Air cooler shootout.

Tonight, I'm going to make a little shootout between the following coolers:

Thermalright 120 extreme lapped and penny modded (cost: $60 for heatsink LGA775, $10 for bracket 1366 and $10 for lapping kit = $80; no fan included; it should be said that it can be had unlapped with the new fan and bracket for $70 but i had it from my LGA775 still, so i needed to buy a bracket and lapping kit)

Xigmatek Dark Knight (cost: $39 heatsink with LGA775 and 1366 bracket, fan included but loud and obnoxious)

Scythe Mugen 2 (cost: $37 heatsink with unversal bracket for everyhting, fan included, very quiet)

All will be mounted with the same paste (Noctua) and temps will be measured with the same tool. Also, each of the heatsinks will be equipped with the same fan. I'll remount and apply paste 3 times for each cooler and take measurements each time to eliminate the wrong application of thermal paste. Posted will be idle temps and Prime95 temps after 20 minutes.
Of course i will let my computer idle after each change of cooler to achieve the same case temp before taking any measurements.

Processor used is a Core i7 overclocked to 3.8GHZ. I won't include the stock cooler nor will i post temps without overclocking as this is the [H].

We'll see what happens.

PS: There will be nothing scientific about it. I won't post screenies or anything, just the bare temps. Believe it or not, up to you. I just thought it would be nice for some people looking to buy an I7 cooler to see which of these popular models perform in relation to each other.

First test is done.

Room Temp 24C

Xigmatek Dark Knight cooler scored

Idle test 1
Cores 1,2,3,4: 46/44/47/43

Prime95 20 minute torture test 1
Cores 1,2,3,4: 79/78/80/78

Gaming (Left4Dead) 20 minutes test 1
Cores 1,2,3,4: 68/65/68/64


Idle test 2
Cores 1,2,3,4: 51/48/52/47

Prime95 20 minute torture test 2
Cores 1,2,3,4: 79/77/82/77

Gaming (Left4Dead) 20 minutes test 2
Cores 1,2,3,4: 70/66/69/65


Idle test 3
Cores 1,2,3,4: 46/44/46/44

Prime95 20 minute torture test 3
Cores 1,2,3,4: 79/79/82/78

Gaming (Left4Dead) 20 minutes test 3
Cores 1,2,3,4: 69/65/69/63

All temps in degree Celsius measured by RealTemp. Don't so much pay attention to the temps per se but more to the differences later on between the coolers as that is the main reason why i'm doing this.

After each test, the HSF was removed and thermal paste applied once again. Room temp stayed the same for each test.

I think i won't be able to finish my testing for all 3 coolers today as this seems to take quite a bit longer than anticipated. Tomorrow i hopefully have the results for the other two as well.

However, results so far seem consistent. Core3 is obviously the hottest of the 4 cores, followed by core 1. Anyways, next in line is the Mugen 2 cooler.

Thanks for taking the time to do these

Subscribed.

Second round is done.

Room Temp 24C

Scythe Mugen 2 cooler scored

Idle test 1
Cores 1,2,3,4: 43/41/44/40

Prime95 20 minute torture test 1
Cores 1,2,3,4: 74/72/74/70

Gaming (Left4Dead) 20 minutes test 1
Cores 1,2,3,4: 63/61/64/60


Idle test 2
Cores 1,2,3,4: 44/43/44/42

Prime95 20 minute torture test 2
Cores 1,2,3,4: 74/73/75/73

Gaming (Left4Dead) 20 minutes test 2
Cores 1,2,3,4: 65/63/65/62


Idle test 3
Cores 1,2,3,4: 45/41/45/42

Prime95 20 minute torture test 3
Cores 1,2,3,4: 71/70/73/70

Gaming (Left4Dead) 20 minutes test 3
Cores 1,2,3,4: 62/61/63/60


All temps in degree Celsius measured by RealTemp. Don't so much pay attention to the temps per se but more to the differences later on between the coolers as that is the main reason why i'm doing this.

After each test, the HSF was removed and thermal paste applied once again. Room temp stayed the same for each test.


