Scythe’s e-Otonashi fanless EPIA-M cooling case

Table of Contents

 

Feb 23, 2004 by Mike Chin

Product
e-Otonashi fanless mini PC case
Manufacturer / Supplier
Scythe Co., Ltd
Price
US$198

Scythe Co., Ltd has been the source of some unusual cooling products. Most recently, SPCR reviewed the Heatlane Zen NCU-1000, a passive 6″ tall CPU cooler that used a heatpipe as a core design element. In the e-Otonashi, Scythe has turned again to a Heatlane heatpipe, this time in a somewhat more ambitious product: A compact case for VIA EPIA-M Mini-ITX boards that offers a fanless CPU cooling system as an integral part of its design.

The resulting system is completely fanless, because the power supply makes use of a small external brick style transformer and a small DC voltage regulation circuit in the case, neither of which have fans. The only source of noise in this system comes from whichever 2.5″ notebook drive (and slim notebook optical drive) you choose.

If all this sounds familiar, it should be: The e-Otonashi is a kind of poor man’s do-it-yourself kit version of the recently reviewed Mappit A4F, a small, prebuilt fanless, EPIA-M based PC that uses a notebook drive. There are distinct similarities, but also many differences. More on that later.

FIRST IMPRESSIONS

The review sample came in a sturdy carton illustrated with some drawings and promotional information about the product inside. It was well packed for shipping. When unpacked, the contents consisted of the following:


The carton contents were: Chassis, PSU, Installation Manual, AC Adapter, cables for a slim optical drive and a notebook drive, and set of screws and hardware.


A shot of the small brick external PSU: Works off 100-240VAC; output is 12V, up to 5A (60W).

*

SPECIFICATIONS
Model Name
e-OTONASHI (also Otonashi-1000)
Manufacturer
Scythe Co., Ltd. Japan
Motherboard Compatibility
VIA EPIA-M (Mini-ITX)
VIA EPIA-M10000 (Mini-ITX)
Chassis Dimensions
280(W) x 190(D) x 90(H) mm
Weight
2350 g (System: 3200g, approximate)
Drive Bay
2.5″ Drive Bay (1), Slim Drive Bay (1)
Power Supply
60W A/C, Comply with CC & FCC
I/P: 100V ~ 240V, 50-60Hz, 1.8A Max.

O/P: 12Vdc, 5.0A
SPS Adapter Included

Noise Level
0 dB

Possible system configurations and approximate component costs in US$ (based on Pricewatch):

Component
US$ Price (approx)
e-Otonashi fanless mini PC case
$198
VIA EPIA M6000
VIA EPIA M10000
$145
$200
Fujitsu 40G 4200RPM 2.5″ HDD
Fujitsu 80G 4200RPM 2.5″ HDD

Seagate Momentus
40G 5400RPM 8MB Cache 2.5″ HDD

$120
$215
$150
Samsung Slim 8x8x24 CD-RW with 8X DVD Combo
Toshiba SD-R6112 Slim 4X DVD-R/RW
$80
$174
256MB PC2700 DDR SDRAM
512MB
PC2700 DDR SDRAM
$45
$75

So, the most budget system including M6000, CD-RW, 40G HDD and 256MB memory would be well under $600 (at least, in the US). The most high-end system with M10000, DVD-RW, 80G HDD and 512MB memory would be well under $900. The price could drop substantially if a non-recordable or no optical drive and a smaller HDD were considered.

Here is a closer look at the e-Otonashi, from the front and from the back. It is made entirely of aluminum. All of the exterior has a pebbled finish which hides fingerprints much better than these flash photos suggest. On the lower right of the front panel, there is a round power button and a reset button accessible through a very small hole. The rectangular outline area is the cover for a notebook-style optical drive. There are grill openings on top, the front, the sides and the back. A small input jack for DC voltage is visible on the back. In case you didn’t pay close attention to the specs above, the case is small. The metric dimension are given above; it is just 11″ wide, 7.5″ deep and about 3.5″ tall.

