Spire Gemini Rev. 2 Review

airman - 2011-11-04 11:19:34 in CPU Cooling
Category: CPU Cooling
Reviewed by: airman   
Reviewed on: December 1, 2011
Price: $52.99


Though a little less mainstream, Spire has been a part of the cooling community for as long as many of its other competitors. Having branches in countries all over the world, Spire's manufacturing plant and corporate office is located in ShenZhen, China. Not only does Spire manufacture CPU coolers, but it also has products in the categories of cases, power supplies, hard drive enclosures, and other accessories such as speakers, card readers, and other peripherals. Today I will be testing Spire's Gemini Rev. 2. Launching in late August of this year, the Gemini Rev. 2 has received good feedback and has produced good results in comparison to other high performance coolers on the market.

Sporting six 6mm copper heatpipes on a copper base and stacked with 52 dimpled aluminum fins, the Spire Gemini Rev. 2 is packaged with two 120mm 9 blade fans that flow up to 71CFM at ~25dBA. It is compatible with every recent CPU socket, even the fresh socket 2011 by Intel. Being the first Spire heatsink that I have had in my possession I am looking forward to getting the Gemini Rev. 2 on the test bench and loading up the processor and seeing what it can do. In this review, I will be performing an in-depth evaluation of this cooler from unboxing to a full performance evaluation in idle and load scenarios, at both stock and overclocked speeds. Without further ado, let's get started!

Closer Look:

The packaging of Spire's Gemini Rev. 2 is nothing too flashy and is a functional design. There is a black plastic handle on the top of the box along with a plastic window on the front of it. It contains the flagship burgundy-red color on the majority of the box. Through the window you can see one of the fans, which is black in color. The left side of the box displays technical specifications of the cooler, which can be found on page 3 of this review. On that page you will also find a list of features as provided by Spire. The back of the box is rather plain compared to other companies' retail boxes, in the sense that there is only a short paragraph that talks about the cooler and the importance of keeping the processor cool. There are also four small pictures beneath that paragraph of different comparable models that Spire has available.









Upon opening the box and unpacking the accessories and mounting hardware, I can say that I was a little intimidated by all of the different pieces! After reading the manual, however, I was able to have any confusions cleared up about its installation. Included in the box aside from the cooler and two fans are: required hardware for AMD and Intel sockets, two pairs of fan clips, a fan speed controller that mounts to a PCI slot, a 2-way splitter to power both fans on one plug, thermal paste, and the user's manual. Everything was neatly packed in small plastic bags and nothing was out of place.



With everything out of the box it's time to get started on a closer look of the cooler itself. I've always liked the look of nickel plating on heatsinks — not only because of the color and sheen but also for its functionality. Stay tuned on the next page and further along for the testing.

Closer Look:

The appearance of the Spire Gemini Rev. 2 is far from plain or simple. Like some other tower coolers, the cooler is bifurcated (split into two sides) on either end of the U-shaped heatpipes. The dimpled fins not only give the cooler a textured look but also increases the surface area of the cooling area which raises its effectiveness. On the opposite side of the CPU contact area is a small, passive heatsink - further increasing the overall working surface area of the cooler. I have seen this in a few other coolers previous to this one, so this is nothing new to myself or the market but is still an effective use of workable area. Two of the sides are sealed off due to the construction of the fins, keeping the air flowing from the entrance to the exit without any escaping. This too is a previously-seen feature, and keeps the air within the heatsink's volume before it exits the other side.

















Checking out the top of the cooler shows a good picture of the dimpled pattern in the fins and the decorative caps on the tops of the heatpipes. Due to the way that heatpipes are constructed, the ends are generally unappealing visually, so some manufacturers put caps on top of them to clean up the overall look of the product. Looking at the bottom, the heatpipe layout can be observed to be a V-shape. This shape helps get more airflow around each heatpipe, rather than a linear design which can cause airflow blockages on later heatpipes. When laid on its side, another look of the passive heatsink on the top of the base can be taken. The fins themselves are actually milled into the base material itself, which means that it is all one piece and lacks any performance-hindering interfaces like glue.




When looking at the quality of the base, it is indeed flat — but the machining marks are quite evident. The less polished that a heatsink base is, the less contact area between the heatsink and the CPU which hinders performance. Thermal paste assists with this, but it's always better to have as smooth of a base as possible. In the case of the Spire Gemini Rev. 2, the base certainly isn't perfect. Having been plated over the raw machined copper base, this does smooth out the surface slightly. However by looking into the reflection produced by the base, a "ghosting", or duplicate image can be seen due to the micro ridges on the surface.



When looking closely at the fans, the material used to produce them is a semi-transparent, dark gray plastic. Both sides look rather similar due to there being an identical sticker on both sides, as opposed to them usually being different — I.e. the Spire logo on the front and one with specifications (voltage, current RPM, etc.) on the rear. On one side of each fan there is a foam-rubber damper in each corner. This pre-packaged addition will help silence vibrations produced by the fan, only producing a maximum rated noise level of 25dBA. The rubber dampers will quiet the fans further. Both fans use 3-pin, non PWM connectors and draw a current of 0.15A at 12V. Operating at a speed of ~1700RPM, each moves approximately 70CFM.



