Intel Core i7 Heatsink Roundup Q309 @ [H]

vengence said:
Oy, now I'm confused! :p

I think the issue is these temperatures are being measured by a probe at the heat spreader which is going to be several degrees cooler than measuring the die temperature using something like Core Temp. Furthermore, it doesn't help that they were running an open motherboard test setup. I wouldn't be surprised if these were 10-15C away from what Core Temp would report on the same setup in a case.
Click to expand...

Exactly what I was thinking. Compare the numbers in this test to what Kyle got in one of the first heat test on the i7 920 using core temperature readings.
Looks a lot more like the temps we see at home using Core Temp or RealTemp.
 
StillFunkyB said:
haha, nice.... maybe he thought he was at (s)oft/forum?

I was at Microcenter a couple weeks ago and someone was grabbing a Zalman cooler and decided to confess to me how great they are and that they are the "highest rated" cooler you can get. He proceded to put this Zalman cooler in his little basket. I reached down on the bottom shelf and picked up a TRUE and placed it in his basket and said "Now you have the highest rated cooler". Guy just looked at me, although I didn't see him put it back.

Anyhoo, nice review, thanks!
Click to expand...

LOL, that's awesome.
 
Wolfkin said:
Exactly what I was thinking. Compare the numbers in this test to what Kyle got in one of the first heat test on the i7 920 using core temperature readings.
Looks a lot more like the temps we see at home using Core Temp or RealTemp.
Click to expand...

Well, keep in mind this is a heatsink performance review, not a coolest CPU performance review. I understand that some of you may have a little trouble differentiating that, but that is the way it is.

I know some of you want to see it done in a chassis, but the fact is that would be worthless to us. How can we keep specific ambient temperatures in a chassis when we are changing fans, fan direction, cooler orientation, etc inside of a specific chassis, that may or may not be relevant to your chassis at home. It simply introduces too many variables and would make the test tremendously far from objective.
 
I just don't understand the choice of Prime blend, tbh. Of the 6GB of ram that you have installed it is only using 1600MB. Not really an apple to apple comparison but with a P45, quad, and 4x2GB of ram I have been 12+ hours prime blend stable, and 40 runs of linpack stable and my MCH still not 100% gaming stable. Blend dosen't even stress the MCH too hard. Even Memtest 3.8 and to a lesser extent linpack heats up my MCH considerably more. Memtest even keeps my cores as hot as a blend test.

Although, a prime blend with all of my available ram enabled in custom heats up my MCH much more than anything and shows errors fast. HyperPi would also be a better choice.

I think that you guys should have atleast included a small fft run in the review. I just find it very hard to beleive that the stock cooler from the i7 920 would keep it operable at 3.6ghz with 1.45VCC nevermind under 73c at full load.
 
Last edited:
I have just installed the COGAGE True Spirit. I first read Kyle's article and the spec. look good so I jumped on one. After installing this cooler, I have second thoughts tho. The mounting for this nice cooler are terrible. The mounts do not hold this cooler secure at all. Just with the small vibration of the fan it moves around. In other words it rotates around the center screw without any effort. Great for seating the paste but not so good for latter on when you need to to be mounted without sloppy fitting mounts. After all is said and done and all is working you can simply just move the cooler around the center mount screw very easly. I do not like this at all. If I was to buy another mount it would not be this one.
Sorry Kyle the cooler is nice but the mounts suck. Big time. :mad::mad::eek:
 
Last edited:
crewzen said:
I have just installed the COGAGE True Spirit. I first read Kyle's article and the spec. look good so I jumped on one. After installing this cooler, I have second thoughts tho. The mounting for this nice cooler are terrible. The mounts do not hold this cooler secure at all. Just with the small vibration of the fan it moves around. Great for seating the paste but not so good for latter on when you need to to be mounted without sloppy fitting mounts. After all is said and done and all is working you can simply just move the cooler around the center mount screw very easly. I do not like this at all. If I was to buy another mount it would not be this one.
Sorry Kyle the cooler is nice but the mounts suck. Big time. :mad::mad::eek:
Click to expand...

