Megahalems vs. TRUE Heatsinks: Stock Comparison on Overclocked i7 Platform

12/17
2009

Megahalems vs. TRUE Heatsinks: Stock Comparison on Overclocked i7 Platform

Posted by Baditude_df in Cooling, Editorials

Two top performers in air cooling are the Prolimatech Megahalems and the Thermalright Ultra 120 Extreme (TRUE). This is an “out of the box” comparison, and is strictly to test the thermal performance of these two heat sinks with stock base and mounting with no modifications. The comparison was done in an environment that is as controlled as possible without having laboratory conditions or equipment.

Test Setup

Both heat sinks will be tested on my bench in open air. Ambient temperature was measured in the room where the tests were performed.

If you want a rough breakdown of fan specs at each voltage setting just divide into three equal divisions, keeping in mind that this controller only gives 11.3v rather than a full 12v at max setting:

Voltage [actual] (v)	Speed (RPM)	Air Flow (CFM)	Noise (dBa)	Pressure (Pa)
12 [11.3]	2400	95	35	55
10 [9.42]	2000	70	30	45
8 [7.54]	1500	50	25	30

Fan voltages were measured with a multimeter. The rest (RPM, airflow, pressure, noise) can only be estimated, but from what I can tell it’s pretty darn close.

The Bench

The Heat Sinks:
- 1 TRUE Black (Original)
- 1 Prolimatech Megahalems
Fans:
- Sanyo Denki San Ace 109R1212H1011
  - shrouded 25mm, push pull added later on for 4.4 and 4.6 GHz stops
  - Current: 0.52A
  - Watts: 6.24
  - RPM: 2600
  - Flow CFM (Max): 102.4 CFM / 2.9 cu meter[m3]
  - Static pressure (in. H20): 64.7 Pa (6.6mm)
  - Noise (dBA): 39.0
Controller:
- Sunbeam Rheobus 6 Channel
TIM:
- AC MX2

Test Methodology

Each heat sink was tested from 8 (7.54)v to 12 (11.3)v fan settings in 2V increments.
Each heat sink was tested for 10 minutes at each fan speed with Prime95 small FFTs, with resulting core temperatures recorded as measured by Real Temp 3.36
For graphical purposes, temperatures have been averaged. I’ll warm things up prior to the 8V test to create the same conditions the rest of the fan setting tests will see.
After the single push fan test, each heat sink was fitted with a second fan for push/pull testing.
Each cooling setup was tested at the same ambient air temperature of 21C. Temperature was measured before and after each heat sink was tested to confirm a consistent temperature.
Each heat sink will be tested on an un-lapped i7-950 @ 4.0GHz with a vcore of 1.264v.

Prolimatech Megahalems

Thermometer Shot: Megahalem

Megahalems TIM Application

Megahalems TIM Spread

Temperature Results:

Single Fan: 8v

Single Fan: 10v

Single Fan: 12v

Push-Pull: 8v

Push-Pull: 10v

Push-Pull: 12v

Thermalright Ultra 120 Extreme (TRUE)

TRUE Thermometer Shot

TRUE TIM Application

TRUE TIM Spread

Temperature Results:

Single Fan: 8v

Single Fan: 10v

Single Fan: 12v

Push-Pull: 8v

Push-Pull: 10v

Push-Pull: 12v

Overall Results

Pretty much as I expected really, a close match. Both coolers performed very well. However, there are a few things I’d like to mention:

The base on my TRUE is very flat. For a stock TRUE out of the box, this would tend to be the exception rather than the rule. I think that pitting the Megahalems against any other stock TRUE would likely yield different results.
The mounting mechanism for the Megahalems, while well designed and easy to install, did not provide as much tension as I’d like. From what I observed (see TIM spread pictures) the TRUE did provide more stock mounting tension at max. However, both coolers did get a good even spread and adequate thermal bond.

Further Testing

Well, TIM application method definitely makes a difference with the Megahalems. I wasn’t expecting quite this much though. The TRUE was not tested with the line method, as I know this heat sink does not take well to it at all (typically adds ~5C).

TIM Application: Line Method

TIM Spread: Line Method

Overall Results after new TIM application

Some interesting information has come out here I think (and forgive me if this is already common knowledge to some):

The TRUE seems to be superior at lower fan speeds. I find this odd since the original TRUE has a denser fin pattern than the Megahalems.
The Megahalems shows a much larger temperature spread between fan speeds (in particular between 8v and 10v). Whereas the TRUE seems to best it at low fan speeds, the Megahalems quickly steps things up, and even passes the TRUE as fan speed increases to ~2000RPM.
With a single fan configuration, the TRUE and Mega seem to perform the same at the max fan speed setting of ~2400RPM. In push pull, the Mega seems to outperform the TRUE after ~2000RPM.
Please keep in mind, that even though the Megahalems is leading the TRUE in this test, we’re talking about less than a degree throughout the fan speed range. The larger differences are actually seen while comparing the same heat sink in single fan vs. push pull configurations.

