Often when I’m working a project, I’ll work through a series of ideas before settling on something that fits the bill. I try to write down these ideas and keep them around on the off chance that I could utilize them for a future project. Not long ago, I was working on a different computer project, assessing the cooling requirements, and I thought to myself, “It sure would be a cool to build a fully functional, small scale wind tunnel as a case for a computer.”
Unfortunately, I didn’t have time to investigate this wind tunnel-computer idea, and it sat around in some distant corner of my brain, until I decided to build a new computer. A medical research computer that would donate its time to cancer research.
In 2011 I became actively interested in grid computing, and specifically in World Community Grid.
The idea that I could build a computer, or use existing computer resources and donate their power so scientists and researchers could process medical and humanitarian research was extremely interesting.
By donating computer processing time, you actively contribute towards a great cause. World Community Grid has numerous projects available; finding cures and treatments for cancer, AIDS, malaria, muscular dystrophy, etc.
As I became more interested I found myself connecting all the computers I had in my home. Eventually I decided I wanted to do even more; I wanted to build a computer that could donate all of its time to processing medical and research data. That project culminated in the creation of the Lego Folding Farm. This system housed three separate computer systems, all running in one giant Lego case. The system went online in July of 2011 and has been running 24/7 ever since.
In the past year, World Community Grid moved closer to enabling a GPU (Graphics Processing Unit) based project, the Help Conquer Cancer project. GPU computing allows a project to utilize the full processing power of the GPU. In most cases, a project written to run on a GPU is significantly faster than its CPU counterpart. In the Help Conquer Cancer project, a single CPU workunit took around an hour for my other computer to run, where the GPU enabled version completed in less than 10 minutes.
While the GPU version of Help Conquer Cancer project was being developed, I felt personally challenged to donate more towards cancer research. I’ve had people in my family affected by various forms of cancer, and I’ve always felt like I wanted do more to help. But beyond donating money to cancer causes or doing cancer walks (which are also a great way to help show support) it’s hard for regular folks like myself to feel like I’m helping contribute. It seemed like perfect timing that World Community Grid was bringing onboard a project that would allow for significant increases in the speed at which cancer research could be completed. I felt very compelled to actively pursue adding more resources towards this cause.
My Lego Folding Farm is a farm of CPUs and was built to process numerous different medical and humanitarian projects. At the time it was built there was not a GPU enabled project that ran on World Community Grid, so I focused all of my efforts on CPU processing power. The GPUs that the Lego folding farm uses are nothing special. While they can process a GPU enabled project they don’t really have a lot of processing power for that task. With that in mind I decided to build a GPU processing system that would solely dedicate its time towards cancer research.
I investigated the expenses to build and run this system and realized that the component costs would be high, as were the added electricity costs required to keep the system running 24/7. I had already added a significant expense 2 years ago with the electricity required to power the Lego folding farm. I wasn’t sure I could afford the cost of the system itself, as well as another electric bill increase.
Unable to solve the funding problem on my own, I turned to my resident creative genius department @thetinnishflash and explained the issue. After some discussion, we decided to try and build this system with donations from friends, family and others that wanted to be involved in the fight against cancer.
We turned to the fundraising site Indiegogo and setup a project campaign to raise funds to defer the costs of the components as well as the electricity costs. After a couple of weeks working we were able to raise a significant sum of money, almost enough to cover the costs of the components themselves! Outstanding! I’m still so very appreciative to all those who donated and helped contribute to this project happening, without their donations it would have been a hardship to complete.
Getting the of support others, with friends and family coming together to help make something happen, was my favorite part of this project. It’s a great feeling being involved and getting to build something that others have contributed towards.
Once the fundraising was completed I began gathering up components for the build. I dubbed this project “Cancer Supercomputer” and I had fully intended on outfitting it with the components needed to be a GPU processing powerhouse. After hours of research I settled on a list of components.
CPU- Ivy Bridge 3770K
GPU- Radeon 7970 (Sapphire Dual-X cards)
RAM- Corsair Vengence 8GB 2133Mhz
SSD- Mushkin Callisto 40GB
PSU- Rosewill Fortress 650W Platinum
Motherboard- Gigabyte Sniper M3
CPU Cooler- Phanteks 140mm
My plan was to configure the computer and then overclock the hell of it. With this in mind, I started wondering what route I wanted to go as far as cooling. I debated going with water cooling, but decided instead to see how far I could push air cooling.
During the design phase of the Lego Folding Farm, I decided that the best method of cooling many hot components in a tight space was a very short path for airflow. Conventional computer cases generally intake air from the front which then travels through the case and rises upwards. Depending on the configuration there are numerous paths for the hot air to exit. There is almost always a rear exhaust fan, and in some cases a top mounted exhaust fan, as well. Additionally, some air will exit through the power supply fan. The problem I see with this method of air cooling is that the path that the airflow must take is long and winding.
To work around this, I built the Lego Folding Farm with a very short path from intake to exhaust. The intake fans sit directly in front of the components and blow air across the CPU cooler, the GPU cooler, the motherboards on both top and bottom, the heatsinks, the power regulation circuitry, the power supply. These are directly cooled by unobstructed airflow. The exhaust side is a mirror of the intake path, where air is directly exhausted out of the case. This method allowed for a much shorter path to get fresh, cool air in and blowing directly on the components.
However, with the Cancer Supercomputer I knew the graphics cards were very long and had very big coolers on them, so I needed to change my approach. I also knew from testing that under full load the GPUs could output a lot of heat, especially with multiple GPUs in a tight space. I needed a solution that allowed for more space, while also allowing me to move a lot of air.
Additionally, I wanted to pay close attention to air pressure. A conventional computer case has a lot of large “dead-space” areas. These areas don’t direct or force air to flow towards the components that are being cooled. I wanted to avoid dead space. I wanted the case to closely mirror the shape or profile of the finished motherboard with components. This way all the air flow would be forced to travel across the components surfaces, as opposed to flowing past large empty areas of case.
(My drawing skills are clearly unmatched)
Wind Tunnel Concept
Knowing I needed to maximize cooling, I explored a concept I’d tinkered with previously; a wind tunnel computer case. I had done research on wind tunnels back when I first investigated using a wind tunnel to cool components. There was a lot of information, much of it only helpful to those with a physics background. I’m not a physicist, but given enough time I can usually understand theories well enough to apply them in the real world and test them out, and that was the case here.
I used an anemometer (velocity/airspeed meter) for testing, as well as different fans and basic shapes made out of cardboard. I was able to conclude that I could increase air velocity through a scale wind tunnel.