It should not be doing that, are you dead certain the push pins where in the install position and are all fully engaged ?
 
yes the push pins are in all the way. I use a magnifier and very bright light to check afer they click down. Its in a PC-343 CASE and there is a lot of working room in there. What is going on is that the mount is not all the way down on the cooler because of the little plastic spacer they use to cover the swivel lock for the pin locations. The small center screw is in place but the cooler still moves there is nothing to stop it at all. I would take it that if I remove all the paste and mount it dry the friction would stop the movement but then there isnt as much transfer for the cooler to cpu for proper cooling. It realy needs a screw in mount badly.

PS I have had many push pins before but there has always been a positive mechanical lock somewhere to secure the cooler, but not on this COGAGE TRUE Spirit, and it is made soley for the 1366 socket. You would think it would be full proof. Just have to remember that it is made by thermalright but to COGAGE standards and not thermalright's. That may be the difference and why thermalrights name isnt on this thing.
 
Last edited:
crewzen said:
yes the push pins are in all the way. I use a magnifier and very bright light to check afer they click down. Its in a PC-343 CASE and there is a lot of working room in there. What is going on is that the mount is not all the way down on the cooler because of the little plastic spacer they use to cover the swivel lock for the pin locations. The small center screw is in place but the cooler still moves there is nothing to stop it at all. I would take it that if I remove all the paste and mount it dry the friction would stop the movement but then there isnt as much transfer for the cooler to cpu for proper cooling. It realy needs a screw in mount badly.

PS I have had many push pins before but there has always been a positive mechanical lock somewhere to secure the cooler, but not on this COGAGE TRUE Spirit, and it is made soley for the 1366 socket. You would think it would be full proof. Just have to remember that it is made by thermalright but to COGAGE standards and not thermalright's. That may be the difference and why thermalrights name isnt on this thing.
Click to expand...

Well it is made by Thermalright engineers so thats not too surprising. I would washer mod it just like a TRUE and probably pick up a bolt-through kit.

Other than the terrible base the TRUE is also famous for the same type of crappy mounting. I have to use two washes on mine.
 
crewzen said:
I have just installed the COGAGE True Spirit. I first read Kyle's article and the spec. look good so I jumped on one. After installing this cooler, I have second thoughts tho. The mounting for this nice cooler are terrible. The mounts do not hold this cooler secure at all. Just with the small vibration of the fan it moves around. In other words it rotates around the center screw without any effort. Great for seating the paste but not so good for latter on when you need to to be mounted without sloppy fitting mounts. After all is said and done and all is working you can simply just move the cooler around the center mount screw very easly. I do not like this at all. If I was to buy another mount it would not be this one.
Sorry Kyle the cooler is nice but the mounts suck. Big time. :mad::mad::eek:
Click to expand...


When I installed the COGAGE there was some movement allowable but nothing as extreme as you describe. As is the case from time to time you may have received a less than perfect unit. I would try again and rotate the push pin assembly so that the push pins are inserting through different holes and see if you get different results.

If the results are the same then an exchange is necessary. I am sorry it happened to you.
Doesn't mean it's a bad cooler just that you got a bad sample. :(
 
BababooeyHTJ said:
I just don't understand the choice of Prime blend, tbh. Of the 6GB of ram that you have installed it is only using 1600MB. Not really an apple to apple comparison but with a P45, quad, and 4x2GB of ram I have been 12+ hours prime blend stable, and 40 runs of linpack stable and my MCH still not 100% gaming stable. Blend dosen't even stress the MCH too hard. Even Memtest 3.8 and to a lesser extent linpack heats up my MCH considerably more. Memtest even keeps my cores as hot as a blend test.

Although, a prime blend with all of my available ram enabled in custom heats up my MCH much more than anything and shows errors fast. HyperPi would also be a better choice.

I think that you guys should have atleast included a small fft run in the review. I just find it very hard to beleive that the stock cooler from the i7 920 would keep it operable at 3.6ghz with 1.45VCC nevermind under 73c at full load.
Click to expand...

The choice to use blend was made because of the memory controller being moved onto the CPU. We wanted to add the extra stress of memory I/O to the mix. While it did not put as much heat into the CPU as small FFTs would, we felt as long as our testing procedures remained constant useful comparisons could still be made.