Please Note: The black finish of this case obscures details in photos terribly, forcing much reliance on digital image editing in Photoshop. As a result some of the colors may seem a bit off. This is the price for getting the visual detail you’d want to see.

THE COOLING ENGINE

The really interesting part of the e-Otonashi is the bottom, however. It is what it looks like: The entire bottom of the case is a massive black-anodized aluminum heatsink. Two of the fins, one at the front and one at the back, have a hard rubber sleeve over the entire length: These act as the feet, and elevate the rest of the fins up a few millimeters, presumably to aid in air circulation and avoid possible heat damage to the desk surface.

Three thumbscrews on the back hold the cover in place. The cover is fairly thin. Here’s what the interior looks like.

The top piece with the circular cutouts is the mounting bracket for both the hard drive and the optical drive. It comes off with the removal of 4 screws on the side frames.

With the mounting bracket removed, the core CPU cooling engine is revealed. It is this silver colored J-shaped aluminum bar, clamped with brackets on the ends to the flat surface of the case bottom, the big heatsink. It is an aluminum heatpipe with a flat profile.

Here’s a closer look at the heatpipe.

And the mating surface of the aluminum heatsink below it.

Only the heatsink surface beneath the heatpipe is exposed. The bottom sheet metal of the case has a cutout just big enough to fit the the heatpipe.

QUESTIONS ABOUT HEATPIPES

Just how this heatpipe is affixed to the CPU on a Mini-ITX board will be shown in the following photos. In the meanwhile, consider the flat heatpipe itself.

TS Heatronics, who manufactures this heatpipe, says their technology is a significant improvement over conventional heatpipes. A conventional heat pipe transfers heat through a cycle of vaporization, transportation, and then condensation of a working fluid with a low boiling temperature. It depends on gravity to allow the condensed water to flow back down to the heat source, and thus must work with the evaporator (and heat source) at a lower point and the condenser (radiator or heatsink) at a higher point. According to TS Heatronics, their…

Heatlane technology is called meandering capillary tube heat pipe, or self-excited oscillation (pulsation) heat pipe or Akachi pipe named after Hisateru Akachi, the inventor who is former vice president (now consultant) of our company. The technology is, with its epoch-making features, attracting deep concern of companies, universities and research institutes all over the world, resulting in the award of the technology prize of the Japan Association for Heat Pipe for the year 2002… The technology realizes effective heat transfer conducted not only horizontally but also from top to bottom (top heat) as well as better heat transfer capability than the conventional heat pipe.

Before assembling a system in this case, I thought about how I could witness the heat transfer property of this heatpipe first hand. The answer came to me when I was boiling some water for a cup of tea.

I found a piece of light architectural aluminum channel in my workshop and cut it to the same length as the flat heatpipe from the e-Otonashi. I poured piping hot, just boiled water into a large coffee mug. I placed the aluminum channel piece in the cup of hot water and held the end with my thumb and index finger. In about 10 seconds, it became too hot to hold. The aluminum channel was removed, and the TS Heatronics heatpipe was put in its place. It took only 2~3 seconds before the end became too hot for me to hold! This was a dramatic and simple illustration of the thermal transfer power of the heatpipe.

The same thing was also tried with a glass of ice cold water. Again, the speed with which the heatpipe became freezing cold was astonishing. My only regret is that I do not have any conventional heatpipes to compare against the Heatlane version.

SYSTEM ASSEMBLY

Attaching the Heatpipe to the Motherboard

A VIA EPIA M10000 was chosen to go into the e-Otonashi case. It is the hottest of VIA’s CPUs, and one specifically approved for compatibility in this case. The CPU cooler was removed by popping out the two spring-loaded pins and pulling gently to loosen off the HSF from the hardened thermal goop.

The CPU and the heatsink base were cleaned off as much as possible with an old credit card, pure alcohol, tissue paper and cotton swabs. You’ll see that some discoloration remains on the outer surface of the CPU die above, but it cannot be felt with the fingers. Although some thermal interface material was provided, Arctic Silver Ceramique was substituted instead.

