Thermaltake Silent Boost RX-K8



Thermaltake has expanded their product line tenfold in the last few years. A company which originally gained fame with its introduction of the original Orb HSF has now grown into an OEM giant. Thermaltake's product line now includes the following: CPU air-coolers, high-tech/sever style chassis, power supplies, innovative liquid cooling, silent DC fans, several storage devices, a cornucopia of accessories and a introduction of Desktop Media solutions. Thermaltake's rapid growth and cutting edge inventory exemplifies this company has thrown a substantial amount of profit right back into R&D. In my assessment this is Thermaltake's formula for success.

Today we test the RX-K8. This HSF is an all copper cooler based on a reverse flow fan designed to produce maximum air flow with minimal noise. Sadly this cooler has been met with mixed reviews many of which found the “Silent” in Silent Boost to be lacking. Unfortunately this may be one case in which this heatsink's shortcomings mask its true potential. In their desire to build an active cooler featuring near silent operation the heatsink had to be extraordinary in its ability to transfer and dissipate thermal energy. This isn't so much because of the fan's 38CFM rating, there are lower out there of course, but the fan's design almost excludes a "shroud" focusing the air in a more concentrated fashion. If I've done my job correctly we'll discover two important facts which have plagued this particular heatsink and the reviewing process on the whole where noise levels are concerned.

First, criticizing Thermaltake's claim this heatsink-fan emits just 15dBA using an SPL meter in any environment other then an anechoic chamber would be analogous to criticizing the Weather Bureau's data on a Hurricane flood surge by measuring the overflow of your swimming pool.

Second, we cannot expect any degree of accuracy so long as there are no criteria or industry wide standards for sound level tests. After spending years in the High End Audiophile industry there is one thing I know for sure, the number of research level anechoic chambers around the world can be counted on your hands (addition only *hehe*).

As an example, the audiophile loudspeaker industry had been dominated by a single pair of studio recording monitors which are still found today at Abby Road Studios, B&W 801 loudspeakers. These were the most coveted transducers by audiophiles whom could afford speakers costing 5x as much, although B&W 801's sold at around $5,000 and thousands of 20-year old pairs remain in audiophile systems today. The audiophile world saw some dramatic changes between 1980 ~ 2000 with some truly exceptional speakers originating from Canada. The sole reason behind this movement was due to Canada's rapidly expanding NRC Facilities which was built to keep the Canadian economy on the cutting edge and Canadian business competitive in the world market. In particular its anechoic chamber (seen in the article below) has contributed to the success of Canadian loudspeaker companies, such as the relationship described in the following article about Axiom (speakers) and the NRC.

How does all this fit into a heatsink review? In the field of psycho-acoustics our test methods are rudimentary at best and this is not to say we should blindly trust what each company claims in their specifications, only we should be as skeptical of our own. Certainly when we make observations as “Technical Writers” there are those whom entrust in those observations, which is why the description below which violates a few laws of physics is better left unsaid.

As soon as I started up my system the first time, I knew there was a problem. It was the loudest heatsink I’d ever had. It wasn’t so much that the fan was loud; it was that it was vibrating my whole case, and it acted like a sub woofer. I couldn’t leave it running, so I took it out and looked for the problem....

Packaging



As is evident from the photo above the RX-K8 is a fairly compact HSF compared to it's sibling the Blue ORB II which I reviewed here back in January. Admittedly the ORB which supports a 120mm fan is an extra large HSF. RX-K8 packaging is atypical Tt, ready for retail display at your local Computer Shop. The package includes an instruction manual, thermal paste and the RX-K8 itself which is intended to accommodate the Athlon 64, Athlon 64 FX, Opteron and Sempron on Socket-939 / 754.

RX Factor



All fans produce two types of unwanted noise:

1.) Mechanical - caused by the motor, bearings, or blade imbalance.
2.) Aerodynamic - resulting primarily from turbulent air as each fan blade passes over the mounting struts between the fan blades and heatsink cooling fins.

The struts are necessary to secure the motor and blade assembly to the outer shroud. In fact the shroud itself perpetuates turbulence, ironically by doing its job which is to focus airflow down onto the cooling fins. Air from the trailing edge of each blade creates further turbulence which the shroud then refracts back into the environment. Noise created by the spinning blades is characterized in frequency or pitch dependent upon the fan speed, the number of fan blades and the number of struts the fan blade passes over. Multiplying fan speed in rpm/60 (revs per sec) times the number of fan blades times the number of struts each fan blade passes over in one revolution determines frequency of pitch in cycles per second or hertz (Hz).