And to be fair the amount of heat produced by running small FFTs is much larger then would be seen by many reader, though your point is still taken.
 
Marc Adams said:
The choice to use blend was made because of the memory controller being moved onto the CPU. We wanted to add the extra stress of memory I/O to the mix. While it did not put as much heat into the CPU as small FFTs would, we felt as long as our testing procedures remained constant useful comparisons could still be made.

And to be fair the amount of heat produced by running small FFTs is much larger then would be seen by many reader, though your point is still taken.
Click to expand...

I agree with you which is why I never use linpack. It's just that I think that readers would also see higher temps under normal usage even while gaming. I just would have liked a test using a higher heat load since that can make a big difference in a coolers performance.

Edit: I just saw your other post and wanted to add that I wouldn't touch your testing conditions. Like you said variables are never a good thing in a heatsink performance review and you did a great job removing all possible variables.
 
just wanted to stop by here and let you guys know that i think the way you test the heatsinks is just fine. Adding extra testing methods would take too much time in my opinion for not much data. Replacing this test with other tests would take away a lot of useful info and give me little in return.
 
BababooeyHTJ said:
It's just that I think that readers would also see higher temps under normal usage even while gaming.
Click to expand...


You are quite simply wrong in your thoughts. Should you want to introduce proof otherwise, I am all ears.
 
So I read the article and took a closer look at the COGAGE... It looks identical to a TR Ultra 120... And I find myself wondering if it is any different? As I have one and would love to know that it's almost as good as a TRUE and that it would be a complete waste of money to buy something different when upgrading to 1366.
 
WintersFury said:
So I read the article and took a closer look at the COGAGE... It looks identical to a TR Ultra 120... And I find myself wondering if it is any different? As I have one and would love to know that it's almost as good as a TRUE and that it would be a complete waste of money to buy something different when upgrading to 1366.
Click to expand...
it has less heatpipes and its a little slimmer.

http://i32.tinypic.com/2jdjmyx.jpg
 
vertigomhs said:
oh my bad, it still looks to be slimmer than than the non-extreme version, and i think the cogage uses a push-pin system, not sure if the TRU120 does as well.
Click to expand...

Apparently the TRU is the same size as the TRUE, just two less heatpipes per side. So I guess in theory it should cool better than the True Spirit unless they designed it in a completely different manner.

And the TRU and TRUE use a bolt-thru kit to mount to the 1366 socket.
 
Hey guys, just built my first PC (See sig) and I installed the Noctua NH-U12P SE heatsink and I was wondering if these temps are ok. I'm currently using stock settings (i7 920 @ 2786.03 MHz)

Real Temp shows cores at 42 36 38 35. The first core is always hotter.. about 6-7 degrees.

EVGA E-LEET Tuning shows CPU at 26C, VREG at 42-44C, and System at 31-33C.

I was also wondering which orientation would be best for the Noctua cooler.. horizontally or vertically.

Sorry if these questions are dumb, i'm new at this :p

Thanks for any help!
 
jspan said:
Hey guys, just built my first PC (See sig) and I installed the Noctua NH-U12P SE heatsink and I was wondering if these temps are ok. I'm currently using stock settings (i7 920 @ 2786.03 MHz)

Real Temp shows cores at 42 36 38 35. The first core is always hotter.. about 6-7 degrees.

EVGA E-LEET Tuning shows CPU at 26C, VREG at 42-44C, and System at 31-33C.

I was also wondering which orientation would be best for the Noctua cooler.. horizontally or vertically.

Sorry if these questions are dumb, i'm new at this :p

Thanks for any help!
Click to expand...

The temperature sensors work like bonkers at low temps...the are only "semi"-accurate when close to the thermal maximum...people should stop thinking that their low temps readings are accurate.
 
