read in reviews and seen in actual videos of tests, the H50 performs with (and sometimes out) the magehalem (sp). Not knocking the Noctua at all, but I am loathe to have it hanging off my processore. I also am getting info from others that their temps are very reasonable with the H50. Again, grain of salt since I don't have all their info. Yes, I know they are not 'true water coolers'. I have no need for that as I don't game. I convert and process video. Stll leaning towards liquid....
Here are some reviews for the H50:
Overall, it is more or less the same as high-mid high end air cooling.
skishop69 said:I've read the links, and I have to say, I have issues with one persons test without all of their system information, no baseline, specific thermalpaste or other such info. I'm not saying I don't believe it, but what I've read in reviews and seen in actual videos of tests, the H50 performs with (and sometimes out) the magehalem (sp).
See, the thing you aren't taking into consideration is that a reviewer is given the product to review and wants to continue receiving products to review. So don't you think he will try to be as favorable as possible to the product in question. The people who have to buy there own products and share their results have nothing to gain one way or another, hence, honesty is a given.
It sounds like you have made up your mind anyway, so I wish you luck with your system. Just know that video encoding is fairly cpu intensive (I'm sure you already know) and will strain your cooling choice and will have an impact on your OC and as a direct result the time said encoding will take you.
Another H50 review. With actually using the ever more popular psh pull fans.
I was a H50 hater being a big monster watercooler setup guy. But for what it is, it's pretty good.
The coolit has never been worth the money. Poor, just never was ever worth the money.
Ok, I just put everything into an Antec 300 tower with 120mm fan and a top mount 140mm fan. I took the piece of crap fan off the Titan and replaced it with a better (not best, but I'll pick one up) 90mm fan. The temps are now down to 44C at idle at 4GHZ and 1.4vcore (actually reads 1.34 in BIOS). That's 13C cooler than I was at 3.8GHz. I know I still need to run prime95, but not til I get the temp down a little farther. I still have a lot of tweeking to do to get it to stabilize. OS loading is a little slow.
As for making up my mind, one thing I do is a lot of research before I jump into something (not counting that d*mn Titan. It was a freebie...). The big thing that gets me from the above mentioned overclock forum on the H50, is that he never says he changed the thermal paste. The majority of the reviews I read, they wiped off the factory garbage (lets face it, manufacturers have to save costs somewhere) and used Arctic Silver and had anywhere from a 7-12C drop. I do value the opinions of those who have been at this and know their sh*t. I say this with respect RJR; (and not to start a pissing contest) I have read some of your posts and you appear to be a tad biased towards water cooling. Maybe I'm reading into that. If I had the extra money and was going to really push this system, I'd go with it. Almost 1/2 of the reviews I read were done by people who forked out their own money for the product. True there are some who will skew their findings to keep getting their shiny new toys, but you just have to try to take all of it and come up with a (hopefully) intelligent decision. After more research, I have found that Coolit does have issues of leaking, showing up dead and dropping dead in short order. It's out....
And speaking of liquid cooling, the other thing that got me about the overclockers H50 test was that somewhere along the line, someone piped in about the thermal dynamics of water cooling. There were several incorrect statements made. I may not know PC cooling, but liquid cooling I know a bit about. I'm a driveability technician and I build high performance motors on the side. The size of the reservoir has nothing to do with the cooling capability. It does affect how long the system takes to reach thermal equilibrium taking into consideration the system is 'sized' right for the application. There are four main factors in liquid cooling:
1) 'Size' of the radiator... This doesn't necessarily mean overall dimension, although it is taken into account, but does include material and number of cores or passes. Too small or too few cores; overheat. Too big/ too may cores; too long to reach optimum operating temp. Not to mention a waste of money, because at a point, the system reaches equilibrium and can't cool anymore efficiently than it is.
2) Airflow... Too little; back to overheat. Too much; believe it or not, overheat. The latter only applies to colder ambient air temps. It is harder to transfer heat to colder air molecules than to warmer air molecules. The ones that are 'napping' don't want to wake up vs the ones who are up 'cleaning the house'. lol That also depends on the speed not CFM of the airflow through the radiator. Too fast and you can't transfer the heat. The window for correct airflow is pretty large so for real world applications, you don't have to worry about the being too cold issue. Coolant flow also plays into it. Which brings me to...
3) Coolant flow... Too fast; can't effectively transfer heat from the 'engine'. Too slow; It picks up so much heat that it can't dissipate it through he radiator. Two ways to control flow: Thermostat and pump speed/volume.