There are two types of wind tunnels; the one I’ve built is a subsonic wind tunnel. This type of design involves a contraction section which is used to increase velocity (airspeed) through the test section. This increase in airspeed was what I looking for, a way to increase airspeed over the computer components.
There are numerous factors to consider when it comes to wind tunnel design and testing, things like Reynolds numbers, turbulence, boundary layer air, and heat from surface friction. I did extensive testing with various designs, materials and configurations until I settled on a final design.
In a perfect world I would have been able to build a larger wind tunnel. There were many constraints on this project to make it workable. The biggest of these was space. I’ve slowly converted my basement to a lab with toys, computers, Lego, robots and all kinds of various geekery. Therefore, I only had one area with enough space to construct a wind tunnel. The space itself was about 72″ long by about 26″ wide, and whatever I was going to build had to fit within those confines.
Next, I needed to fit computer components in the test section of the wind tunnel. I needed the test/contraction section to be big enough to accommodate the motherboard/CPU cooler/GPUs and also fit the power supply and a SSD. After tinkering with several designs I settled on a shape and size that allowed room for further growth. The contraction section of this wind tunnel has enough space for two full mATX motherboards plus the PSU. This allows for the possibility of adding another system in the future and cooling two systems inside one wind tunnel.
Next I had to contend with fan noise, as I would with any other air cooled pc. I debated using various sized PC fans and configuring them in a grid pattern at the inlet. Using fans around 220mm would’ve allowed for multiple fans at the inlet without too much noise. But when testing those fans I found that the airflow and CFM (cubic feet minute) produced weren’t adequate. At that point I changed direction and decided that I would use a single large fan with multiple speeds. This would produce a much higher airspeed as well much more CFM. I tested a couple of different fans and settled in on a box fan. It only consumed 96 mA and allowed for peak flow in the 2200CFM range with air speeds of 14-15MPH.
Wind Tunnel Construction
I looked at a couple of different options for building materials. I debated using sheet metal, and then looked at plastics. At one point I was going to make everything out of carbon fiber as I had some spare fabric, but the fact that carbon fiber is conductive dissuaded me, as the computer components are relatively sensitive to static discharge. In the end I choose wood, specifically 1/2” MDF board composite. It offered a nice surface to finish and paint and was strong enough to build the entire tunnel structure.
I had a fair amount of MDF board in my garage and that made for an excellent basis to start. I cut out the basic shapes for the inlet and outlet as well as the back panel of the contraction section. I also cut an entire piece that would serve as the base.
After gathering the pieces I screwed everything together to form the basic shapes needed to mock everything up.
Once I had the basic shape configured I went about tweaking and adjusting the layout of the pieces. Next, I moved on testing the inlet section with the fan to verify.
Then I moved over to working on all the finishing and edge pieces. I bought some cheap pieces of angle aluminum from the hardware store and cut and fit them at all the edges. I wanted a transparent section for the contraction/test section where the computer was going to reside. I used a couple of pieces of Lexan, cut to size to form a transparent window and top section.
Next, I moved onto prepping and painting all the pieces. I prepped the MDF for paint by using a sealer. From there I sanded the surface smooth before painting. I went for a darker metal color called anodized bronze. After some initial mockup with the painted pieces I later sanded them to achieve a smoother finish.
Once all the pieces were in place and the basic mockup was done I moved onto the wiring and gauge/switch assembly. I wanted a series of switches I could use to control the various functions of the wind tunnel and computer. I also wanted switches in place for future expansion so I could add a second computer.
I found some nice quality switches online, for cheap. There are 6 total switches. The first two switches with big red buttons are momentary switches that turn on the computers. (Only one used now, the second for a future computer)
The lever switches control both the fan and the power to the LED temperature gauges
The key switches are wired to the computer power switches and also to the fan switches. The key switches prevent anyone from turning anything off, but most importantly of all, I always wanted two key switches so that I could pretend I had a nuclear submarine that required two keys in order to launch. I always use my Sean Connery accent when using the key switches in order to maintain authenticity.
I also added some nice LED temperature gauges that I found online for cheap. They have remote probes and only require a 12v power source. I placed them on the ends of the panel so that I could easily monitor the inlet and exhaust temperatures of the wind tunnel. For power I used an old Dell 12v speaker power adapter and rewired it to the switch to provide power to the temp gauges.
The panel is a piece of aluminum that I cut holes in, to which I mounted all the gauges and switches. I’m pretty happy with how it all turned out. All the power, plug-ins and connections are beneath the switch/gauge panel.
All the angled aluminum is secured by 10-24 screws which I drilled and tapped into the MDF. It’s surprisingly a very strong combo. I would recommend using a step down on the drill size when you tap into MDF as it’s not as robust as actual wood. After what felt like days upon days of drilling and tapping I finally finished installing all the trim and screws.
The aluminum edging itself was polished with varying grits of sandpaper to achieve a brushed look. The edging at the bottom of the inlet and exhaust sections firmly holds both sections in place.
The aluminum braces with steel tubes that are on the top of the intake and exhaust sections are decorative. I picked up some extra scrap stuff from the metal yard for another $5 dollars. The stainless tubes contrast nicely with the brushed aluminum.
I debated buying a monitor stand, but since I was attempting to do this as cheaply as possible, I thought it might be a better idea to make my own. At the metal salvage yard I had purchased a junk, cutoff piece of aluminum for $5. Using that piece, I machined a stand to mount the monitor to, and then attached a tray I made for the keyboard.
The monitor itself was purchased on eBay and only cost $15 because it was broken. After adding a couple new capacitors it worked just fine. It’s a 20″ viewsonic.
Wind Tunnel Air Intake
I designed the air intake section to try and minimize fan noise. It seemed moving the fan deeper inside the intake section would cancel out a little bit of the blade noise. Additionally, I added a couple of layers of open grating which helped to minimize fan noise. With a decibel meter I measure around 65-67 decibels, which is significantly quieter than the 75-77 decibels that the Lego folding farm measures.
The fan sits about 4 inches inboard of the intake. I used some molding to form a radius at the inlet to smooth the airflow coming in around the inlet opening. I used some open “egg-crate” style plastic drop-ceiling panels, cut them into sections and placed them in front of the fan. For the intake opening I used more of the ½” angle aluminum, cut a bunch of pieces and placed them with the corner facing out. I spaced them very close together allowing a small 1/8” or so space between them.
This minimizes airflow restriction, because of the very large surface area of the intake inlet.