The next step, of attaching the looped part of the heatpipe to the CPU, and then reattaching the original CPU cooler on top of the heatpipe, was too fiddly and convoluted to capture in photos. It is a rather delicate operation requiring a lot of manual dexterity best done with three hands, but I managed with two without dropping or breaking anything. Arctic Silver Ceramique was used on both mating surfaces. The two screws that hold the entire assembly together seem a bit too small and fine to me, but I guess they are sturdy enough for the low weight and relatively low pressure they are subjected to. If the heatpipe had been 1~2 mm narrower, larger, easier-to-handle screws could have been used.

The end result is that the short looped portion of the heatpipe is pressed tightly between the CPU below, and the original CPU cooler above. The long part of the heatpipe goes underneath the board and extends out about a centimeter past the opposite edge of the board.

One question was whether it is possible that the heatpipe below could make contact with the circuit board traces and short anything out. The answer is a firm no. There are 3 reasons:

1) The motherboard is mounted on risers that place the bottom of the board several millimeters above the bottom of the case.

2) The heatpipe is clamped to the case bottom heatsink so that it is pulled away from the underside of the board.

3) There are two spring-loaded plastic pins (for the big chipset heatsink in the middle of the board) along the path of the heatpipe under side. The ends of these plastic pins stick out far enough that they prevent any contact between circuit traces and the heatpipe. (See photo below; the yellow arrows point to the pins.)

Installing the Heatpipe-equipped Motherboard

There was no mention in the instruction sheets about putting TIM between the heatpipe and the big heatsink, but I thought it would probably improve the thermal transfer function, so I applied ordinary silicone goop. It seemed too large an area to apply the expensive Arctic Silver stuff.

The motherboard was installed with four screws to the tapped risers on the bottom of the case, and the heatpipe clamped to the big bottom heatsink. Finally, the ATX cable from the PSU was connected to the motherboard.

Installing the 2.5″ Notebook Hard Drive

A Seagate Momentus 40GB, 5400 RPM notebook hard drive with 8 MB cache (ST94811A ) kindly loaned by SPCR sponsor Frontier PC in Vancouver was used for this system. A notebook optical drive was not installed; instead I used a USB external optical drive for software installation.

Although a data cable was provided for connecting up a slim optical drive, no data cable for the HDD was provided, only the notebook to standard IDE adapter. This may have been an oversight or a simple anomaly in packaging. I borrowed a folded, short HDD data cable from the recently reviewed AOpen XC Cube SFF PC; this cable was perfect for the application.

The multi-connector DC power cable on the bottom left in the photo above is provided with the e-Otonashi case. I decided that this was too messy to use when all I needed was a single 4-pin connector and only the 5-volt line, So I scrounged through the parts bins once again and came up with a quick mod that would do the job more tidily: This is the 2-lead cable in the bottom right of the photo.

The hard drive is shown mounted in the photo above, with four screws from the underside. There are 2 small thin pads affixed to the plate beneath the HDD; these appear to be mostly for thermal conductivity.

Here’s the end result, all neatly assembled, just before the cover was put back on.

PERFORMANCE & NOISE

Rather go through the details of the VIA EPIA M10000, I refer you to the linked review of the integrated board. It is a fine PC for the great majority of PC users who are non-gamers, with plenty enough power for office-type apps, web-browsing, email, etc.

Windows XP Pro SP1 installation went fine with the external USB optical drive, although as with other M-ITX setups I have done, installation of the enormous number of XP updates and fixes can be excruciatingly slow. I believe this has something to do with the system scanning most of these updates do as they’re being loaded. There is something the VIA boards don’t like about Microsoft’s procedure. After it was all done, I spent a few days using the system off and on to get a good feel of it before starting to measure, analyze and write.

Cooling

The following tools were used to test the thermal performance of the system. Links to all of this software can be found in SPCR’s Software section of Useful Web Links:

  • CPUBurn processor stress testing utility
  • DTemp hard drive monitoring
  • Motherboard Monitor 5 motherboard monitoring utility

One other testing device is a Kill-a-Watt AC power meter. The ambient temperature during testing was 22°C.