Next up is the TRUE lapped with penny mod. This will take me all day for 2 reasons:

1. I wanna take one round of temps with the TRUE unlapped. Not 3 rounds but just one. Then I'm gonna lap it, which takes 2-3 hours alone.
2. The TRUE has the most annyoing and hard to install mounting mechanism of all of these by a VERY wide margin. The stupid x-clamp that thermalright uses is downright mean and nasty and it takes me quite a bit to get the cooler installed properly.

Ah this is why I switched to WC. Swiftech GTZ's mounting system is the secks. I wish air coolers would adopt this model....

soulesschild said:

Ah this is why I switched to WC. Swiftech GTZ's mounting system is the secks. I wish air coolers would adopt this model....

Click to expand...

It's really only the thermalright mount that is so annyoing. The other 2 coolers mounted without any problem or effort.

ScYcS said:

It's really only the thermalright mount that is so annyoing. The other 2 coolers mounted without any problem or effort.

Click to expand...

Oh but if you haven't seen the GTZ mounting...you're missing out . The backplate has some kind of tape that sticks to your mobo and stays there. Then you're free to take it off whenever and the backplate stays in place just all nifty like. Add in the screws that prevent over tightening or under tightening....you've got a winner I remember my old Thermalright SI-128. Great cooler, but my god it was hard as hell to tighten it because you had to use a little wrench and it was hard as hell to screw it tight while holding the backplate to the mobo.

BTW if anyone is wondering....i am posting this from my second PC as the other one is (once again lol) disassembled and ready to mount the Thermalright Xtreme unlapped for test number one.

Subbed. Any chance you could get your hands on a Prolimatech Megahalem?

nissanztt90 said:

Subbed. Any chance you could get your hands on a Prolimatech Megahalem?

Click to expand...

Lol no......here's what happend and why i am doing this actually:

I bought a new I7 setup but could not use my Thermalright cooler because i had no mounting mechanism for it and also needed to have a cooler for my "old" LGA 775 setup. So instead of buying a mounting bracket for the TRUE and another 775 cooler, i decided to leave the TRUE in my old machine and get a new 1366 cooler instead. I opted for the DarkKnight from Xigmatek after reading reviews and such as i did not want to spend another 70 bucks on a second TRUE. Mounted it and was a little disappointed by the temps. Then, i looked and ordered a Mugen 2 and was just about to buy it from newegg but they went out of stock. Shopped around and found it at Jab-tech.com for a lower price. Upon checking out, i figured, wth, i just take a mounting bracket for my Thermalright with it and make a little shootout.....

So, thats why i have these 3 coolers here. I don't intend to buy more of them and i think it's kinda hard to get your hands on a Mega right now anyways anywhere in the US.

Ah, gotcha.

Third round part one is done (only one test for the unlapped for comparison to lapped later).

Room Temp 24C

Thermalright 120 extreme unlapped cooler scored

Idle test 1
Cores 1,2,3,4: 45/42/45/42

Prime95 20 minute torture test 1
Cores 1,2,3,4: 72/71/73/69

Gaming (Left4Dead) 20 minutes test 1
Cores 1,2,3,4: 60/59/60/57


Next three tests will be done with the TRUE after i lapped it (using 400grid, 800, 1200,2000 and 2500 grid and polishing it with MAAS polish when done. Mirror finish is not necessary but looks so nice and shiny lol ).

PS: My TRUE is not flat at all. Compared to both, the Xigmatek and especially the Scythe, the base is rough and non-flat. This might work for many chips out there that are not flat (and, in fact my Q6600 was not flat and thats why i never lapped my TRUE before) but my I7 920 seems totally flat, so lapping the TRUE is almost a must in order to make good contact. Out of the box, the Mugen 2 is by far the best quality as far as the base flatness goes. It even has a decent mirror finish.

PPS: I might not be able to use the Penny mod for the TRUE. Not sure why, but i can't get it mounted with the penny under the clamp anymore. Either the 1366 x-clamp is different or something else is different, not sure what. As it stands, i will have to continue the testing without the penny mod.