Using my anemometer (wind speed tester) I’ve been able to record airspeeds and velocities. The following numbers were measured with the fan set to its lowest setting
Inlet Air Speed- 0.6 MPH or .88 ft/s (Measured at inlet grill)
Exhaust Air Speed- 1.4 MPH or 2.05 ft/s (Measured at exhaust opening)
Fan Output Air Speed- 5.0 MPH or 7.33 ft/s (Measured free standing, no tunnel)
Contraction Section (Without Computer) Air Speed- 12 MPH or 17.6 ft/s (Unobstructed flow- Empty Test Section)
Contraction Section (With Computer) Air Speed- 9 MPH or 13.2 ft/s (Obstructed Airflow)
CFM (Cubic Feet Minute)- 1200 CFM
This data shows that the basic design of the wind tunnel (inlet, contraction and exhaust section) does an excellent job at increasing air velocity.
The gain in airspeed from the shape of the tunnel is around 240%. This is a very low speed example of the potential of a subsonic wind tunnel. I also tested the fan at its highest setting and recorded air speeds of around 26-30mph through the contraction/computer section of the tunnel. This is just with a cheap box fan. Adding a higher speed fan would possibly allow for greater velocities, though the cooling benefits on the computer components would need to be tested at such high speeds.
The goal is removing as much heat as possible from the computer components. There is an ideal point where airflow would be moving just fast enough for heat transfer to be most efficient.
Airflow & Fan Electricity Efficiencies
Another interesting point is the efficiency of one large fan as opposed to numerous smaller fans. Comparing the wattage, costs, and flow reveals the following:
Single Box Fan
1200 CFM = 96 Watts
Purchase Cost- $20
To replicate 1200 CFM, I would need roughly 15 smaller 120mm (Based on Scythe Gentle Typhoon 120)
Scythe Gentle Typhoon (Single Fan)
80 CFM = 12 Watts
Purchase Cost- $20
Scythe Gentle Typhoon (15 Fans)
1200 CFM = 180 Watts
Purchase Cost= $300
Stepping up to an 200mm fan would require roughly 10 fans (Antec Big Boy)
Antec Big Boy (Single Fan)
120 CFM = 36 Watts
Purchase Cost- $15
Antec Big Boy (10 Fans)
1200 CFM = 360 Watts
Purchase Cost= $150
Using a single larger fan cuts costs considerably. The box fan is roughly half the wattage of the 120mm fan arrangement and uses roughly one quarter the power of the 200mm fan arrangement. The initial investment cost of the box fan is also considerably cheaper than buying numerous smaller fans.
While its not possible to use a box fan with most computer builds, in this application where large airflow volumes are required it’s clearly a superior arrangement. The numbers listed above are with the box fan on the lowest setting, the smaller computer fan numbers are all listed at their max setting. Turning the box fan up to the highest setting allows for over 2000 CFM of airflow.
Additionally, I wanted to test flow patterns and turbulence around the computer components. In order to do this I needed a way to visualize the path of the air. Just like in a full scale wind tunnel this can be accomplished by adding smoke to the air. I purchased some smoke sticks (I used smoke testing sticks for HVAC systems, as they are non-toxic and leave no residue) that release smoke at a controlled rate. This allowed me to add smoke into the flow path and watch the air path as it moved around and through the components.
My plan was to test the system in standard configuration with the fans in place on the CPU and GPU coolers. This was accomplished with smoke tracing and produced some interesting results. I thought at lower speeds that the smaller fans on the coolers would restrict airflow. However, testing showed that at lower tunnel speeds (5-10MPH) very little turbulence was created. At higher tunnel speeds (15+ MPH) there were small vortices that would appear around the coolers, and as speeds increased the vortices became unpredictable during testing.
My next step is to test the system with passive cooling of the GPU and CPU. I’m curious if the airflow will pass through the CPU/GPU coolers, or if the airflow will move around it instead. More testing is needed.
The motherboard is mounted to standoffs which sit on top of the surface of the tunnel floor; however, only two of these are actually secured. This allows me to move the motherboard around if I need to, without having to drill a bunch of holes. The GPUs are mounted in the motherboard and sit vertically which is better as there is less weight on the PCIe slots. I made a small angle aluminum bracket which mounts to the wind tunnel floor and holds the graphics cards securely in place.
The GPUs are overclocked using a combination of software utilities which allows for voltage, power, temperature, frequency and fan control. At a stock voltage of 1174 mV I was able to push the core clock to 1125Mhz with ease. The temps stayed in the 48-50 °C range on GPU 1 and in the 40-42 °C range on GPU 2. Pushing further, I increased the voltage to 1220 mV and achieved a core clock of 1225 MHz with temps increasing 4-6 °C degrees. Given the temperature and voltage headroom still left on the cards, I plan to push both further and see what kind of clocks I can get too.
(For this application the GPU memory is actually downclocked as it has no effect on the processing of research tasks)
GPU Test Numbers
First Overclock Settings
Core Clock- 1125 MHz
Core Voltage- 1174 mV
Core Temp- 41 ˚C
VRM Temps- 41-43 ˚C
Core Clock- 1125 MHz
Core Voltage- 1174 mV
Core Temp- 49 ˚C
VRM Temps- 45-47 ˚C
Second Overclock Settings
Core Clock- 1225 MHz
Core Voltage- 1220 mV
Core Temp- 46 ˚C
VRM Temps- 46-47 ˚C
Core Clock- 1225 MHz
Core Voltage- 1220 mV
Core Temp- 56 ˚C
VRM Temps- 52-53 ˚C
I was hoping to get to 1275-1300 on the core clock and keep the voltage under 1300 mV. That’s my goal for the next test session; I think with the temperature headroom I have I might just be able to get there.
Overclocking on the Intel Ivy Bridge 3770k platform has been a learning experience. Unlike the Sandy Bridge 2600k CPUs I worked with previously, there is a much harder voltage wall with Ivy Bridge. Additionally, the thermal interface material (TIM) and attachment method of the Ivy Bridge heatspreader are different than the Sandy Bridge CPUs. Through testing it has been confirmed that large reductions in temps can be achieved by “delidding” an Ivy Bridge 3770k CPU.
This process involves carefully removing the factory applied heatspreader with a blade, and replacing the thermal compound that Intel used with something of superior quality. It appears that the real reason a large reduction in temperatures occur has to do with the reduction in the gap between the die of the CPU and the heatspreader itself. Temperature drops of 20-25 °C are not unheard of with “delidding” a 3770k.