For example: a fan running at 5000 rpm with 7 blades rotating over 3 struts generates a frequency of 5000/60 x 7 x 3 = 1750hz



Reverse Flow may read like another marketing ploy; however, the design is proven, pragmatic and used throughout the industry. I must confess when I reviewed Arctic Cooling's Slinetium-2 PC-case which utilizes reverse flow technology, was completely unaware of the science behind the fans, which worked well and were very silent. The design works to eliminate aerodynamic noise by eliminating strut turbulence and minimizing shroud turbulence. If there are deficiencies in this design they most likely result from the absence of a full shroud concentrating the air-flow where it needs to be.




Specifications



The RX-K8 fan is rated at 38.7CFM which isn't the greatest amount of air-flow, however; where silence is a goal sacrifices must be made. Removing the fan assembly shows the absence of an atypical shroud. Further details can be found by clicking on the thumbnails below.





Looking at the exposed mounting base and pseudo shroud reveals mounting holes for an 80mm fan, although I fail to see the need for this as it would inevitably increase noise and decrease air-flow compared to its included 92mm fan. I intentionally left the dust particles in place to exemplify the "dead zone" which is the Achilles Heel for all fans mounted in this fashion, albeit reverse flow or the usual fan design. Design alternatives are numerous including the use of vertical mounted fans which move air across a fin assembly in what are much taller units, these are also known as tower coolers and will often integrate heat-pipes. Even with its Achilles’ Heel it seems mounting the fan so air is forced down onto the fins and base plate works best.





The fan mounting enclosure encasing the fin section seems to be made of aluminum, the fin and base plate section is constructed of copper with the exception of an aluminum "riser" bar. This center section which screws into the base plate acts as a spacer between the mounting retention bracket and base plate, it’s there to keep the mounting clip in place.






The retention clip-style mount is about as simple as it gets providing solid footing for the copper base plate and adequate pressure for contact between base plate and CPU IHS (Integrated Heat Spreader).



Finish is always a critical issue as the thermal paste provides a microscopic layer which must fill every striation on the base. If the finish is too rough there's bound to be air-gaps and while their minute pockets they do nonetheless raise temps.

TT’s base-plate finishes have been relatively consistent which is a good thing; however, their finish is representative of a mass production process. I would say this finish is "just" acceptable and if I were to live with this unit for any length of time would most likely lap it. Spending an hour lapping the base plate will lower temps and effectively raise the performance value of your $25 HSF.



After removing the RX-K8 note the perfect spread of the thermal paste. As proven more effective, I applied a single dollop or rise sized grain of thermal paste to the IHS, then mounting the heatsink as usual. This method uses the tension of the mounting hardware to "compress" and naturally disperse the thermal paste rather then "spreading" the paste. Spreading thermal paste will most often incorporate air-pockets into the paste.



Although unlikely in a cooler at this price range, just once I'd like to see an HSF with a base plate recess able to accommodate a thermistor. Since thermistor sizes vary ever so slightly this can easily be solved with the manufacturer purchasing in bulk as usual and mate thermistor to recess. Even more effective would be an impregnated base plate fabricated around the thermistor such that the thermistor is located just a mm or so above the processor’s IHS (Integrated Heatsink) and core. When it comes to thermodynamics in the world of micro-electronics, a few mm's distance from the surface of the processor to its core would be akin to measuring the distance from the Earth to the Sun using a Yardstick.

Thermal Transfer



Thermal Conductivity: For our purposes a more appropriate definition will be Heat Transfer by Conduction. This is the transfer of kinetic energy (heat) through a material without any movement of the material. The rate at which heat is transferred through a material depends upon the temperature gradient and the conductivity of the material itself, for example copper possesses higher thermal conductivity then aluminum. Another term pertinent to this section; Specific Heat is defined as the amount of heat required to change a particular unit of matter by one degree in temperature.

The RX-K8 is made of copper which offers the most cost effective balance of specific heat and thermal conductivity. In order to take full advantage of copper's thermodynamic properties the material must be formed and constructed in just the right shape. Many ingredients combine for just the right thermal recipe, base-plate thickness, surface condition, number of cooling fins. Cooling fins must be formed at a specific thickness, length, width and surface texture to take full advantage of the air passed over them to dissipate heat effectively. Fins are not the only surface shape used for this purpose, a few HSF manufacturers including Alpha and Swiftech made extruded pin design HSF famous with their PAL8045 and MCX-478V (.pdf) respectively. Fins seemed to be the surface structure of choice over the last few years and Thermaltake's RX-K8 features 46-machined surface cooling- fins press-fitted into a copper base.