Kyle, True Spirit, Shin-Etsu7783-$59 delivered & (installed it with old TRUE 1366 bolt thru bracket, not the push-pin 1 included), looks like a TRUE Mini-me, removed old TRUE with OCZ freeze, using same Delta fan at 2280rpm (goes up to 3125rpm). Prime95 blend test with Cpuz 1.52 & Real temp 3.0. Core i7-965 stock at 3207x24x133 at 1.216volts with Ram at 1600, 1.66v, 8,8,8,24,88,1T. True min.45,45,43,44 max. 73,73,71,73. Spirit min.40,39,38,38 max. 63,63,61,62. Can't believe it but that's what I got, granted ambient is cooler today but for a cooler this size that's pretty amazing. Fits on P6T6 with12 Gb Corsair XMS ram with no problems could even use a 120x38mm fan w/o blocking ram slots. If I crank up Delta fan I lower temps by 1c to 2c. Excellent stuff glad you tested it I'll be using this cooler from now on.
 
Last edited:
Kyle, Do you have some quick voltage #s I could try to OC the i7-965, my multiplier is unlocked so that's not a problem, but what voltages should I increase to get a postable OC. Then I'll repost my True Spirit temps.
 
This looks pretty good, but the fairly new Coolermaster Hyper 212 Plus seems to be the best value in my opinion.

It can be found here for Canadians:
http://ncix.com/products/index.php?sku=41337&vpn=RR-B10-212P-GP&manufacture=COOLERMASTER

Here for Americans:
http://www.newegg.com/Product/Product.aspx?Item=N82E16835103065&Tpk=RR-B10-212P-GP

It fits all sockets, including the core i5 socket that is about to come out. It's ridiculously cheap and comes with all necessary gear to mount it. I wish this was out when I bought my TRUE-120.

Hardware Canucks has a review of it on a Q6600 and an I7 920 against all the main contenders (TRUE-120, Megahalems, Noctua, HDT-1283, OCZ Vendetta). It placed near the top on all of them and can be price matched for $25 Canadian.
 
Would it be possible to do an ifx-14 on the i7 or wait to compare it and the Noctua look-a-like cooler (NP-14 ?) that supposedly is going to be out soon?

I know the cooler has been done once before, but I keep hearing good things about it and other than xbitlabs.com no one else seems to have testing it on the i7. Plus there is much dispute over that article.
 
I've a query about the noise levels that were reported for the TRUE-120, did the TRUE-120 test cooler have one of the new 120mm Ultra low noise “Fluid Dynamic Bearing” fans fitted?
 
Dantrax said:
Kyle, True Spirit, Shin-Etsu7783-$59 delivered & installed it with old TRUE 1366 bracket, looks like a TRUE Mini-me, removed old TRUE with OCZ freeze, using same Delta fan at 2280rpm (goes up to 3125rpm). Prime95 blend test with Cpuz 1.52 & Real temp 3.0. Core i7-965 stock at 3207x24x133 at 1.216volts with Ram at 1600, 1.66v, 8,8,8,24,88,1T. True min.45,45,43,44 max. 73,73,71,73. Spirit min.40,39,38,38 max. 63,63,61,62. Can't believe it but that's what I got, granted ambient is cooler today but for a cooler this size that's pretty amazing. Fits on P6T6 with12 Gb Corsair XMS ram with no problems could even use a 120x38mm fan w/o blocking ram slots. If I crank up Delta fan I lower temps by 1c to 2c. Excellent stuff glad you tested it I'll be using this cooler from now on.
Click to expand...

At stock speeds Spirit beat True. So I overclocked my cpu with A20, T/M & C/state disabled all else was enabled. 1st test 133 bclk x 29 cpu multi = 3875MHz with cpu 1.35V, Qpi 1.35V, ram 1.66V I had Temps Min. 40,39,38,38 Temps max. 81,81,78,80. Next test, cpu settings same as 1st, 139bclk x 27 cpu multi = 3761 MHz cpu, with 1672MHz dram, 3344MHz Uncore, 6689GT/s, at 1.33125V cpu, 1.325V qpi & 1.66V dram I had temps Min. 41,40,38,39 & Max. 79,79,76,77. PC stable running Prime95 99.9%-100% load for 35 min. Now I didn't OC my True to compare #s, but what is your opinion of those temps ?
 
I ran the True with Noctua PF fan installed & it's silent when compared to SilentX or Thermalright tr-fdb fans. Noctua moves as much air as others at same speed 1300rpm. But True likes lots of air to cool off & high rpm = high noise. I have a tr-fdb-12-2000 on rear of my case & I'm glad I have it on a speed controller.
 