With that knowledge I fully intend on delidding my 3770k CPU so that I can squeeze every last bit of processing power it has to offer. However, I’ve been focused on finalizing the wind tunnel portion of the project and getting everything up and running. So I’ll delid the CPU in the next couple of weeks, when time permits. Currently I’ve overclocked the CPU to 4.5GHz and left it there for now though I plan to push it further.
Frequency- 4.5 GHz
Temperature- 63-65 ˚C (Depending on ambient)
Judging by the current temperatures and voltage used to achieve 4.5 GHz, I hope to be able to achieve 4.8 GHz, possibly even higher. I plan on pushing hard to see what I can get. I don’t think the temperatures will be an issue, it will be a matter of voltage required to achieve those clocks.
Currently, the system has exceeded my initial estimates for output. The goal was to process as much cancer research as quickly as possible.
With the power of the GPUs, multiple tasks/workunits from the Help Conquer Cancer project are able to run concurrently. The average GPU can run 1, sometimes 2 workunits. A higher end GPU can run 2-4 workunits and sometimes 4-8. However, a balance must be struck with the number of units running and the total time it takes to run each unit.
Each GPU can run 8+ workunits; and currently I’m running 10 on each GPU for a total of 20. I’m still testing the number of workunits to run per card for maximum efficiency. At this pace I can process 40 workunits in the time it takes my other computers to process just one.
World Community Grid awards points as a way of keeping track of the work contributed and to help motivate others to join and donate time. The points are worth nothing, they just give you a way to compare what you’re donating in terms of processing power.
Each computer connected to World Community Grid is assigned a host number. There are over 1.6 million hosts registered and of those there are 220,000 active hosts. (Numbers pulled from BOINCstats data) Of those 220,000 active hosts this wind tunnel computer has been in the top 5 for a number of days and has been as high as 2nd position in most points returned for a single host in one day.
On top of this World Community Grid also awards hours of processing time donated. Each CPU thread is equivalent to one day’s worth of time donated. The more CPU cores/threads you have the more days of processing completed at once. For example, a 4 core processor would achieve around 4 days’ worth of processing time in a single day, an 8 core would achieve 8 days, and so on. This computer running 20 total tasks simultaneously achieves around 20 or so days of processing time in just one day.
Since going online, this computer has already completed over 750 days’ worth of cancer research, or a little over 2 years worth of processing cancer research in just over a months’ time.
Workunits Per Day- 4,000-4,200 double units (8,000-8,400 single workunits)
Run Time Per Day- 20 days (worth of processing)
Total Run Time Donated- 750+ Days
Total HCC (Help Conque Cancer) Results Returned- 147,000+ (as of 1/1/13)
First I plan to modify the CPU by delidding it, then I hope to be able to push my CPU overclock. After that, I’ll move on to pushing my GPU overclock. Additionally, I plan to test the cooling performance of the components with passive coolers. Further down the road I plan to do more airflow testing and use the wind tunnel as a test bench for other projects. At some point, I hope to add a second computer system and double the output of the system.
Every now and then an idea will pop into my head that causes me to pause and think “that would be fun!” These ideas vary wildly, from a hovercraft dog-bed styled like Jabba the Hutt’s sail barge for my dachshund… to a monetary system based entirely around Lego bricks as currency. Sometimes these ideas happen, while other times they get added to my possible projects list.
Often times I find myself at an in-between point with some of the larger projects I have going, and it’s at those very times that a smaller project like this is the perfect thing. Doesn’t require a lot of time, and is a fun trip down the rabbit hole.
At some point or another I got the idea to make geeky business cards for myself. I definitely remember thinking “that would be fun!” I didn’t really formalize just what path the idea would take, but it definitely seemed like something fun to pursue. There are billions of business cards out there, and most of them are probably fairly boring. I have business cards for work, they have my name and they’re just dandy as work business cards. But… what if instead of having just work business cards I had day-to-day business cards too! Granted my day-to-day life outside of work is not really a business, so maybe they should have a totally different name, something that reflects that these cards are me, they display my randomness, they display me.
Seeing as how I had my heart set on this plan I set out to start working on creating geek/business-cards for myself. The first hurdle to overcome was Photoshop and the fact that I have zero graphic design background and am a total n00b when it comes to Photoshop. But I wasn’t about to let that stop me. So I started tinkering, playing around and learning as I went. Admittedly, I’m still a noob, but I managed to learn enough as I went to be able to create something fairly close to what I had imagined in my mind.
When I started this I planned on making maybe 3 or 4 designs and then picking the ones I liked best and having those printed. But…. as time went on I couldn’t seem to stop the deluge of fun ideas and things popping into my head. So 3-4 ended up soon being 7-8… and then 9 or 10. Before things were done I had made over 25 different designs. Needless to say things got a little away from me there for a while. These things happen though. After I finished with each design I found fonts that fit each theme and added my email to the back of the cards in the respective theme font, Blade Runner, Star Wars, SEGA, TMNT, etc.
In the end I narrowed it down to 15 designs I liked and had those printed out. I majorly lucked out that @thetinnishflash worked at a print shop and could print me the cards dirt cheap, like 75 cents for 10 cards of one design. So for all 15 designs I ended up with 10 of each design or 150 cards total and it only cost me $11.25.
When this whole geeky business card idea had first come up I never thought much of who I planned on giving them to. But when I started to get closer to actually having them I realized I could give them to anyone, didn’t matter who they were; new people you meet, friends, and family, whoever you want. Since I’m not really trying to sell anything and am just having fun it doesn’t really matter who you give them too.
I’ve already given almost all of them away, and people really seem to love them. The cards are excellent conversation starters. It’s been a fun little project and I’m reaping the rewards every time I give a card away and get to geekout with somebody about Blade Runner, TMNT, video games, Lego and all sorts of other geeky topics.
With every passing project I feel like my basement is being converted from a living only area, to a work and project area. Computers being built, gadgets being taken apart, Lego projects all around. I’m not complaining by any means, but I do feel as my basement becomes populated with more and more tech based projects that the environment is missing something organic, something natural to balance things out.
I’ve been interested in working on a project for while that would combine something tech-based with something organic. Mixing the two elements intrigued me for a while. I’ve wanted to add flowers, or maybe plants of some sort to blend a little bit of nature into the space. I do receive a bit of sunlight through a glass block window, but the temperature in the basement is usually on the colder side. In the winter I don’t really heat the space because the folding farm outputs enough heat to keep the temperature comfortable. I didn’t want to change that aspect so I needed to come up with something that would tolerate slightly cooler temperatures and limited sunlight.