Activity
CPU
System
Hard Drive
AC Power
Idle
33°C
29°C
40°C
23W
CPUBurn
72°C
34°C
41°C
40W
CPUBurn2*
72°C
36°C
38°C
40W

*Case turned on its side, for vertical orientation, HDD on bottom.

The only temperature that’s a bit high is the CPU; as mentioned in the past VIA CPUs have a reputation for excellent toughness and performance at high temps, and 72°C is nowhere close to a danger zone for the VIA CPU. Both System and HDD temps are very modest, probably due to the generous ventilation that the case features on five panels.

It’s interesting that the CPU temp remains completely unaffected when the case is turned on its side, which tends to substantiate TS Heatronics’ claim about their heatpipe not being affect by gravity. My guess about the System temp rising is that the sensor for this is “downwind” of the CPU HS when the case is turned on its right side. In other words, the heat from the CPU is rising up to that sensor, and causing the System temp to register higher. A similar effect is happening with the hard drive: The heat from the internal PSU circuit no longer comes straight up from below it, but rises towards the CPU and motherboard area in this vertical setup.

Here’s a comparison against the similarly outfitted prebuilt Mappit A4F PC:

System
CPU
System
Hard Drive
AC Power
Mappit A4F
70°C
41°C
53°C
41W
e-Otonashi
72°C
34°C
41°C
40W

The CPU temps are close enough that variances in CPU samples and assembly could reverse the numbers, so I’d consider the CPU cooling of the two systems equal in performance. The differences in System and HDD temps are easily explained: The e-Otonashi’s cover is very openly ventilated while the Mappit is completely enclosed.

Noise Analysis

The parts in my sample of the e-Otonashi don’t make ANY noise. The PSU circuitry is dead silent, and if the power brick makes any hum, it is too low to be noticed. There are no fans, no hum, no tinkling of water — no noise at all from either the cooling system or the PSU.

The point is here simple: Whatever noise comes from a Scythe e-Otonashi system will depend entirely on your choice of drives. Choose a quieter drive and you will have lower noise.

The noise level of the system with the Seagate Momentus hard drive was very low, but marred by the drive’s high frequency whine, which is clearly audible from several feet away in a quiet room. Missing is the lower frequency hum that emanates from any PC with a 3.5″ hard drive mounted normally to the chassis. The vibration level of this notebook drive is minuscule in comparison to any 3.5″ drive, so the difference in vibration-induced chassis noise is dramatic. Most of this noise occurs in the low frequency area around 100~120 Hz.

The Seagate Momentus is not in the same noise class as the amazingly quiet Fujitsu notebook drive in the Mappit A4F: It has quite a bit of high pitched whine, and its idle whirring is higher than that of the Fujitsu. It also seems to go through fairly frequent heat-resetting types of noises, a bit like the Hitachi/IBMs, but much softer.

Naturally I had to try using the whisper quiet Fujitsu MHT2040AT 40G 4200 RPM single-platter drive (from the Mappit A4F) in the Scythe / Heatlane e-Otonashi.

The result was noise performance maybe 2 dBA louder than the vanishingly quiet Mappit. The very quiet Fujitsu drive does not change its noise output, but the open vents all over the case of the e-Otonashi allows much more of the acoustic noise to get out of the case. The all-closed Mappit blocks more of the acoustic noise from getting out.

Rough Noise Measurements

I decided to try and measure the sound pressure levels. The testing was done in the lab, an acoustically reflective converted kitchen, at around midnight when ambient noise is low.

The actual SPL reading in the room with all noise producing equipment turned off was ~16 dBA, which unfortunately is still too high for really accurate measurements of these systems. The SPL readings of the gear were only marginally above the ambient level, so their accuracy is definitely not assured. Good acoustic practice requires that the ambient noise level be at least 6 dBA below the SPL of the noise source being measured.

