Review: Thermalright Ultima-90 ES vs TT120 (56k warn)

[Battle_Rattle]

Review: Thermalright Ultima-90 ES vs Tuniq Tower 120

Thermalright has been releasing some interesting new products recently. I'd like to show you an engineering sample of what could be described as the “little brother” to the yet to be released Ultra 120 Extreme, the Ultima-90.

The Ultima-90 with 92mm Fan (Fan will probably not be included in production models)


Every review needs a known entity to compare a subject to and I chose the Sunbeam Tuniq Tower 120. The Tuniq has been one of the gold standards of high end air cooling for the past year.

The two coolers. The Ultima-90 is mainly intended for 92mm fans, but the sample I received included 120mm fan clips. I'll focus on 120mm fan performance.


Profile of the Ultima-90 and Tuniq


The Base of each, pre-lapped


Heatpipe configuration


Heatsink Dimensions

Thermalright Ultima-90(i)
48 fins 6 heat pipes
Heatsink size (LxWxH) 115 x 55 x 138
Heatpipe width = 6mm
Weight without fan = 453g

Sunbeam Tuniq Tower
51 fins 3 heatpipes
Size (LxWxH) 131x108x153
Heatpipe Width = 6mm
Weight with Fan = 965g

Testing Configuration

- Open Test Bench with Asus P5B-D
- Both Heatsinks lapped
- Lapped e6400
- Temperature measured by CoreTemp beta 0.94
- Load temperatures established by Orthos smallftt set at priority 9 for 5 minutes
- Idle temps established by idle time of 15 minutes with minimal XP background applications
- Ambient 20c-21c
- Arctic Silver 5 Thermal Grease with 48 hour “burn-in.” Note: The best application of TIM was used for each Heatsink.
- Bolt-thru mounting for each HS. The Ultima-90 should also offer push pins in the kit.

The test bed


Testing:

I'll dispense with stock settings as they shouldn't apply to this level of performance and move to some decent overclocking of an e6400 (L630A992) at 3.5Ghz and 3.7Ghz. Vcore is the issue, so let's see the results.

Test 1) E6400 3.5Ghz (7x500) - 1.36 VCore - As5 - 20c-21c Ambient
Idle and Load Temperatures
Tuniq (Stock Fan) – 27c idle – 51c load
Ultima-90 (Scythe S-Flex) - 27c idle – 47c load

Tuniq Load at 3.5Ghz screenshot - http://s144.photobucket.com/albums/r...-36v_1900_.jpg

Ultima-90 Load at 3.5Ghz screenshot -
http://s144.photobucket.com/albums/r...1c_1-36v_1.jpg

For test 2 I decided to use the stock Tuniq Fan at full rpm for each heatsink, reviews that ignore this methodology always concern me.

Test 2) E6400 3.7Ghz (8x463) - 1.4875 VCore - Arctic Silver 5 TIM - 120mm Fan (used Tuniq Fan on both for 1 to 1 comparison) - 20c-21c Ambient
Idle and load Temperatures
Tuniq - Idle 33c - Load 62c
Ultima-90 - Idle 32c – Load 57c

Tuniq at 3.7Ghz – 1.48 vcore - Load


Ultima-90 at 3.7Ghz – 1.48 vcore - Load


Looking at the Ultima-90 I bet “push-pull” has entered your minds, so let see a result from two Scythe S-Flex-F fans in a push pull configuration.

Test 3) – Same as test 2 except Ultima-90 is in push pull using Scythe S-Flex-F fans at full rpm.
Ultima-90 - Idle 32c – Load 54c

Screen Shot of Test 3
http://i144.photobucket.com/albums/r...1c_1-48v_S.jpg

Test 4) - Same as Test 2 except a single high rpm/high cfm 92mm fan (Panaflo 70cfm) was used
Ultima-90 - Idle 32 - Load 53c

Screenshot of Test 4
http://i144.photobucket.com/albums/r...1c_1-48v_9.jpg

Noise:

Noise is a matter of taste and mine leans towards the quite side. The Scythe S-Flex was very silent into the 1300rpm ranges. The Tuniq fan was quite loud in the upper ranges as is any fan. One thing worth mentioning was that the S-Flex was noticeably less efficient while in the Tuniq.


Final Remarks

The Ultima-90 has delivered some note worthy results in the two weeks spent testing. The Ultima-90, in conjunction with an e6400 is outperforming a heatsink that dwarfs it in size - why? The answer could be in the heat pipe number combined with the large amount of fins. There are 6 heatpipe coolers such as the Scythe Ninja and 45+ fin coolers like the Tuniq. The performance differential between the Ultima-90 and Tuniq appears to come from combining the two. A 48 fin cooler with 6x6mm heatpipes seems very efficient at heat dissipation. Therein is the difference, never in my searching have a found a heatsink that so densely packs the two together. Distance between the base and the bottom fins may also come into play. Heat does not have to travel long on the Ultima-90 before it reaches the lowest fins. Lastly, the Ultima-90 heatpipes converge right at the beginning and travel the breadth of the base.

For the future, I look forward to more examples of employing 7-9mm heatpipe solutions as has been done on the Enzotech Ultra-X and Thermalright IFX-14.

At this point, I have to say the Ultima-90, once delivered, should be an outstanding high performance cooler at what will probably be a reasonable price.

I'd like to thank Thermalright for sending me this engineering sample and SVC.com for sending the Tuniq out quickly and in good condition.

Many thanks also must go out to ChaosMinionX at Xtremesystems forums for helping me in the methodology of this review. I promise I will get a run of coollabs liquid pro in soon!

If there are any requests or questions please ask them. I'm hitting a bit of a wall with my e6400 at 3.8Ghz, but I could do some suicide runs with 1.5-1.6 vcore...

Cheers ~ Battle_Rattle

Addendum:

Several people expressed interest in performance differences between heatpipes oriented vertically (as they are in the original testing) and horizontally, as they would be in most computer cases. I tested this, here are the results in degrees of change.

Tuniq
Vertical Idle to Horizontal Idle = Delta rise of 1c
Vertical Load to Horizontal Load = No Change

Ultima-90
Vertical Idle to Horizontal Idle = Delta rise of 1c
Vertical Load to Horizontal Load = Delta rise of 1c