Very cool thread. I also had an uneven Q6600 I had to lap along with an uneven heatsink. Resutling load temps in the end were ~8C cooler on the hottest core after hours of lapping. For what its worth I used a washer to increase pressure on my heatsink. This let the center piece fall into the center of the washer while still giving increased pressure. I even lapped the washer to ensure it was flat Looking forward to your lapped results!

btw, testing is almost done. I have to make one more dismount/remount and run the temp tests and then i'll post my results. probably in 2 hours or so.

Third round part two is done

Room Temp 24C

Thermalright 120 extreme lapped cooler scored

Idle test 1
Cores 1,2,3,4: 41/40/41/38

Prime95 20 minute torture test 1
Cores 1,2,3,4: 68/66/69/66

Gaming (Left4Dead) 20 minutes test 1
Cores 1,2,3,4: 54/52/54/51


Idle test 2
Cores 1,2,3,4: 41/41/43/39

Prime95 20 minute torture test 2
Cores 1,2,3,4: 70/67/71/67

Gaming (Left4Dead) 20 minutes test 2
Cores 1,2,3,4: 55/52/56/52


Idle test 3
Cores 1,2,3,4: 42/40/42/39

Prime95 20 minute torture test 3
Cores 1,2,3,4: 69/67/70/66

Gaming (Left4Dead) 20 minutes test 3
Cores 1,2,3,4: 54/52/55/52


Conclusion:

1. All three coolers are totally adequate for I7 920 overclocked to 3.8GHZ.

2. The Xigmatek has by far the easiest mounting mechanism

3. The Thermalright has BY FAR the bitchiest, rediculous stupid retarded mount ever.

4. The Mugen 2 is a better performing heatsink than the TRUE 120 extreme when the TRUE is not lapped.

5. Lapping a TRUE can get you tremendous gains (or better: losses) in temperatures. I did not expect such a dramatic difference

6. After mounting and remounting the TRUE 4 times today, i will never ever mount it again.

7. The size of both, the TRUE as well as the Mugen 2 make my first row of memory slots unusable if memory with tall heatspreaders are used. The Xigmatek lets me use all 6 rows.

8. The Xigmatek DarkKnight does NOT fit in a Abit IP35-pro as it interferes with the Northbridge cooler and has to be mounted horizontally instead. It's cooling my "old" Q6600 now.

9. The application of Thermal paste is vastly different with coolers with direct touch heatpipes (e.g. Xigmatek DarkKnight; one must put paste in between the pipes and the base first and wipe off the excess before applying the paste normally; learned it the hard way as temps were way over the top when i did not)

10. Due to Thermalrights stupid clamp mount, drinking alcohol is a necessity when performing the mounting process more than once. It should be noted that with multiple mounting, the alcohol level increased in my blood, the swearing was that of a sailor and i'm well advised to never visit Thermalrights facility as i would prolly throw feces at them.

10a. The TRUE 120 extreme is overpriced. Really it is. But i like it anyways. But please Thermaright......change the mounting mechanism. It sucks. It really really does. For something this good, a mounting mechanism THAT bad just leaves a bad taste in my mouth.

Overall FOR ME, i will keep my lapped TRUE 120 extreme on the I7. I was able to increase the voltage with this cooler to achieve a 200MHZ higher overclock to 4.2GHZ total compared to the Xigmatek while staying at the same temp. The Mugen 2 is the best value of the 3 if you can live with it's size. It's simply HUGE. Is 200MHZ worth it? Ask Intel. Compare I7 920 and 940 prices. YEAH.

For everything else, there's Mastercard.

PS: for anyone interested:

Asus P6T Deluxe V2
I7 920
G.skill 6GB F3-12800CL8TU-6GBPI
eVGA GTX285 SC
300GB Velociraptor
640GB WD Blue
HX620 PSU
Antec P-180SE Case

Fan used to test these babies: Scythe KAZE-JYUNI "SLIP STREAM" 120mm Case Fan - SY1225SL12H - 1600rpm

If you are ever bored, it would be cool if you lapped the other coolers as well to give an apples-to-apples comparison as well. Thanks again for doing these.

Gomit said:

If you are ever bored, it would be cool if you lapped the other coolers as well to give an apples-to-apples comparison as well. Thanks again for doing these.

Click to expand...