- Baditude_df

18 Comments:

splat

12-17-09 09:20 PM

thanks for the great article, baditude

mdcomp

12-17-09 09:27 PM

Sweet article...unfortunately comment threads are a little messed up at the moment. We are working on it!

Matt

Editor's Note: Comment thread fixed!

MIAHALLEN

12-17-09 10:56 PM

Nice testing Pat! Thanks

I.M.O.G.

12-18-09 01:51 AM

Nice work baditude!

I see the Megahalem performed much better with the line application method. How did you apply TIM in the initial test? Sorry if I missed it, but I didn't see this mentioned.

baditude_df

12-18-09 02:33 AM

In a dot, same as the TRUE. It's the second picture under the Megahalems heading.
http://www.overclockers.com/wp-conte...a-1024x768.jpg

Thx for your comments guys, it was my pleasure.

dejo

12-18-09 02:58 AM

I have been thouroughly impressed with my mega, even with the scythe cheepie single fan. with the old evga board could do any bench at 4.7 with the i7 920

overvolting

12-20-09 03:15 AM

so whatever happened to covering the whole surface area on both surfaces with a thin layer of thermal paste?

baditude_df

12-20-09 03:28 AM

If it was ever standard practice to coat both surfaces fully, it was WAY b4 I ever got into computers.
However, I find that spreading the TIM around on the surface of the IHS till I get full coverage, yields about the same results as using the dot or line method.

I.M.O.G.

12-20-09 03:28 AM

It was baditude, but it's a legacy method. Most thermal paste manufacturers updated their recommended procedures around the time that integrated heatspreaders became commonplace.

By the way, there's a thread in the news team forum which has a screenshot of your article, mentioned at 3dcenter.org.

wickedout

12-20-09 06:17 AM

This is one great cooler. I'm so glad I finally got it. Awesome review by the way.

When I applied the TIM (MX-2) I did the line method as shown in this article and then I did smooth it out with an old credit card.

I still have yet to run some tests on it.

ozzlo

12-20-09 07:46 AM

Wow, that's alot closer than I was expecting...

i'd be interested to see how they cool with 2 120CFM fans in push-pull cuz that's what i'm runnin right now

bing

12-20-09 08:11 AM

Really enjoy the writing, thanks badtitude !

Btw, just curious about the "shrouded 25mm fan", what kind of shroud did you use ? A pic will be appreciated.

the_cultie

12-20-09 01:10 PM

Nice article. Will be getting one of these heatsink for the next builds so all the info you provided is just what I needed

Thanks!

baditude_df

12-20-09 07:25 PM

Was that question directed towards me?
If so, the first picture of the article shows the test setup: http://www.overclockers.com/megahale...-i7-platform/#
There you can see it is simply a 120x25mm fan body. )

bing

12-20-09 07:35 PM

Of course its for you, just curious whats your experience with and without that shroud, at the performance and noise wise.

.

baditude_df

12-21-09 07:45 AM

Performance and sound wise, no difference to speak of. The fan blades still push air over the part of the sink that caries the most heat. One thing's for sure though, the shroud doesn't reduce the fan's performance, and as long as I'm running these Huge hub Denki's and I have a few old 25mm fan bodies here, I will throw one on to help reduce dead spot.

buckyball

01-04-10 06:40 PM

Timely article, as Santa

was building a machine for my son. I chose the TRUE to avoid possible clearance issues near the socket and a Delta QFR1212GHE-PWM Ultra high-speed fan:

http://www.heatsinkfactory.com/delta...speed-fan.html

My current top results are:
Core i5 750 @ 4140 Mhz

=================> Cores Active @100%
-----------------------> Idle ---- 1 ----- 2 ---- 3 ---- 4
CPU Temp (C) ------- 27 ---- 49 ---- 53 -- 60 --- 64
Fan Speed (RPM) - 3950 -- [ 5700-5800 (read: *loud*) ]

Gigabyte GA-P55A-UD3P
Corsair XMS3 PC1600 8-8-8-24 (stock, not tested settings)
Gentoo Linux + 'lm_sensors' + 'gkrellm'

I was unable to divine a way to monitor both the Tcase Max and Tjunction Max temps on Linux (a la Real Temp on Windows). I therefore extrapolated per the Intel doc that Tjunction Max = Tcase Max + 5C.

This means that at 4 active cores I am right on the 70-72C threshold. With the PWM fan control, the noise is just bearable at idle and the fan speed cycles up and down with load. At 4 cores max, the fan sounds like a leaf blower.

My son plays WoW (a lot)

, and with the game sound FX cranked up a bit, the fan noise is pretty much un-noticable.

(Edit - formatted the "table" for readability.)

baditude_df

01-04-10 07:06 PM

Ahh, good deal!

I love the TRUE personally, and would buy another in a heartbeat.

Overclockers

Most Popular Articles