HELP! Some one on this thread has to have their old i7 stock CPU cooler around. I modded the fan of my stock CPU cooler into a great memory cooler. I love it so much I started sleeving it today. I forgot to write down the wire order before I slipped the 4 pins out. I can't run my current settings without this fan on my memory. Please tell me the order and from which side you are looking, or post a pic if it's easier.
 
Brock10 said:
HELP! Some one on this thread has to have their old i7 stock CPU cooler around. I modded the fan of my stock CPU cooler into a great memory cooler. I love it so much I started sleeving it today. I forgot to write down the wire order before I slipped the 4 pins out. I can't run my current settings without this fan on my memory. Please tell me the order and from which side you are looking, or post a pic if it's easier.
Click to expand...

That's a good idea, If you are using 3 wires it's [ Black, Yellow, Green ] then the Blue pwm wire. The [ & ] indicate the "3 wire" fan guide tab lock on the plug. That's with an i7-965 fan. With 3 wire fan guide tab lock on top of receptacle, (looking at the holes you will plug the wires into) that's BK on right side, YL, GR, then BL on left side. Do you have any details of how you hooked the fan to the ram?
 
Last edited:
Dantrax said:
That's a good idea, If you are using 3 wires it's [ Black, Yellow, Green ] then the Blue pwm wire. The [ & ] indicate the "3 wire" fan guide tab lock on the plug. That's with an i7-965 fan. With 3 wire fan guide tab lock on top of receptacle, (looking at the holes you will plug the wires into) that's BK on right side, YL, GR, then BL on left side. Do you have any details of how you hooked the fan to the ram?
Click to expand...

Thanks! I cut off the leg attachments but left some stubs. Those stubs fall on both sides of the memory. One stub sits on the NB heat sink. My Rampage II Gene is notorious for blazing NB so I wanted to cool them both. This is a new build. When I settle on a configuration I'll attach it more securely.
 
vengence said:
Any chance of you guys picking up a dial indicator to measure flatness in the future? Also, how did you guys measure thermal paste thickness to ensure each heat sink was had the same thickness?
Click to expand...

How would you use a dial indicator to measure the flatness on a heatsink? How could you easily make sure the heatsink is sitting flat to get a true reading, like an Ultra 120 for example with its heatpipes poking out of the top? Good idea, hard to pull off in my opinion.

Nice review, Megahalem coming my way because of all of the reviews you guys do.
 
Just out of curiosity, what fans were used on the TRUE and Megahalems prior to the apples vs apples test?
 
got the cogage true spirit on my 920 @ 3.6 and the temps are about the same as [H] supplied in the review
 
I got the Cogage True Spirit on sale from crazypc.com for $37 and the Thermalright 1156 bolt thru kit. Since the bases appear the same, it mounted just fine and the temps are great.

Price is great. I have no comment on the push pins since they were for 1366. I really hate push pins, so I would have grabbed the Thermalright 1366 bolt thru kit if I had gone with the 1366 i7s.
 
jimbean21 said:
My Cogage just arrived today, couldn't wait to install it, got it in and going from intel stock heatsink to this is a huge improvement (obviously) I used AS5 so there's probably a few more degrees to be had but I'm idling at around 38-40 C per core w/ the fan on full @ Turbo speeds (which keeps the vcore @ 1.25 @ 3.06 Ghz automatically) no speedstep / c1e. Doesn't load over 57C per core. All in all happy so far, when the AS5 is cured we'll see where I'm at w/ the same settings.
Click to expand...

Did you yank the mobo out of the box to install it or did you install in-place?

otapnam said:
got the cogage true spirit on my 920 @ 3.6 and the temps are about the same as [H] supplied in the review
Click to expand...

Ok, I'm ordering this. I've been told that I need to OC my 920 above 3.0 to get the best out of the 5970 I got. If I can get 3.6 with a $40 heatsink, I'm in.
 
So, I ordered a Cocage TRUE Spirit today, which should hopefully arrive next week. I was just wondering whether I should consider lapping it before install? The review mentioned that it wasn't polished, but it seemed to perform excellently anyway.
 
You must log in or register to reply here.
Back
Top