CoolerMaster Hyper 6 mod: 4-way fan mounting (photo)

wumpus · Post by **wumpus** » Tue Jul 13, 2004 7:47 pm

I think this deserves its own thread. Read the official SPCR review of the Coolermaster Hyper 6 first.

Now, my modification:

You can read a little background on the third page of Hyper 6 review thread.

I first used the tin snips to remove the alu shroud, by cutting at the base. This also had the benefit of helping with clearance on my mobo, because the shroud pushed the VR heatsink assembly over a bit. Now it doesn't.
I took an unused 60 to 80mm fan adapter and mounted it by making "screwholes" between the copper fins. I mounted as high as possible, eg, the upper mount point of adapter is between the highest and 2nd highest fins. That provides maximum clearance for DIMMs, etc (and I can verify, my DIMMs install fine directly under it). This is surprisingly secure and only bends the copper fins a bit near the screw areas. I can now "mount" the fan on any of the four sides, so I chose the side that directs airflow back and out of the case. Note that the two opposite sides ARE smaller (rectangle), directly mounting the 80mm fan is not possible there.
Mounted 80mm fan in the newly installed 60 to 80mm adapter. Using the adapter also helps remove the "dead spot" behind the fan's motor, and "ducting" the fan into the heatsink is an old technique to get reduced noise (ala Scythe Samurai). So the adapter should be a net PLUS.

Why is this mod worthwhile? The Hyper 6 is arguably the best performing ultra-low airflow heatsink currently on the market. The SPCR review confirms this:

It's a downright massacre, isn't it? The cooling advantage of the Hyper 6 [over the SP94 using same fan] is 4°C at 12V, which is pretty big as heatsink performance goes. But it gets bigger as airflow is lowered: 7°C at 7V and a whopping 18°C at 5V.

Well, we have a tie in cooling performance. The Thermalright SP94 is able to match the performance of the Hyper 6, but with a >50% airflow advantage [using 92mm fan] and higher noise. It's true that the noise difference at 7V and 5V is much smaller than the 6 dBA difference at 12V, but it is still there; the 80L is a quieter fan.

The only real obstacle to using this heatsink is A) the weight concern and B) the limited fan mounting that required a specific motherboard layout. I've provided what is, IMO, a trivial mod that removes the fan mounting restriction completely.

As for weight, this is obviously not a heatsink you want in a system that will be moving around a lot, such as a LAN party machine. But I remain unconvinced that there is any real, non-theoretical danger for a normal system installation. The mounting is unconventional and kind of annoying, but it's secure and through the motherboard holes, supported with a back brace. Rounded IDE cables are theoretically way out of spec, but everyone uses them anyway.. and we've been installing way-heavier-than-recommended heatsinks for years now.

Sledge · Post by **Sledge** » Wed Jul 14, 2004 6:38 am

The next step is to make a custom duct from the back of the Hyper 6 to that nice quiet 120mm aluminum fan

Edward Ng · Post by **Edward Ng** » Wed Jul 14, 2004 6:56 am

Sledge hit my idea right on the head.

nannygoat · Post by **nannygoat** » Wed Jul 14, 2004 9:57 am

The hyper-6 looks like it was practically made for a rear duct mod....
I like the shiny fan wumpus - do you have to wipe the blood off everytime you're fiddling around inside, or do you just whack it up to full speed and let physics do the work?

wumpus · Post by **wumpus** » Wed Jul 14, 2004 1:30 pm

It's not really aluminum; the frame may be, but the blades are just painted silver. It's the Evercool aluminum 120mm, which is (IMO) the quietest 120mm fan I have ever heard (sample size ~9 brands).

One other comment on mounting. I noticed-- when removing the heatsink after my first install-- that the Ceramique paste I used creates a pretty darn strong suction bond between the flat plate of the bottom of the heatsink, and the flat plate of the athlon 64 heatplate-- both rather large surface areas.

I had to actually wiggle the heatsink back and forth to get it to "break free" of that suction bond. Pulling straight up, I was actually afraid I would rip the CPU out of the socket.

Anyway, that's just another counter to the criticisms of the fiddly mounting mechanism of this heatsink. Again, no way I am proposing this HSF for a LAN machine-- that would be just plain stupid-- but for a normal, sedentary system install you'd have to be a total wussy girly-girl* to be worried about mobo damage.

* That's right! You heard me, ladies!

wumpus · Post by **wumpus** » Mon Jul 19, 2004 2:41 pm

Update: using some heavy plastic folders scavenged from Office Depot, I built a an "L" shaped duct around the 120mm fan, which extends from the back of the case all the way past the front mounted fan on the Hyper 6. So it covers the side and the top completely, forcing all air from the 120mm fan to blow directly across the hyper 6.

