Today I've the Prolimatech MegaShadow on the bench. It's real heavy metal, for real heavy cooling-but how hot can it handle? I'm throwing the new and improved tech bench at it, and we'll see if it turns out to be a Megatron, or if there's something more... Prime about it! ....More here
Prolimatech isn't a name you hear unless you like heatsinks so large they can be used as self-defense weapons, under normal circumstances. They're comprised in great part of former Thermalright engineers, and their initial offering, the Megahalems, took the world by storm, oftentimes beating even a lapped and washer modded TRUE. This is its evil twin, the MegaShadow. What's different? Improved mounting, extra fan clips for push/pull, and it's black. Does it make it cool any better? Nah, but man it looks good.
Something about that logo looks familiar...
It comes packaged in what looks for all the world like a limited edition anime box (shades of Fire Convoy, anyone?) Immaculate packaging, all white, with a black plastic band and an inset window to view the strangely Decepticon-like logo on top. This cooler has earned itself many names due to that, and just looking at it freed of its packaging, it's pretty evident why. It's a twin tower of all-black cooling power (try saying that three times fast) that positively scared the other coolers in my lineup despite being sold bereft of fans.
Now, a lot of people have an issue with buying a cooler that has no fan, but I've got some numbers for you that might just change your mind-hold tight. As to the rest of the innards, there's the combined mounting system for S775/1156/1366, a pair of sets of fan clips so you can mount two fans in push/pull, and a full-sized tube of TIM. I've put that aside for a later date (I know of something coming up that might need it) and substituted AS Lumiere, as I've done quite a few cooler changes in the last two days. Its faster curing time is an incredible boon to reviewers, lemme tell ya. It let me actually finish testing this year!
And to think, that all goes on one board
A word on mounting systems. I've seen them all, I swear. Some good, some bad, and some so miserable it's not even funny. As a matter of fact, I design and make mountings for quite a few coolers, and I make my own waterblocks, so I'm not unversed in how to do it. This is how to do it, folks. It looks complex, but every part I received was 110% perfectly machined, easy to assemble, and so *insert string of expletives here* over-engineered that it brought tears to my engineer's eye. The backplate is cast aluminum, not stamped steel, with inserts for the threaded areas in stainless steel so as not to gall or strip. The padding for the back of the socket, instead of being neoprene foam, is a silicone/EPDM sandwich material with the PERFECT amount of give-and it sticks to the backplate, not the board. The screws were all high quality heavy stainless, and the rest of the mounting is cast aircraft aluminum-no flex as opposed to stamped steel. The mounting is designed to last through crashing on a primitive planet-I mean probably a hundred years or more. Serious overengineering.
This is NOT your stock cooler
I keep thinking of Galvatron, for some reason
That's flat. Real flat
The die doesn't lie
No distortion at all
It takes two hands to mount, mostly for the strength of the screws. That's a real positive however, as its main competitor is only a competitor after being lapped and washer modded for stronger mounting pressure. The base was perfectly flat (I used a subject that would have shown me any distortion if there had been any there) and under a layer of the same nickel as the rest of the heatsink. Really glad it's flat, I'm sure you've read my articles on the issues of bowing in the case of heatsink bases and waterblocks. Once mounted, it has just a bit of give-it holds very well to the upper end of the Intel clamping spec.
So how does it fare? That's a complex question, considering that it ships with no fan, and we've not attached one. So, to set this issue up, we're going to take a break, look at our new testing companion, and see what we learn. We'll rejoin the show in a bit.