Heat is kinetic energy or the collision of molecules and as explained above since the heatsink is a fixed object it doesn't move, nor do the molecules it is composed of except in relation to each other. In other words rapidly "vibrating" copper molecules do not "travel" from the base-plate to the fin surface, what does travel is the "energy" as each molecule collides with its neighbor. Since heat usually travels from a higher temp region (faster vibrating molecules) to lower temp region (slower vibrating molecules) heat from the CPU is conducted through the base plate, up through the fins where forced air has "cooled" these fins, the process is perpetual whether or not the fan is on or not.



A critical aspect of forced convection cooling (air) is a concept known as laminar air flow. Read any number of PC-forums or reviews pertaining to cooling and you'll find the term kicked around more then a Hacky Sack at a Hemp convention. The term is often used out of context as Laminar airflow is one of two types of airflow interaction (or molecular interaction) as it passes over an object. Air molecules closest to the surface tend to adhere to that surface, air molecules in motion above the first "stagnant" layer colliding with it travel slightly slower. As these layers rise further away from the surface their speed increases as resistance decreases. This pattern continues until the air-stream is uninterrupted. Problems can arise where the air-molecules come in contact with a surface as these act as a Boundary Layer. Semantics behind the name boundary layer are apropos since this layer can have a deleterious impact on thermal transfer. The chart below summarizes much of this explanation.



The theory behind cooling fins is very simply to provide as much surface area as possible in the least amount of space. While this statement might seem antithetical it explains why the fins are placed so closely together, often just a few mm's apart. The unit is svelte compared to some of the LGA-775 behemoths weighing in at 530gr to 938gr (or 2lbs). On the DFI LAN Party these usually require through mount hardware to safely secure the unit to the mainboard and of course this requires removal of your mainboard which basically involves disassembling your entire system. The mounting hardware as mentioned on the previous page uses the stock retention mount (RM) making installation an ease.



From the photo above you can see DIMM-1 looks to be close, this is an attribute of the LAN Party UT Nf4 mainboard layout and not the RX-K8 which would be considered svelte compared to today's monsters. Turning our attention to the photo below the heatsink leave plenty of room for the memory and uncommonly tall sticks such as Corsair XMS Pro or XMS XPert should be fine.



The RX-K8 delivers what is commonly referred to as secondary air-flow. Thermaltake didn't set out to reproduce a cooler such as Thermal Integrations Dr. Thermal looking at the T.I.T.I.'s base-plate below the CPU contact area or "foot-print" is a minimalist design. Where most heatsinks have a large rectangular base-plate into which the fins are pressed, Dr Thermal minimal foot print allows forced air from the fan above to freely flow down onto the socket area and the devices in close proximity to the socket.



Comparing Dr. Thermal above, to Thermaltake below, we might say the former is a design in which aerodynamics plays a greater role in the heat transfer process while Thermaltake uses a straight forward approach with the exception of its inverted fan for silent cooling.

Testing the RX K8 Silent Boost involved several comparisons. I installed the cooler onto a DFI LAN Party main board and Opteron 148 retail processor. The main board and CPU combo had been tested in the Titan Robela water-cooled case (reviewed here). This allowed me to compare both the Robela and RX-K8 in cooling ability and in silent operation, upon which the Robela is partly based. Below the RX-K8 installed on the DFI and in the TITAN Robela.

AMD Test System
CPU	Opteron 148 Retail Socket-939
Mainboard	DFI LAN party UT nF4 Ultra-D
Memory	GSkill HZ PC4000 (2x1024MB DC CL3-4-4-8)
Graphics	Gigabyte GV-NX68T256DH 6800GT
Power Supply	InWin IP-460 460W
Cooling	TITAN Robela water-cooled case Thermaltake RX-K8 Silent Boost AMD Stock Cooler for Opteron S-939
Operating System	Windows XP SP2/3

In addition to comparing the RX-K8 to the Titan Robela and AMD stock cooler I switched out the "reverse flow" 92mm fan with a standard 90mm Thermaltake model TT-9025A. Specifications for the TT-9025A can be seen below.



Compared to the RX-K8's reverse flow fan (PL-2096) which displaces 38.7CFM, the TT-9025A-2B displaces 56CFM. It would seem logical the TT-9025A-2B will produce lower temps and a higher noise level. Below I've installed the TT-9025A-2B installed on the RX-K8.



Test Methods

As mentioned above the RX-K8's performance was tested on the DFI/Opteron combination measuring the CPU core temp using the processor's on-die thermal diode as displayed through Smart Guardian. In addition I inserted a thermistor coming from Robela's on-board temp-monitor LED, between the processor IHS and Heatsink base plate a few mms so not to disturb mounting pressure. The same was true of the Robela's water block. Absolutely no changes were made to the DFI / Opteron test system so that the only variables throughout this test were changing out the Stock AMD cooler for the Robela's water block and exchanging fans on the RX-K8.

