Scythe Mugen 2 CPU cooler install tip (and results)

4trees

Distinguished
Dec 4, 2009
147
0
18,710
Pointless post unless you are considering the Scythe Mugen 2.

This may be unnecessary, but from the number of reviews I read about the hastle involved with installing this chubby hsf, I thought posting my experience could be of some use to a few. Basically trying to install it by yourself can be a challenge. This method might make it work and not screw up your effort in precise spreading of the thermal compound.

Components:
Cooler Master HAF 932
a8n-sli
amd 3800x2
G Skill 4x1gb (model#F1-3200PHU2-2GBNS)


I will skip to the hastle part...

So you've got the 2 side brackets attached to the base of the mugen 2, the thermal compound is spread on the cpu (I used artic silver 5) as best as possible, the mobo is facing up laying on desk with the CPU closer to you.

step1
Position the backplate under the mobo with the 2 long screws sticking thru mobo up at you.

step2
Grab the mugen 2 and position it over the cpu as if you were going to set it right on top of it, you can't set it onto the cpu because the 2 screws are obstructing the union.

step3
(You'll have to be on your knees for some of this) Carefully slide the mobo to the edge of the desk until it is sticking out so far over the lip of the desk that 1 of the 2 screws inserted into the backplate is visible from below (you will need to be waiting for that screw to show itself, if you aren't careful it will fall to the floor allowing the mugen 2 to rest on the cpu).

step4
With the screw driver in place below the mobo start to screw it into the threads about a 1/3 of an inch. In the mean time your other hand will need to be balancing the mugen 2 on the 2 screws.

step5
Then slide the mobo further off the desk until the 2nd screw is visible, screw this one in 1/2 an inch or so. Be careful not to screw either screw too far without adjusting the other, the mobo may bend if done unevenly.

step6
Now both screws should be in the threads and the mugen 2 is not fully tightened yet. Keep working the screws a little at a time until you make contact with the IHF of the cpu. Before it is too tight gently twist the mugen 2 back and forth several times to thoroughly spread the thermal paste onto both surfaces. Finally tighten the screws until you are comfortable that the connection is firm yet not torquing the mobo too much.

The rest is up to you. I would have taken pics of this process but I was by myself.

FYI
Currently running EVEREST system stability test to compare the difference of temps before and after. They aren't fantastic but I'm plenty happy. I didn't have a heat problem before, I'm buying this HSF for a new build in a couple of months. Why not make use of it now. What I'm most pleased with is how quiet the case is now. See some numbers below that I have recorded so far.

Using a Xilence Fan Speed Controller so I can't tell you the rpms of the fans, sorry. All I can tell you is percent of max power, I can't assume what that translates into rpms. This is just my own testing please don't critique my lack of detail here. :pt1cable:

AMD 3800x2 overclocked to 2.5 at 1.43v
GTX 260 SC (896-P3-1262)
Cooler Master HAF 932 (3x230mm fans, 1x 140mm fan)
Scythe Mugen 2 (single fan pushing towards the back of case directly at the 140mm)
Corsair TX650 (1x120mm fan)

With stock HSF these were my max temps with EVEREST stability test (HSF was max rpms, the case fans were at 50% max voltage):
ambient temp 24c
core1 50c
core2 57c
GPU 39c
mobo 25c

Max temp results after mugen 2 installed, (mugen 2 set to 40% of max speed which is the lowest I can set it with my controller, both side and front 230mm fans are off, top 230mm fan at 25% max voltage and rear 140mm fan at 25% max voltage) trying to see how hot cpu will get with as little help from case fans as possible. Again these are the max temps only, the averages are 4-6c less. Run for 1 hour 56 minutes.
ambient temp 26c
core1 45c
core2 51c
GPU 40
mobo 27

Max temp results with all fans off except the mugen2 (still at lowest setting 40% of max power), mobo chipset fan (auto), and the PSU fan (auto). Run for 58minutes.
ambient temp 25c
core1 50c
core2 54
GPU48
mobo 39


 

4trees

Distinguished
Dec 4, 2009
147
0
18,710
You may be right. I could use the practice.

The above conditions aren't friendly towards exhausting heat and passing air over surfaces.
The 2nd set of results I listed are with a closed case (a lot of exhaust grates are built into the HAF though) 2 of the fans off, the other 2 at 25%, and the Mugen 2 at 40%.
The 3rd set of results are with a closed case and all fans off.
With that and an increased Vcore at 100% load.


One problem I thought of just now, I didn't clean the baseplate of the mugen 2. I don't know if any residue from that plastic wrap remains on the plate but I would think it might. I also checked the IHF and base of the Mugen 2 for level. They aren't bad but a little light was shining thru when I had a razor over the top of each. If I take it off in a couple weeks I'm going to lap the Mugen 2 and ignore the cpu. Won't have it long enough to matter.


The as5 spreading was about as thin a layer I could make without making significant patches of bare metal. I don't think any dust or debris got in there but I wouldn't be able to tell anyway.

 

4trees

Distinguished
Dec 4, 2009
147
0
18,710
My system was at idle all night, ambient temp was 21c (big difference vs. the 26c ambient last night) at the time of starting stability test again. This time I ran the Mugen 2 at 100% power all the other tower fans were set to 100%. Ran at load for 45min.

Idle temps (ambient 21c):
mobo 25
cpu 29
core1 27
core2 34

Load temps (ambient 21c at start 25c by end because the heater turned on):
mobo 30c
cpu 39c
core1 40
core2 47


I'm not going to show the difference between the two columns, you can do that yourself. Anyway it wouldn't be that great of a comparison because the ambient room temp changed by at least 4c during the test. I don't intend to run the cooling system like this. I want it as quiet as possible so I will probably run all fans at the lowest possible setting and yet still be spinning. The point to my stability test is I wanted to be sure my cpu wouldn't overheat if I left all case fans off and the HSF to lowest point. In that test I am easily safe. Core1 and 2 were 50/54c respectively for almost an hour of load time (with the ambient temp as high as it gets 25-26c). The crit temperature for my cpu is 71c but should stay around or below 65c. So I'm safe.

I think the next and final series of tests I will do is unplug the fan on the Mugen 2 and see how it does at passive cooling in worse/moderate/good tower air movement.

I hate my chipset fan! Want to rip it off the board.
 
Well, you need to have some air moving through your case for ANY CPU hsf to work best, and it looks like the Mugen-2 is working well for you.
If that's an older nForce4 mobo with a chipset fan, you can take off that fan and put one of those Zalman passive heat sinks on it with the fins that fan out. You'll need some airflow through your case for it to work, but I used one a few years ago and it worked well.

Edit: Here's a link: http://www.newegg.com/Product/Product.aspx?Item=N82E16835118014
 

4trees

Distinguished
Dec 4, 2009
147
0
18,710
Thanks for the link jtt283, i bookmarked it for future. I will probably live with it for now. Going to change mobo in a few months anyway but now that my system is considerably quieter than before I find myself being more picky with other issues.

Can anyone explain this...

PC was off for several hours, on startup my ambient temp was 18c, i opened EVEREST and SPEEDFAN asap and initial temps of the hardware was much hotter than ambient. Do things heat up that fast in the few moments of booting?
mobo 23
cpu 27
core1 23
core2 32
gpu 34
 

Conumdrum

Distinguished
Nov 20, 2007
2,397
0
19,960
Ohh yea. These parts would burn up in a minute or even less without the cooling on them.

Have you tried Real Temp, Core temp and HW Monitor as your monitoring software?

You might want to use Prime 95 small FFTs as your loading test. It's better at stable heat gereation than Everest which checks stability, not generating temps like Prime 95 does.