The first thing we're going to do is examine our test setup. It now consists of:
Intel Q6600 G0, lapped to 1500 grit and checked with engineering rule/gage block
Gigabyte GA-EP45-UD3P s775 P45 board
2x2GB GEiL Esoteria Limited Edition DDR2-PC6400
nVidia Geforce 8800GT 512MB
WD Caviar 160GB SATAII HDD
Ultra X4 1050W PSU
NZXT Lexa S, no top 140mm fan (removed for clearance issues)
It's evolved some since we saw it last, no? I've upgraded to a quad so as to give a better view of the average enthusiast's system, as well as to crank out the heat. I also lapped it because the processor came so concave it was sickening. I had to apply TIM like frosting beforehand. Now, it sticks to gage blocks through force of attraction. The MSI board is hot, sure, but it also has limited quad support, while the VRM on the EP45 is much more robust, and avoids the nVidia stability issues as well (read as I tried without success to get the 680i to play and it wasn't worth it.) I've also upgraded to 4GB of RAM and 64-bit Windows 7 Ultimate as 64-bit places more of a stress on components, as well as the fact that no self-respecting techie with the means to do it would stay 32 bit anymore, anyway.
Now, as to what we're about to see here-the temps generated here are all collected from the on-die temp sensors on the chip, as monitored by OCCT Perestroika v3.1 (the latest release.) The proper Intel-supported way is by using a special processor, which has had a groove milled between the cores and a thermal sensor thermal epoxied in, then lapped back down. I don't have that. I'm looking into getting one made, but for the moment, this will have to do. All tests are taken with 45 minutes for TIM to cure (AS Lumiere has a 30 minute cure time) so as to be as similar as possible. Ambient in my office hovers around 24C as measured by FLIR IR thermometer. Hygrometer reads rather dry for a house, often well below 50% humidity (the office is the driest part of the house oftentimes.) With the ambient determined, let's start wading through the numbers.
First off is the real baseline, an Intel stock Q6600 heatsink. It's decently capable, being a finned aluminum extrusion with a copper core pressed into it. The PWM fan it comes with has decent static pressure as well (I don't have solid or non-conflicting info on the fan or you'd get it here.)
Early on, it's really evident that the stock cooler, um, sucks. With a peak of 76C, it's not in danger areas, but it's getting mighty close. This is still a tenable solution, for businesses or massive deployments, but it's really evident that there's better out there. Also notice that it gets hot and can't scrub the heat. That's not a thing you want to see on a higher end heatsink.
Next is another pushpin mounted heatsink, but the best one of its type ever made-a Scythe Mugen. This heatsink has fewer heatpipes and a looser fin cluster than the Megahalems, which should translate to decent but not stellar temps. The heatsink is tested with the original fan from Scythe, with specs of:
Fan Dimensions:
120 x 120 x 25 mm
Fan Noise Level:
25.00 dBA
Air Flow:
68.54 CFM = 117 m³/h
Fan Speed:
1,200 rpm (±10%)
And as you can see, it's an improvement. It peaks at 48C, with a more evident rise and fall of temps, as it's more efficient at getting rid of the heat. OCCT's near sinewave pattern shows up here, as the cooler is able to mostly keep up with the cycling of the CPU. However, the pushpin mounting lets it down, as it simply can't put enough force on there to do the job right.
However, once I finally pulled the big boy out of his box, was it able to handily defeat all comers, or was it more of a race?
First off, I tested the MegaShadow with a low-speed Yate Loon, for those of us who enjoy our quiet (or our music.) The model is a D12SL-12, from Petra's Tech Shop. Its specs are:
Model: D12SL-12
Fan Size: 120mm x 120mm x 25mm
Speed: 1350 +/- 10% RPM
Airflow: 47.0 CFM
Noise: 28.0 dBA
Bearing: Sleeve
Voltage: 12 V
Current: .3 Amp Max
Fan Life: 30,000 hours
Connector: 3pin and 4pin pass-through connector
Mounted and ready, so let's start cooking! And the numbers are better than to be expected-the Yate has a lower static pressure and CFM rating, but the numbers don't lie. The greater efficiency of the cooler means it records the same top temp as the Scythe, but with fewer trips there and a more clearly defined waveform.
So what happens if we mate the Scythe with the MegaShadow? Omega Prime-I mean a Megashadow with the rather nice Scythe fan. How does this combo fare?
Now only Core 1 ever sees 48C (hot core, as you can tell by the tendencies shown over the collection of graphs) and the rest average up to 2C lower. It's still improving, so let's get a heftier fan.