To produce a full LOAD effect I used the CPU stress test program S&M 1.7.6. Beta since this test, the version is now 1.8.1.

Ambient temps were consistently between 18~20°C



Sound Level Tests

Since the RX-K8 Silent Boost title denotes quiet operation, sound level tests were indicated. As I stated earlier this was the primary reason I chose not simply to compare this cooler to a H20-system but actually installed the RX-K8 into the TITAN Robela enclosure, thereby minimizing the number of variables. To measure noise levels we use the Intellisystems AR824 SPL meter at a distance of 50cm. Ambient noise level was measured and fixed by turning off all sources of noise and taking measurements during the evening hours. For our purposes the figures below show dBA levels over ambient.



Thermaltake has utilized the fan found in our RX-K8 with the Mini-Typhoon which our Editor reviewed here. While I didn't exactly find the RX-K8 to be dead silent I still felt it met the criteria for "silent" cooling. Of all the reviews we researched on the RX-K8 and Mini-Typhoon there were quite a bit of mixed results.

As I stated in the introduction, the reviewing/testing community on the whole lacks the capability to perform truly accurate, objective sound measurements. This goal will be nearly impossible to achieve without constructing an anechoic chamber. For those interested in the cost and complexities involved in building an anechoic chamber, this article on McIntosh anechoic chamber is a MUST read. If there are degrees of accuracy among anechoic chambers (which there are), imagine the inaccuracy among ad hoc tests performed in most reviewer's homes? For this reason the best method I could come up with was to measure the ambient noise in my room then record the increase in sound (if any) the device in question produces. If your ambient sound level measures 30dBA and you find a particular HSF emits 60dBA , essentially your proposing the HSF being tested emits 30dBA above ambient. Since this is always an approximation based on an infinite number of variables we should treat these figures with relative skepticism.

Summary

PRO
Silent Operation.
Decent Overclocking ability.
Easy Mounting system.
All copper construction.

CON
Mounting Limited to AMD.
A more powerful fan with rheostat would be preferable.

Price at time of writing: Thermaltake's RX-K8 Silent Boost can be found through FROOGLE at numerous shops for as low as $27.00.

Conclusion

Thermaltake RX-K8 Silent Boost is an HSF which offers a slightly higher level of performance under slightly less noisy operation. As we've seen, several reviews had the RX-K8 producing 40dBA+ running flat out, which is fairly noisy. This is also the reason I choose to eschew testing and measuring performance using a motherboard fan-controller program. Had the RX-K8 came with a rheostat I would have measured temp/noise-levels at different speeds. As it stands the degree of variance from one motherboard BIOS fan-controller to the next is simply too great for my figures to be applicable, not to mention the number of people who don't even bother to enable these programs.

The RX-K8 as is, won't appeal to Overclockers although I was able to overclock the Opteron 148 as high as 2750MHz. Since this involved raising Vcore I've only included overclocking data at default Vcore. In the final analysis I believe the tendency to focus on the "Silent" in Silent Boost leads us astray from the real benefits this cooler has to offer. Thermaltake has built a high quality copper cooler with substantial surface area; proof of this can be found where its original fan was substituted.

Can I recommend this cooler? To be honest it depends on what you’re looking for, there are better alternatives at the same price in silent category and in the overclocking category. If this HSF were sold with a more powerful fan and a rheostat I wouldn't hesitate to recommend it, if you can find the RX-K8 for under $25 I would recommend it.

Questions/Comments: forum thread

From the 13 reviews found on the Mini Typhoon and RX-K8 (which both use the same fan) we gathered 4 ? noisy ? user experiences and 9 ? silent ? ones. Could this be solely due to sample variance? Not all reviews used the same method to evaluate the acoustic performance, those that used a dBA meter found the fan to be clearly audible. The silent reports came from those reviews where there was no accurate measurement of the noise generated.

Mini Typhoon

Hartware - ?@12v 44,9 dB(A) sehr deutlich h?rbar (very clearly audible)?

TechPowerup - ?@12v: 56dBA @7v: 39dBA (very loud)?

Overclockercafe - ?Quiet operation? but no more detail

Motherboards - ?nearly silent? but no more detail

Madshrimps (CPU roundup) - ?@12v: 51dBA @7v: 36dBA (very loud)?


RX-K8

Virtual-Hideout - ?slightly over ambient?

Tweaknews - ?virtually silent at maximum speed?

aphnetworks - ?I can hear it very clearly?

Futurelooks - ?I also noticed the PC got a tiny bit more noisy?

moditory - ?incredibly quiet operation?

modthebox - ?This is a great device for low noise applications?

techtastic - ?a really quiet operating heatsink/fan combo?

hardware-review - ?runs silently as advertised?