I can’t exactly recall when the idea came to me, but at some point I started wanting to use the heat from a computer as a way to warm the soil and help with germination/growth. I’m about as far from a botanist as it comes, I did some reading online and became pretty interested in the effects of soil temperature on germination/growth. I read different studies and papers from various universities. It was not too long into that process that I became hooked on the idea of using computer heat as a way to control the soil temperature of some sort of living plant life.
As the idea developed further I started looking into wheatgrass as a plant option. There is something clean and natural about the look and idea of a piece of grass growing in my basement. I thought the look would alter the space a little bit and add a bit of color along with something more than just metals and plastics. After reading enough studies and papers on the effects of soil temperature and germination with wheatgrass I felt like I had a good enough handle on the basics to tackle this.
The first step was finding a computer to adapt for the project. Luckily I had a lot of donated computers that people had given me for various project purposes. Most of them didn’t work, but by cobbling together various components from a bunch of different old computers I was able to come up with enough good parts to make a working computer.
At the time I was starting to get a formalized plan together for this project, another idea popped into my head. I’ve got a 5 yr. old who is already fairly geeky (just ask him to do his Jabba the Hutt impression for you) but he really does not get that much computer time. He’s getting to an age where he’s more inquisitive about tech stuff and I think he’s ready for his first computer. Since I was already well underway with this project it seemed like a perfect opportunity to orient the computer itself towards a learning tool for him in addition to using it as a way to blend something organic with a piece of technology.
With all of these goals in mind I started tinkering away. The computer hardware itself was nothing fancy, definitely outdated, but perfectly suited towards this project.
CPU- Pentium 4 (3 GHz)
Dell Factory Motherboard
Various Ram Sticks- 2gb
Old Maxtor 120gb IDE Hard Drive
Old FPS Power Supply
Donated Old Computer Case
If you’ve spent any time around computer hardware in the last decade you’re probably well aware that Pentium 4 has a reputation for running hot. The 90nm Pentium 4 was named Prescott and it didn’t take long before the nickname Pres-hot popped up. In this case though that extra heat is going to be put towards good use.
Once the hardware was all in order I started working on the layout and arrangement for the case. I had originally envisioned the grass growing out of the top of the computer case. This seemed like it would provide a good blend between the hard edges of the computer and the soft feel of the grass. This also worked well with putting the soil area in the upper portion of the case, where the most heat should collect. After doing some initial tests with this configuration I found the measured temperature near the top of the case was the hottest. The placement of the CPU near the top of the case and the lack of airflow in that area contributed to these higher recorded temps.
After finishing with some initial tests I decided to completely strip the case down and start removing all the unnecessary brackets and pieces inside of the case. Like most computer cases all the internal brackets and mounts are riveted together, so I drilled all the rivets out of the components I wanted to remove and pulled them out. The gutted case looked pretty barren after being stripped.
In keeping with my original plan of having the computer blended with the organic grass I wanted it to be easy to see the inner workings of the computer and also the soil from the grass growing above. I needed something that was translucent, which left me with either glass or acrylic/polycarbonate as my options. I ended up choosing acrylic because I was able to find a cheap, used, die-cast model car display case top that was the perfect dimensions for the top of the case. Couldn’t go wrong for 5 bucks.
I test-fit the acrylic over the top of the computer case and marked my layout on the top of the case. My plan was to cut open the top of the computer case and insert the acrylic display case into the section I cut out. To cut open the top of the computer case I drilled a hole as a starting point and then used tin snips to cut through the thin gauge metal.
Once I had the metal opening cutout, I needed something to put along the edges of the freshly cut opening to create a clean and finished look. For $2 I was able to find some black, car-door-edge, plastic molding that worked perfectly.
After getting the opening of the top of the computer handled and the acrylic case properly fitted, I moved onto some more testing with soil in the acrylic and the computer running. I wanted to see what kind of heat transfer I would get and how it would affect the soil temperature. I knew that using acrylic over glass would make it a little more difficult to heat the soil as acrylic is not that good at transferring heat. Luckily though I didn’t really need to alter the soil temperature that much, I just needed a little extra heat. Testing showed that heat transfer with just the single acrylic display case as my soil container was slow. So I decided to add some acrylic tubes. This would allow more surface are for soil to acrylic contact and also give me an area where I could add a mixed substrate to allow for soil drainage.
I went online and found a 3ft long section of 2 1/4″ clear acrylic tubing for $8 dollars. I wanted to add a couple of hanging cylinders off the main of the acrylic display case in the top of the computer case. To do this I needed to cut holes into the bottom of the display case. Cutting holes in acrylic is not always easy, it’s likely to chip and craze at the edges. To get around this I used a buildup of masking tape on both sides of the acrylic, to provide a little bit of resistance and strength. I placed the acrylic display case back into the computer and marked my layout for the tubes. I needed to leave clearance for the CPU cooler, power supply and hard drive mount.
After marking the layout, I used an air-powered pencil grinder and carbide tipped bit and cut through the acrylic case along my marked lines. After the initial shape was cutout I used a tube piece as a template and slowly worked on grinding the holes as perfectly round as I could get them so that the tube had a snug fit all the way around. The fit is pretty important as the acrylic cement that I used needs there to be a tight fit so that the acrylic can bond correctly.
I placed the tubes into the acrylic case and mounted it in the computer and test fit the placement and height of the tubes. After getting everything lined up and oriented correctly I moved towards bonding the acrylic together. Using a needle dropper applicator I applied the cement to all the joints. After letting it set overnight I came back the next night and applied a thin coating of clear silicone caulk around all the joints just to ensure they were completely water tight
The next step was sealing up the bottom of the acrylic tubes. I looked at a couple of different options before realizing that the discs I had cut out of the acrylic case to fit the tubes would be almost the perfect size. Using a bench grinder and a test piece of tube I slowly ground each disc to the perfect size for a tight fit all the way around. I placed the clear acrylic discs into the bottom of each tube and repeated the same cementing and silicone caulking procedure to ensure they were water-tight. After a couple of days’ worth of drying I tested the case by filling it with water overnight and checking in the morning to make sure that no leaks had popped up.
In order to hold the weight of the acrylic case after adding soil, I drilled and installed a machine screw with a rubber cap at each corner of the computer case to bear the weight of the acrylic case.
Once all the work was done with the acrylic case section I started reinstalling all the components. The motherboard, hard drive and power supply were added and then wired up.