Not possible. Here's why:

1. The Xigmatek is a direct touch heatpipe cooler. If you lap it, you will possibly harm the cooler base.

2. The Mugen 2 was (is) absolutely flat with a mirror finish already. I could not possibly make that cooler better by lapping it.

ScYcS said:

Also, each of the heatsinks will be equipped with the same fan.

Click to expand...

What fan was used?

ScYcS said:

Fan used to test these babies: Scythe KAZE-JYUNI "SLIP STREAM" 120mm Case Fan - SY1225SL12H - 1600rpm

Click to expand...

Always wanted to quote myself

BTW apropos fans: I now run my Thermalright with the fan that was included in the Scythe Cooler. It's the same fan as the one i used for testing but runs at 1400rpm instead and is a lot quieter and did not increase my idle temps and only 1C for load. It's a special model that only comes with the Mugen 2 cooler and can't be purchased seperately.
Oh and the Xigmatek DarkKnight fan......let's just say.....it's in my garbage. Whiney high pitched noise at full speed and worse performance at 2500rpm than the Scythe at 1400rpm. End of story.

I never had any trouble with my TRUE mount, do you have any plans on lapping your CPU? I also don't think it is over priced at $49.99, for as good a cooler it is I was willing to pay it and would again.

Not gonna lap the CPU as it's flat. And $49 does not buy you a TRUE for 1366. Nor does it make it lapped or even flat. Unlapped, the heatsink is on par with the Mugen 2, which is far cheaper.

soulesschild said:

Oh but if you haven't seen the GTZ mounting...you're missing out . The backplate has some kind of tape that sticks to your mobo and stays there. Then you're free to take it off whenever and the backplate stays in place just all nifty like. Add in the screws that prevent over tightening or under tightening....you've got a winner I remember my old Thermalright SI-128. Great cooler, but my god it was hard as hell to tighten it because you had to use a little wrench and it was hard as hell to screw it tight while holding the backplate to the mobo.

Click to expand...

FYI, that sounds a lot like the Xigmatek 1366 kit. The backplate comes with adhesive pads to keep it stuck to the board. The spring-screws are only threaded up to a certain point, so you just screw them all the way in and the spring tension (between the screw head and the cooler's mounting brackets) is actually what holds it.

ScYcS said:

Not possible. Here's why:

1. The Xigmatek is a direct touch heatpipe cooler. If you lap it, you will possibly harm the cooler base.

2. The Mugen 2 was (is) absolutely flat with a mirror finish already. I could not possibly make that cooler better by lapping it.

Click to expand...

There are a few people on here that have lapped the xigmatek. I can't remember if it helped or not but you can lap it. Just cant hit it for 10 min with 400 grit and not expect to bust through the pipes. As for the TRUE did you manage to try anything to increase the pressure? Awesome thread!

Copyright said:

As for the TRUE did you manage to try anything to increase the pressure? Awesome thread!

Click to expand...

No, i could simply not get the screws mounted if i put ANYTHING as a spacer.....but the temps seem ok without after lapping. And thanks.

Awesome thread. It's really unfortunate that the Prolimatech Megahalems is so hard to come by. I would have enjoyed reading your take on it.

I would have to agree with you about the mounting system of the TRUE. It's TRUE-ly a severe pain in the, well...you can do the math. That being said, I applaud you for mounting and remounting it as many times as you did. XD

ScYcS said:

No, i could simply not get the screws mounted if i put ANYTHING as a spacer.....but the temps seem ok without after lapping. And thanks.

Click to expand...

Intresting. I wonder if the bracket is different on the 1366 vs the way they built the 775 version. I didnt have much of a problem getting the spacer on mine but mounting that thing was never any fun.

You probably would have gotten better performance, at least on the TRUE120 with a fan with better static pressure, like a YL High speed. Slipstreams have awesome CFM ratings but have very low static pressure so they're great for case fans but awful for heatsink/radiator fans compared to YLs. Just a thought.