  • Turning on the e-Otonashi system with the Seagate Momentus drive raised the noise level to 21~22 dBA at 1 meter distance from the unit. It is very quiet except for the high frequency whine, which is annoyingly, easily audible.
  • Turning on the e-Otonashi system with the Fujitsu drive raised the noise level about 2~2.5 dBA, to 18~18.5 dBA at 1 meter distance from the unit. It is audible but extremely quiet.
  • Turning the Mappit A4F on (w/ Fujitsu drive) raised the noise level just 1 dBA, to ~17 dBA at 1 meter distance from the unit. The overall noise signature is similar to the e-Otonashi; it is the same drive being measured. But overall, the sound is quieter still, more subdued and muffled.

Obviously, there is a tradeoff between heat and noise in any system. In the e-Otonashi, the case seems geared more for good cooling of added drive components than the lowest noise. According to Scythe, this is because in Japan, their first market, ambient temperatures can reach 35°C or higher in the summer. They wanted to ensure that the system would be able to stay cool even under such conditions.

NOTE: The manual has a comment about recycling the 40mm fan from the original EPIA CPU heatsink: There is a place for a 40mm fan on the back panel just behind the PSU circuit board. The fan from the original heatsink can be installed here for additional cooling if necessary.

CONCLUSIONS

The Scythe e-Otonashi integrated fanless CPU cooling case and PSU for VIA EPIA-M Mini ITX boards works exactly as claimed. It is completely silent: It cools the EPIA M10000 CPU well enough, and its power supply works noiselessly. Whatever noise from a system built around this product will come from the other components you choose to put in it — namely the drives.

This product shares with previously reviewed Scythe products a kind of fussiness, along with a quirky and inventive design sense, which is ultimately effective. Some assembly details might be improved. There is, for example, a question of how tight is tight enough for the two screws that sandwich the CPU between the original heatsink, the heatpipe and bottom heatsink. And the top front edge of the cover does not align right up against the back edge of the front panel, leaving a tiny gap that seems a bit glaring. The fit with the VIA board is precise, however, which is nice to see.

Given the somewhat fussy procedure for assembly and the loss of motherboard warranty that CPU heatsink removal probably entails, this product might well end up being sold in a prebuilt system and warrantied by some retailers or system integrators. This assumes that the product is more widely distributed than it is today.

The asking price of US$198 seems well justified, especially given the absence of any alternatives to an integrated noiseless case, cooler and PSU combination for the EPIA M boards. Just finding or creating a CPU heatsink that will cool EPIA M10000 CPU silently is a challenge; as far as we know, no one has offered an aftermarket heatsink replacement for this part (which uses a fan that’s too noisy for PC silencers.)

Pros
* Makes no noise at all
*
Fanless CPU cooling
* Fanless universal voltage power supply
*
Compact, efficient design
*
Good workmanship, fit and finish
* Assembly is not that difficult
* Fair price
Cons
* Doesn’t block HDD noise
* No front panel ports

* Only fits VIA EPIA M boards
* Uses more expensive notebook drives
* Instruction manual is a bit skimpy
* Assembly is not that easy

The e-Otonashi is definitely not a product for everyone. It is essentially an enthusiast’s case, for someone who is:

  • satisfied with the performance of EPIA M6000 or M10000 integrated motherboards — actually this probably covers 80% of non-gaming PC users.
  • handy and willing to assemble a system on their own as a kit, with some modding-type tasks that will void the motherboard warranty (though most people have friends who are technical and handy and willing to help with a project like this)
  • looking for a compact, minimalist PC that functions without extras — or noise

For its target market, Scythe have scored a bull’s eye with the e-Otonashi. Highly recommended — for the right buyer or user.

Our thanks to Scythe for the e-Otonashi sample and their kind support.

* * *

Comment on this article in our Forums.

Silent PC Review is reader-supported. When you buy through links on our site, we may earn an affiliate commission. Learn More

Leave a Comment

Your email address will not be published. Required fields are marked *