At this point all that was left was to make a clear panel for the side of the case. I had some left over acrylic sheets from a different project that were really close to the right size, after test fitting, marking and cutting I then had a clear side panel to see into the case. (I left this off for the pictures because it caused a reflection)
The final step was filling the case with soil and adding some wheatgrass. I had been test growing other samples of wheatgrass during the build process. So I had already a couple of different patches of grass growing. I tried to keep the soil level just below the top of the acrylic case edge to make it a little more seamless.
For the computer itself I ended up installing Windows as well as Linux in a dual boot setup. Since the computer was going to be used by my son as something to play with I figured I should put both on there so he can tinker with them and learn as he goes. Besides, for important life decisions like picking an OS (Operating System) I think any good parent should present the options and then let their kid decide for themselves.
In the end I’m happy with the result of the project. It’s been fun experimenting with growing wheatgrass and I like how the natural look of grass is blended with the very modern, inorganic case.
In testing the case temperature versus the soil temperature I found it was easiest to control the case temperature by using a variable fan speed control on both the inlet and outlet fans. By turning the fan speed down I could increase the temperature inside the case and correspondingly raise the soil temperature.
In order to maximize output of heat during these tests I used Prime95 to run the CPU at 100% load.
Using this method to control the case and soil temperature I played around with adjusting the soil temperature and then tracking the rate of growth of the Wheatgrass. The target temperature for peak growth seemed to fall around 66°F (19°C).
Using the fans to control the case temperature allowed me to vary the temperature inside the case from a low of 75°F (24°C) to a high of 91°F (33°C). In order to achieve the desired increase in soil temperature, I had to run the case temperature toward the higher end of the scale.
When the soil temperature was too high, the growth of the wheatgrass would slow.
The wheatgrass I used for my tests was nothing special, it can be found as seeds (berries) in a lot of health food stores and it grows relatively easily.
Caring for the wheatgrass is just a matter of keeping it watered. It’s a pretty easy plant to work with and doesn’t require any special skills.
Wheatgrass is often used as a nutritional supplement; it can be juiced and mixed into smoothies and other beverages. I plan on harvesting some of the grass and making some juices.
While wheatgrass offers a nice, attractive and easily grown option I do plan on trying other plants as well in the future. Since I already have the “planter” built all I have to do is add a different plant and see how it goes. I’m sure there will be some plants that don’t like the lower light environment, but it will be fun experimenting with other plants as well and seeing how they take.
Donated Computer- Free
Clear Acrylic Display Case- $5
Clear Acrylic Tube- $5
Plexiglass Side Panel- Free
I’m an avid Lego geek; I love sitting down for a build session whenever I can fit one in. This often means pulling out the bins of Lego bricks and clearing a spot on the floor or a table and having at it. With bricks flying and hands madly building it’s only a short time before robots, dinosaurs and all kinds of fun Lego creations are scurrying about. The downside to this fun is at the end of a build session you have to clean up and put away the bin of Lego bricks and even….gasp… disassemble your new creations! The horror.
The solution to this problem seemed simple enough; just build a dedicated area for Lego play and building. Only one small hitch, my house isn’t big enough to devote an area for just Lego play. So I started thinking, I needed an area that I could use as a Lego build area and never take it down, but I didn’t have any more space. This got me thinking further, what about my walls? Why couldn’t I just get some Lego plates and stick them to the wall and build on the wall, seems like that could be pretty cool way to do it. I toyed around with this idea in my head a little bit and eventually decided that if I was going to build a permanent Lego area on the wall I wanted it to be both functional and aesthetically pleasing.
Besides utilizing unused space, I also wanted to implement the design in a way that would allow for a little more creativity during building. Don’t get me wrong, there is nothing wrong with a big flat space of Lego plates but I wanted something a little more than that. Since I was going to be building on a vertical surface now I thought why not take it one step further and add some Lego plates to the ceiling as well. Inverted Lego building area! I thought that by having both a vertical and inverted surface to build on it would change things up a bit and offer the chance to do things I had not done before, plus it was just more space that was sitting there going unused, so why not.
The two most important aspects to me for this project were the overall layout of the Lego plates/design and the location of where I would build this. My basement was the ideal place in my house since it was already my principle domain and a favorite spot for all kinds of Lego, computer and general geekery projects. I evaluated a couple of the walls and corners and ended up choosing one that I felt offered the best design possibilities. While a standard corner would suit this project just fine I had a small area in one corner that jutted out and offered a couple of extra surfaces to work from. Plus, I liked the angularity and look. I spent some time tinkering around in my head what I wanted the layout to look like before I finally came to a decision.
When it comes to the Lego plates you have 3 options for the plates themselves. There are two colors of the 10″ x 10″ plates, either blue or green. There is also a larger “15 x 15″ plate that is grey. The 10″ plates are $5 and the “15 plates are $15 dollars. Since I was trying to get maximum bang for my buck I went with the 10″ plates. I ended up choosing the blue plates since I thought they offered a nice contrast to my wall color, though the green probably would have been just fine too. I had saved up a birthdays and Christmases worth of Lego gift cards and took off to my local Lego store to buy the plates. For the design that I intended I would need about 26 plates.
Unfortunately, the area I intended on putting these plates was not an even surface because when the previous owner refinished the basement they used 1/4″ thick strip moldings everywhere. I needed to get the surface flat before I could mount anything. In order to do this I went to the hardware store and bought a couple of small sheets of 1/4″ MDF board. Then I went through and based on my design laid out exactly the lengths I would need in order to build a patchwork of MDF board that seamlessly blended in with the existing 1/4″ mold work to create a flat surface. I also wanted this project to be removable in the future without damaging the walls. Just in case I ever sell the house and some crazy non-Lego loving person moves in; they can always take it down.
The actual attachment of the MDF board uses short drywall screws to fasten all the pieces in the exact locations needed. The Lego plates themselves are attached using Liquid Nails. The key to making the system removable in the future is to not apply Liquid Nails around the areas where the screws are, this way if you want to remove the system you can break the Lego plates off the MDF support boards and then just unscrew the MDF boards from the wall.
After getting all the MDF boards in place and making some final double checks on the placement I moved forward with gluing the Lego plates onto the MDF support boards. Using a very thin amount of Liquid Nails and spreading it with an old plastic card across the surface, I put each board in place one by one. Once each section was done I used 2×4 Lego bricks to align each plate and make sure they were the exact distance apart, so you could seamlessly build from one plate to another. This is an important step, because if you didn’t align the plates before the Liquid Nails dried you would not have the correct alignment to build across plates.