Also, the mounting mechanism must be different for the i7 because my TRUE is incredibly easy to mount and I actually like its mounting mechanism better than any other heatsink I've used. Mine was also lapped and washer modded.

dealmaster said:

You probably would have gotten better performance, at least on the TRUE120 with a fan with better static pressure, like a YL High speed. Slipstreams have awesome CFM ratings but have very low static pressure so they're great for case fans but awful for heatsink/radiator fans compared to YLs. Just a thought.

Also, the mounting mechanism must be different for the i7 because my TRUE is incredibly easy to mount and I actually like its mounting mechanism better than any other heatsink I've used. Mine was also lapped and washer modded.

Click to expand...

I had the fan at hand and did not want to spend more money on anything else.....and as far as mounting mechanism....it's the same x-shaped clam for the 1366 as for the 775 but with a longer base plate. I find this mechanism totally stupid and not very easy at all. It never stays in place when screwing it in, it moves, it does not secure the heatsink very well and just overall is sub-par, at least compared to the rather nice mounts of the other two heatsinks i tested.

Have you mounted the TRUE120 on a LGA775 motherboard? If so, how would you compare the 1366 and 1775 mechanisms?

dealmaster said:

Have you mounted the TRUE120 on a LGA775 motherboard? If so, how would you compare the 1366 and 1775 mechanisms?

Click to expand...

They are identical with the exception of the length of the base plate and the x-clam. Only difference was that on the LGA775, i was able to penny mod it and on the 1366 i could not squeeze anything in.

ScYcS said:

They are identical with the exception of the length of the base plate and the x-clam. Only difference was that on the LGA775, i was able to penny mod it and on the 1366 i could not squeeze anything in.

Click to expand...

I wonder if Thermalright modified the bracket for increased pressure compard to the 775.

That is indeed possible....with the LGA775 the penny mod was possible, not easy but possible. With the 1366 mount, i was afraid i would crack something before even getting close to put a penny there. Either way, the end results seems to be that the heatsink STILL can be moved slightly, even when the screws are tightened completely, which i think is just not a good design. If they would've used a octagonal center screw with edges, then this would already be prevented but no, they used a round one, which of course does nothing but swing back and forth in a circle....not sure if i described this properly but i hope you guys know where i'm getting at.

Nice work, Thank you.

I know from using the Mugen 2 that it works well but I wanted to know how it stacked up against exactly these coolers. I have noticed with a cpu voltage increase when OCing that it requires at least one additional fan due to the massive size. One fan just can't move enough air across it so one side is cool to the touch and the other is hot.

I was finally able to penny mod the TRUE on my 1366. It took a slightly thinner washer than a penny to do it but i got it in. Temps dropped another 1-2 degrees under load with it but i didn't measure the ambient temp at the time, so it might be that it was 1-2 degrees cooler in my room as well.

Nice ScYcS! You do good work!

Due to Thermalrights stupid clamp mount, drinking alcohol is a necessity when performing the mounting process more than once.

Click to expand...

I've got the drinking part down but when the going gets tough, it's time to bring out the big guns...left-handed cigarettes seem to settle jagged nerves quite well...or so I've heard!


Nice job and thanks for taking the time to do it right! (Big Thumbs Up)

good call testing the true last then. saved yourself mounting it a 4th time.

dealmaster said:

You probably would have gotten better performance, at least on the TRUE120 with a fan with better static pressure, like a YL High speed. Slipstreams have awesome CFM ratings but have very low static pressure so they're great for case fans but awful for heatsink/radiator fans compared to YLs. Just a thought.

Also, the mounting mechanism must be different for the i7 because my TRUE is incredibly easy to mount and I actually like its mounting mechanism better than any other heatsink I've used. Mine was also lapped and washer modded.

Click to expand...

You have to rig the yate loons to be able to mount them since they are corners are closed on them. It wouldn't look as clean imo.

Also, the true mounting system totally sucks balls

I don't think the TRUE mounting bracket is too terrible. It's not good, and it's far from great...but it could be worse. (I hear Zalman's bracket on their coolers is even worse.)

Noctua's makes it look like child's play though. Both my NH-U12P and NH-C12P have basically the same mounting schematic, and it is light years ahead of Thermalright's in not only simplicity but also secure-ness. Thermalright's, you can twist and turn the cooler after it's screwed down; Noctua's - you can't.