The fans are secured to both sides of the cooler using the supplied two pairs of fan clips. First I snapped the long part into the slots on the cooler, where they will stay on their own due to the way the slots are cut. Then the fans with the rubber side in are clipped in. It takes a little bit of effort to get them completely snapped in, so I would say the "fan side" of the clips could have been a few millimeters further out so it wouldn't take as much effort to pull them out far enough to hook them into the fans.



I noticed during installation that I thought I was going to have a problem getting the pressure even. With its true 4-point design, it's almost impossible to get all four corners at the same amount of pressure to allow an even contact patch over the entire surface area of the CPU itself. In order for 4-point setups to be as effective as possible, the nuts used to hold down the bracket (which is attached to the heatsink) must be able to bottom out at a specific point - every time. In the case of this cooler, the nuts are threaded too deeply and cannot bottom out before the mounting bracket itself deflects too far for my comfort. This is where good design and quality manufacturing and tolerancing come into play if we want a successful 4-point mount, but why would you? See the designs of the Noctua, Prolimatech, NZXT, and many other coolers. They use 1 or 2-point mechanisms that work beautifully and are completely repeatable with their results every time. I can say that I definitely tried my hardest to get this mounting as close to perfectly-installed as I could but we'll see how the thing performs here shortly. With the cooler put together and installed in the case, it's almost time to load up the testing gear and get started. First, we'll take a look at the manufacturer-provided specifications and features.


131 x 71 x 153 (L x W x H) w/o fan
Aluminum & Copper
Six (6) 6mm all copper U-shaped
Ball Bearing
Rated Speed
600 ~ 1650 +/- 10% RPM
Rated Power
1.8 W
Rated Voltage
12 V
Noise Level
10 ~ 25.0 dBA
Air Flow
71.12 Max. CFM
0.15 A
4-pin PWM
Socket Compatibility
1156 /1155 / 1366 / 940 / 775 / 939 / AM2 / AM3 / AMD FM1 / 2011
130 W
Thermal Resistance
0.090 °C/W
Thermal Paste
BlueFrost – SP802 blue grease (0.5g)
Static Pressure
2.12mm H2O
5 years
Net Weight




Information provided courtesy of Spire-Corp @ http://www.spire-corp.com/main/product_detail.asp?ProdID=1090

Testing and Setup:

Testing of this heatsink will involve applying a load simulated by Prime95, using small FFTs in stock and overclocked scenarios, where both idle and load temperatures will be recorded. Load temperatures will be the maximum value displayed in RealTemp after running eight threads in Prime95 for one hour, and idle temperatures will be the minimum recorded value by RealTemp with no computer usage during a period of one hour. The temperature values for each of the four cores will be averaged and displayed in the graphs below. The ambient temperature is held at a constant 23 °C throughout testing of the Gemini Rev. 2 as well as the comparison units. All the data shown in the graphs below is in degrees Celsius. The included thermal paste from Spire will be used during testing and thermal pastes on other heatsinks from their respective manufacturers will be used. The fans on each cooler will be run at full speed for these tests.

Testing Setup:


Comparison Heatsinks:




Given the cooler's size, weight, material, fan flow, and overall build quality, the numbers are a little higher than where I expected them to be. It ended up right in the middle and tied with the likes of the $20 cheaper Cooler Master Hyper 612 PWM. I can guarantee that this is due to the base and not from its other qualities!


So, what I've determined here is that I have a high-end cooler with a mounting design that needs some work. I have found similar quality mounting systems on $20 or $30 coolers, but with a cost of $52.99 its still a tad high. With a mounting system designed the way this one and others like it are, it's almost impossible to achieve even pressure over the surface of the CPU die! Instead of having a full-faced, even contact patch between the cooler and the CPU, you end up with at BEST a line contact or even worse, point contact — both of which will completely destroy the performance of the cooler. For a better explanation and not to cloud the conclusion with too much explanation, see the end of page 2 for more comments as to what can be improved and what works best.

Moving past the mounting design, the cooler itself is of good quality. The base is flat (although the machining marks are evident), everything is plated nicely, heatpipes are capped off and looked good, and the overall design is appealing to the eye. If only it had a mounting that was on the level of the rest of the cooler! I like that the fans come with pre-applied rubber dampers, and the included fan controller is a sweet freebie. The manual is clear as far as the installation process goes and all of the parts are well-described for both Intel and AMD installation types. For what it's worth, it performs well but not to where I would like it to be for it's price point. The numbers that I recorded in the results section are after 4 or 5 individual adjustments (even counting and trying to match the number of turns on each nut at each corner) to even the pressure at the four corners as best as I could, so no, I did not give up easily! I'm not being a snob, I just know what's out there and what works well versus what doesn't. And trust me, proper designs aren't any more expensive to manufacture. The main flaw with this mounting is that the sprung nuts (as I'll call them) do not bottom out soon enough (and the fact that it's 4-point instead of 2 or even 1-point). The nuts bottoming out at the same place on each nut mean that there will be even pressure on all four corners. Unfortunately, these nuts have to be tightened down far too deeply in order to bottom them out and will cause permanent deformation (I already witnessed quite a lot of deflection before I even came close to bottoming them out) of the hold-down bracket — something I was not willing to do.

As I said at the beginning of the last paragraph, there is a bright side to this cooler and what is provided and what comes with it! The only thing I could wish for is that manufacturers out there would understand how to make a proper mounting bracket. There are designs out there that work that cost hardly anything different to build and are miles (kilometers for a lot of you) more effective.