4) Type of coolant used... There are many different types of coolant and each one is better suited for its application. You wouldn't use liquid sodium in your car and you certainly wouldn't use ethylene or propylene glycol to cool a nuclear power plant! you have to go with the recommendation of the engineers and hope they have it right.
I apologize for my verbal regurgitation. It's been a long week and I'm really fragged. Didn't mean to go on like that. One other thing before I go eat... There is a 10C variance between my ASUS CPU temp and my Realtemp core temps. I know the core temps should be higher, but by roughly how much?
Thanks all for your input. Keep it coming, I haven't decided yet though I am leaning.
Good input. Keep at it. You have the technical ability to build a solid real WC loop and bypass the crap stuff out there.
But don't compare car systems much with PC cooling, even tho there are mahy similarities.
You and your knowledge need to move to more informed and better places.
Look at the sticky I wrote at the top of this forum, read.
My home forum is OC Forums, google it. I have the same user name. Start reading MANY posts and all the FAQS we have.
Never trust the Mobo software. they don't spend much on that in development and they don't use the right sensor info.
Realtemp is aceppted as solid for Intel folks. You can calibrate it.
I thought about building a WC loop, and I just might. I checked out the two TEP modules I have and it just so happens they are the same size as the Titan base! I'm just getting ready to hook them up here at the shop and see what the draw is and the lowest temp when active. If the draw and temp are adequate, I'll find a way to mount it to the cooler base and attach it to one of the molex connectors. I read your sticky before I posted . It had good info. Even printed some of it out. So, Realtemp is pretty much the word of God when it comes to core temps? I know you can adjust it, but how would you know which way. I remember reading something on that subject earlier this week, but now I can't find where. I'll post back with the TEC results.....
So I spent some time at Frye's today and decided to go with the Cooler Master Hyper N520. It has decent ratings just by itself, but I chose it because the top side of the heat sink is flat and smooth and big enough to accommodate the TEC and heat sink for it. I've got the fans off and it looks like I can get the TEC and sink mounted without modifying or removing the fins. I'll take some pics as I go and then post when I'm done. After that, I'll run Prime 95 without the TEC active and then again with it active to see if all this sick, twisted fun of fabrication and modifying is worth it.
skishop69 said:The big thing that gets me from the above mentioned overclock forum on the H50, is that he never says he changed the thermal paste. The majority of the reviews I read, they wiped off the factory garbage (lets face it, manufacturers have to save costs somewhere) and used Arctic Silver and had anywhere from a 7-12C drop.skishop69 said:I say this with respect RJR; (and not to start a pissing contest) I have read some of your posts and you appear to be a tad biased towards water cooling.
Now that makes more sense to me. I didn't know they were using Shin Etsu. I didn't see that fact anywhere. The facts you have on the quad vs. i7 platform are a good argument and I did fail to take into consideration that the heat output was that much higher. I still haven't come to a decision yet, but I did just finish making my hybrid air/TEC cooler. Took a couple of hours to build and I had to run after some hardware. Seems I didn't have any 6-32 screws in the shop. I'm going to shut down after this and mount it up. I took pics, but it looks like I need to post them somewhere else and provide a link, so if anyone is interested, just PM me and I'll send them.
Well, this has been uneventfull..... I tested it before I hooked it up to be sure it was wired right and functional. I actually had a 2C increase at idle but a 2C drop at full load. 42C idle and 65C full load. Granted they say that AS has to 'cure' through three or four heat up/cool down cycles, so we'll see what happens....
I had some time to think about it and realized I had sized the heat sink for the TEC too small. Luckily, I have plenty lying around so I'll be able to modify one to fit. I hadn't taken into account the heat generated at the CPU sink added to the heat generated on the backside of the TEC. I don't know why I didn't think of this since the TEC is acting as a heat pump. Another one of my Polish moments....
Yep, thats a issue. Thats why TECs went out of favor. Even with a massive WC loop most TECs can't pull enough heat to make them even worth it. Back when chip were only 100 watts max, they worked. the technology isn't there for newer chips.
Folks that want sub ambient went to chilled water loop, phase change setups, DICE and Ln2.
It gets really expensive and complicated. Gotta insulate the mobo and socket. Never be able to RMA the board.
Take a big breath and dig in here. Lots of info.
It's ironic that my cheapy Titan cooled better (after I polished it) stock than the Coolermaster with the mods. I actually have two of the same TECs, and it seems I also miscalculated their wattage. I believe I will see what I can do about mounting both of them to the base of the Titan since it's already out and I can pull the installed TEC without removing the cooler. Combine, I should get a 4 degree drop which would put me at 35C. I guess now it's just more of a morbid curiosity then anything.