Next up is an AC Ryan BlackFire4 Limited Edition 120mm fan. That's a lot to say for a simple fan, but the specs aren't too shabby.
Fan specifications:
Clear Acrylic with UV-active™
80x80x25mm / 92x92x25mm / 120x120x25mm ball-bearing
powered via 3pin fan connector or 4pin Molex power connector (both included)
80mm 92mm 120mm
Low power consumption: 1.3W 1.9W 2.8W
Fan speed (RPM) 2500 2500 2000
Noise level (dB) 22.18 25.24 28.95
Airflow (CFM) 32.37 43.40 77.70
The claimed noise is trash, but the 77CFM is a bit closer. After thirty minutes on the torture rack, did it give up any secrets?
It shaved 1C off the hottest core, but especially on it you can see that it's now accruing so much thermal dissipation that 100% load takes a while to plateau in temps-there's a lot of grunt still left in this cooler, in other words. Overall load temps on the other cores have gone down as well, so there's still stuff to be found.
Where else are we gonna go for more gains, though?
Last, but not least, and so strange and dangerous that it gets its own page, is my experiment of stuffing an 86.5CFM 38mm Panaflo on this thing. This involved zip ties, swearing, and turning my left hand into hamburger. It also involved putting the fan on the floor at 100% and letting it chase my cats.
At this point, there is NO mounting for 38mm fans included. I did this at my own risk, simply because I have a sick fascination with overcooling and because it's science, and science is awesome. How does it stack up against the other fans? Let's find out.
First, specs are:
Dimensions: 120 x 120 x 38 mm
Airflow: 86.5 CFM
Fan Speed: 2100 RPM
Noise: 35.5 dBA
Bearing: Hydro Wave
Power: 4.08 Watt
With a thicker fan body, there's a better static pressure, as well. The MegaShadow has a tight fin pitch, so that helps.
And the judges said... Nothing, there are no judges here! However the
temps on this are quite a bit better. A peak of 44C on the hottest core, and averaging 39-41 on the other cores is a marked improvement. I'm still hunting for my AC powered Nidec that can cut off fingers, but until I find it, if you have more fan, it will take it.
So what does this leave us with? Well, a few interesting things.
First, the comparison between the Scythe and the MegaShadow is less than fair, because their mountings are so disparate. The fact that it even tried to get up there shows that Scythe made a great cooler as well, but with the same airflow, the MegaShadow peeled away, and had the capacity to continue to do so even with more CFM, whereas the Scythe doesn't benefit as much due to fewer and more open fins. It's not apples to apples, it's apples to Megatron.
Second, more directly on topic, there was nothing I could throw at it fanwise it wouldn't benefit from. Fun fact, it even ran that quad passively during a discarded test, when a fan dropped off while I figured out how to mount fans right on it. The PC never shut down, but it looked a lot more like the first Intel stock sink test at that point...
So where does this leave me? I want to see 38mm fan clips, and there's an AMD mounting kit this didn't come with as well. For $80 dollars, those aren't too big to ask for. It does, though, come with a full-sized tube of TIM, which is a really nice touch. I struggle to call this thing a true "value" in the strictest sense of the word (I mean it's a bare sink, you know) but in the intended market it's not that abnormal. What is worth paying for is the excellent mounting system that doesn't need any modding to work right, just like the base needs no lapping to work right.
There's no pros and cons section this time. The numbers don't lie. The only real con is the price, but it's well in line with other high-end heatsinks, so it's just picking fights to complain. The missing fan lets you add whatever you want to fit your needs, from mild to wild.
In closing, I give the MegaShadow a 5/5 and the title of Epic. These guys did it all right the first time-give them some love and pick one up. You won't regret it. It's currently available at FrozenCPU.com and Amazon.com for around $80 ...not a bad deal for one of the best air based CPU coolers in the world.
So, Decepticon or Autobot? I'll just say this:
Transform and roll out!
|