I was sorely disappointed in the results: no significant temp change. In fact I may even be seeing a sight temp rise, for some reason, I'm not sure.. but definitely no improvement

So either the air flow was already so close to optimal that it didn't matter, or something else is going on...

wumpus · Post by **wumpus** » Tue Jul 20, 2004 8:25 pm

OK, I know I'm replying to myself over and over, but I have more information

I can confirm that the duct had no real impact on my system or heatsink temps. That surprised me. My theory is that the airflow is already close to optimal: as it is, there's really nowhere for the air go other than across the heatsink and out the back of the case. The only other fan of any significance in this system is the PSU fan, and it is isolated in its own little chamber. (yes, there is a VGA Silencer on the video card in this box too)

More importantly, I just added a NoiseMagic NMT-2 temperature controller to the rear 120mm fan. I placed it on the top left corner, so it is effectively measuring output air temps closest to the base of the Hyper 6. This is particularly advantageous on an Athlon 64 because of Cool n' Quiet; the difference between idle and load temps are much greater than they would be on a P4.

And the NoiseMagic works great! At idle the 120mm slows down a lot; under load the output temps from the hyper6 cause it to speed up (it's already fanmated, so I can set the max speed to taste). The 120mm fan is arguably the loudest fan in the system, so any decrease in speeds results in a direct improvement in the #1 noise factor with this configuration.

I loves me some Cool n' Quiet!

jimveta · Post by **jimveta** » Tue Jul 20, 2004 8:29 pm

hello! have you tried just ducting--as air-tight as possible--the 120mm to the heatsink and forego the 80mm? that is, use the 120mm to suck all of it's incoming air through the heatsink to cool it passively? dell does this ALOT even with really hot cpus. one way i was thinking would be using flexible tubes and a 120mm -> 80mm cone adapter (since all tubes i've seen are 80mm wide) .. but there may not be enough space between the two though ..

wumpus · Post by **wumpus** » Tue Jul 20, 2004 8:39 pm

No, I didn't try that. I bet it would work though!

The thing is, that 80mm fan is only rotating at ~1100rpm, AND it is on the inside of the (damped) case. I'm doubting I would hear any difference from the outside of the case. Compare with the rear mounted 120mm fan, which is definitely the most audible thing in the system currently.

I'm not entirely sure how much the 120mm (fortron 350) PSU fan is registering at this point. With the reversed case config-- PSU at the bottom-- it is much harder to hear than it would be in the "normal" orientation at the top of the case. Anyway, since it is in its own isolated chamber*, I doubt it will ramp its fan much, and my experience has taught me that most quiet PSUs are very quiet until they start to ramp their fan. If necessary I can mod that, move the internal thermistor back or replace the 120mm fan entirely. I'm not sure if that will be necessary though.

* Technically it is merged with the hdd chamber. I drilled out the rivets holding the wall between the hdd chamber and the psu chamber, and removed it. This is a great airflow mod, I heartily recommend it to all lian-li V1000 and V2000 owners..

bomba · Post by **bomba** » Wed Jul 21, 2004 5:27 am

jimveta wrote:hello! have you tried just ducting--as air-tight as possible--the 120mm to the heatsink and forego the 80mm? that is, use the 120mm to suck all of it's incoming air through the heatsink to cool it passively? dell does this ALOT even with really hot cpus. one way i was thinking would be using flexible tubes and a 120mm -> 80mm cone adapter (since all tubes i've seen are 80mm wide) .. but there may not be enough space between the two though ..

You could also try a couple other options. Would be interesting to compare the results. You say the 120mm fan is loud? Then keep the 80, keep the duct and get rid of the 120. Alternately move the 80 to the other side of the heatsink in suck mode, and duct to the de-fanned 120mm hole.

shathal · Post by **shathal** » Wed Jul 21, 2004 10:32 am

I assume this is just my paranoia talking, but aren't you concerned that the mobo will crack, with the HS at that angle?

It's certainly a heatsink I'd not be comfortable to plunk onto my new hardware

.

wumpus · Post by **wumpus** » Sun Feb 13, 2005 9:52 pm

6 months later, no issues whatsoever.

By now, there might be alu-finned heatpipe equivalents to the Hyper-6 that will cause less nervousness amongst the.. uh.. skeptical.

TheWesson · Post by **TheWesson** » Fri Feb 18, 2005 12:33 am

wumpus wrote:Update: using some heavy plastic folders scavenged from Office Depot, I built a an "L" shaped duct around the 120mm fan, which extends from the back of the case all the way past the front mounted fan on the Hyper 6. So it covers the side and the top completely, forcing all air from the 120mm fan to blow directly across the hyper 6.

I was sorely disappointed in the results: no significant temp change. In fact I may even be seeing a sight temp rise, for some reason, I'm not sure.. but definitely no improvement

So either the air flow was already so close to optimal that it didn't matter, or something else is going on...

it is my belief that forcing an duct to be airtight (or nearly so) can actually be counterproductive.

Why?
a) pressure differential (from trying to 'force' the air through resistance) reduces CFM throughput from axial fans a lot.
b) air blowing past your target (from a loose duct (or even no duct?)) helps move air into/through your target anyhow.

my experimentation with different exhaust ducts seemed to bear this out.

but you're probably right that your original system was close to optimal anyhow. Maybe a loose duct would give you 1-2C more, maybe not.

deanjoly · Post by **deanjoly** » Tue Mar 22, 2005 12:03 pm

I'm currently building a rig using the Hyper 6 inside of the Silverstone TJ-06. I can't wait to see how efficient the wind tunnel will be with this heat sink. Also going to have it all connected to the Silverstone fan controller, should be nice and quiet except for under extreme circumstances.

Has anyone else built a similar rig? What did you experience?