With all the plates installed, and the Lego bricks at all the edges lining them up, I let everything dry overnight. The next day I removed the bricks and took a look and was very happy with the result. I lasted all of about 37 seconds before I had my bins of Lego in front of me and was building away exploring the different surfaces and things I could do. I spent a couple of hours, 2-3 days in row, playing each night. It was nice to be able to build without having to deconstruct anything. After about 5-6 hours of play over a couple of days this is what my Lego corner looked like…
In order to facilitate easier builds and to get all the bins of Lego spread throughout my house into one place I bought an Ikea Trofast organizer with bins and put it up right under the build area, this allowed for easy and quick access to all my Lego bricks and makes building an ease.
The biggest problem I ran into during this whole project is that there were a couple of spots where the wall geometry is not perfect and square, so I just kind of had to make due and roll with it. In a perfect world all of your walls will be even and square. My house was unfortunately built over 60 years ago, so I’ll cut it some slack for being off an eighth inch here and there.
In the end, I’m very happy with how this project turned out, I didn’t spend too much money (thanks to gift cards) and I now have a permanent Lego build area using space that was unused before. I really enjoy the new things I can do with the vertical and inverted building that I had not played with before. Plus I think it looks kind of cool, but as a Lego-maniac I might be biased.
Lego Plates- Qty 26 ($130)
MDF Board- Qty 3 ($15)
Liquid Nails- Qty 2 ($5)
Total after using Lego gift cards- $20
This past Halloween I decided that I wanted to go a little more all out on my costume than I had in years past. After a little time spent brainstorming I decided that I wanted to try and find a character that let me use my already messy fauhhawk spikey hair and incorporate it into a costume. With spiky hair as a requirement I quickly turned towards anime and video game characters, from there I further narrowed the list to Final Fantasy and settled in on Cloud Strife. I had played FF in the past and thought it would be perfect. Only one problem, my hair is black and Cloud’s is blonde. Then I quickly remembered that Zack also had awesomely spiky hair. That helped finalize my decision in choosing Zack Fair.
The real fun of getting to do this costume was getting to attempt to build a Buster Sword. I did a bit of research and looked around places on the web to see what I could find. I ran across some sites selling replicas for a couple of hundred dollars, but they were not full size. What’s the fun of having a Buster Sword unless it’s the full-size version? I figured out the scale of the sword and decided if I wanted a relatively accurate Buster Sword I was going to have to build it myself from scratch.
I originally set out with the intention of building two swords, the first would be a prototype and the second a more polished version. But due to time constraints I ended up just putting all the finishing touches on the prototype and rolling with it. The sheer size of a full scale Buster Sword meant that I needed to find a way to build it and keep the weight down so that I could actually wield it. I first thought of using foam board and an overlay and then moved towards a hollow wooden structure. In the end though I settled in on just buying a 2×10 8 foot piece of lumber and whittling it down.
After some initial testing for weight I determined that the single piece of lumber approach would yield a sword of the appropriate weight to wield. I busted out the reciprocating saw and cut out the basic shape of the sword blade.
I then moved to the edge of the blade and cut both sides at the appropriate angle and width to create a blade edge that matched that of a full scale Buster Sword.
Once I had the basic shape and woodwork done I moved to the most labor intensive task of the sword build which was the finishing of the wood. The more difficult aspect of using wood is that it’s hard to paint it and not have the grain structure show. After getting the wood itself to a decent smoothness I used some epoxy filler to fill in the knots and rough areas of the wood surface. I followed up by sanding the surfaces before I applied a base coat of varnish to seal the wood. After applying the varnish I resumed sanding and smoothing the surface. Once this was done the next step was applying a coat of primer to the surface and then resuming sanding. This priming-sanding-priming technique was done numerous times until I was final satisfied with the smoothness of the finish.
After achieving a fully prepped surface ready for paint I experimented with a couple of different paint colors to try and achieve the right look for the sword blade. After a couple of coats of paints followed by some wet sanding I was able to achieve a finish that in most cases looked almost metal like. Enough to fool most at least.
The next step following the sword blade was moving onto the handle and hilt of the sword. In testing out the weight of the sword and the balance I determined that I needed to have more weight near the handle to balance the sword out better. For this reason I ended up using a block of aluminum instead of a lighter piece of wood. I purchased a scrap cutoff of aluminum from a metal shop for cheap ($5) and then using a drill press drilled and tapped the holes needed for the handle itself, the attachment points to the blade and the 5 decorative screws in the face of the hilt.
I sanded the block of aluminum with many steps of progressively finer sandpaper to achieve the metal finish that looked best with the sword. The handle I ended up using was a scrap piece of electrical conduit tubing from a hardware store. I cut it to the correct length to match the sword at full scale and then inserted it into the aluminum hilt. The handle inserts into the hilt with a small amount of press to keep it from moving. It’s secured with the middle decorative screws on both sides of the hilt, which are longer and are drilled and tapped into the conduit pipe.
Using another scrap piece of aluminum bought from a metal shop I made two aluminum pieces for the handle, the one fit over the handle and butts up against the hilt while the other is capped and fit at the end of the handle with a press fit.
For the handle surface I ended up getting a cheap spool of leather lacing and wrapping the handle top to bottom with it to achieve the look of a leather-wrapped handle.
The final portion of the assembly involved attaching the hilt to the sword itself. I was concerned at first that the weight of the wood and length of the sword blade would be too much load for screws to hold the hilt to the blade. So instead of using regular screws I found a couple of unused screws I had purchased for a gutter I had hung on my house. These screws are larger in diameter than regular screws and most importantly they are very long, about 6 inches. Just one of these screws was able to hold the hilt to the blade and support the entire sword without issue. I used 3 of them to ensure that I would have plenty of clamping force and load carrying ability. This allows the sword to be swung around without fear of it breaking.
The finished product came out better than I could have hoped, given that it was just supposed to be a prototype I’m very happy with how the overall finished Buster Sword looked.
The sword materials list and cost are as follows-
2x10x8 lumber- $6
Epoxy Filler- $6
Spray Primer- $6
Spray Paint- $7
Electrical Conduit Tube- Free
Aluminum Scrap piece- $5
Aluminum Scrap round- $5
Spool of suede- $5
Total = $48
I worked with my lovely assistant and costume maker extraordinaire @thetinnishflash and built up a pretty respectable looking Zack Fair costume from pieces gathered at the local Value World thrift store.
The costume itself came out great! I’m not a super picky costume person so I didn’t worry about some of the exact details; I just tried to get it pretty darn close. The same thing can be said about my hair, and the Buster Sword. I know Zack has had a couple of different hair looks in the series so I kind of just went with a mix of what seemed to look decent. For the Buster Sword itself there are a bunch of versions that I know of- Crisis Core, Advent Children, FF10, Kingdom Hearts, etc. I decided to just go with a more original and toned down look. I was going to add the gold coloring and handle embellishments but when I got to the end I really liked how the sword looked as it was, so I ended up leaving it. While it’s not a spot on replica I think the look is close enough that most will easily be able to tell what it is.
The sword was a huge hit and everybody really liked it, I had a couple of people offer to buy it and many people took pictures with it. Not sure if I’m going to sell it to someone who more frequently cosplays Zack or Cloud or just keep it and hang it on my wall.
For correct scale reference to a person and for the sheer awesome geekout factor I had @thetinnishflash do some magical Photoshop work and take one of the photos of me in my Halloween costume with the sword and Photoshop me into a Final Fantasy background. Its silly looking, dorky, geeky and everything else and I love it.
Also here is a picture of me making a really goofy face with my hair still done up like Zack fair but out of costume holding the sword. It gives another example of the scale of the sword; it’s a little over 6 feet in length.
Like Tribbles, Legos seem to have an uncanny ability for multiplying in my house at an almost exponential rate. First, you build models, then it’s Star Wars, then it’s your phone, your jewelry. Before things are said and done you’ve got nooks, bins and chests full of them. I’ve been addicted to Legos for longer than I can remember, so when the opportunity comes up to work on a new project of some sort the question that invariably arises is, “Can I use Legos?”
When I first looked into building my next computer I had no intentions other than taking the system and speeding it up. The once venerable overclocked Phenom quad-core system, with its dual Raptor HDs in a RAID 0 and other hardware was starting to show its age. I decided that this time around it was time to start a new platform. I had been making upgrades to my AM2 based AMD system for a couple of years and it seemed like the platform had served admirably but was reaching the end of the road.
Around the time I began my planning I beginning to be involved in Grid Computing. I liked being able to use one of my geek hobbies in a way to help try and benefit others. Grid Computing (http://en.wikipedia.org/wiki/Grid_computing) allows for using your home computer (through the addition of a small free downloadable program) to use its CPU and or GPU for the purpose of processing data in the form of research problems, equations, and more. Normally, it takes a supercomputer days, weeks, or months to works its way through some of this research. Grid Computing leverages the power of hundreds of thousands of computers whose users donate their processing time to make this happen.
Since I was going to be building a new computer anyway it seemed like the perfect time to maximize my build for Grid Computing (Crunching). My first plans were to focus on a multi-GPU setup that would be a dual purpose crunching/work machine. I encountered a problem in that my Grid Computing program of choice did not offer any GPU compatible projects, only CPU compatible. My main goal was Grid Computing with medical research and humanitarian projects in mind. For this reason I chose to go with IBMs World Community Grid as it offered a lot of these types of research. (Cancer, Aids, Muscular Dystrophy, etc.) http://www.worldcommunitygrid.org/
My plans changed when I realized I wouldn’t be able to make use of a GPU folding farm. The type of Grid Computing I wanted to do required CPU power in the form of multiple fast CPU cores. My first plan was to build a dual CPU Xeon based platform with a EVGA SR2 board. Some of the SR2 based systems I was finding were jaw-dropping; the performance was out of this world. But such performance comes with a high price tag. I had planned on using my normal budget of $1,500-2,000 and doubling it when I decided to build a Grid Computing computer. I decided that the money was a small donation in terms of trying to help a much larger cause.
My initial goals with my new system were as follows-
$2000 budget goal
100,000 Points Per Day (Points are used as a rough estimate of computational power)
Energy Efficient as possible
After pouring through and reading for hours in the forums I started to realize that the SR-2 monsters I had seen huge numbers from were also somewhat fickle beasts with RAM and some other settings. At that point I realized I wanted to go a different direction. I started looking more carefully at the computers I already had in my household and also looking at the electricity costs to run all these systems in addition to adding the new Folding computer. I already had a quad core workstation computer and a slightly lower end intel Core2 system running a touchscreen in my kitchen, and a server. Adding a fourth machine gave me pause.
I began thinking a little harder about the whole project. I wanted this to be efficient, at least relatively speaking. I turned my direction briefly towards building a multi-CPU setup based on a server board. This would have satisfied my requirements for multiple CPU cores. The downside of course was the cost. Server CPUs are expensive, as is the motherboard that supports them.
I was stymied for a couple of days on this whole project. I knew what I wanted, but couldn’t formalize a plan to pull it off. As I was working on other people’s computers and staring at the pile of pulled motherboards sitting on my bench I got to thinking, why do I need to have a bunch of separate cases and power supplies? Why couldn’t I build one system that housed multiple motherboards and CPUs? I could use desktop parts that were cheaper and consolidate all of my PCs into one.
Based on this I started to put together a parts list-
3x CPU Coolers
2x GPU (Only 2 of the systems needed video, the third is remote operated)
3x Power Supplies??
3x Computer Cases??
Multiple Hard Drives and SSDs
Now that I had a plan I started doing some research. I wanted to use the highest rated efficiency power supply available, but I did not want to buy multiple power supplies since that seemed inefficient with the power loss I would have per power supply. When I calculated out the power requirements of each sub-system it seemed like I was going to be right around 350 watts. This started me thinking, why can’t I just get one power supply with output around 1100-1200 watts and wire it for all three systems? This would give me the efficiency of buying the best Gold rated power supply along with saving some money. After a little more research I found what I was looking for; the Antec 1200 High Current. According to reputable online power supplies sites and reviewers this was the bad boy I wanted.
Cases presented my next challenge. I had an idea in my head what I wanted the case to look like, but after researching I couldn’t find something to fit my vision.
This takes me all the way back to the beginning. With every project I do I always invariably arrive at the same point, “Can I use Legos?” VOILA! YES! Lego! Lego and computers definitely sounded like a good combination. In reality the structure of a case built from Legos was going to require a fair bit of thought. I needed to get my case laid out correctly and able to support the weight of all the components without Legos buckling or falling apart.
Now that I had a solid plan I got underway with buying parts for the project-
3X Sandy Bridge 2600k CPUs
3x Thermaltake Frio Cpu Coolers
3X Asus P8P67 Micro atx motherboards
1x Antec 1200 HCP Power Supply
2x Corsair SSD (System 1/Workstation)
1x Mushkin SSD (System 2/Touchscreen)
1x WD HD (System 3/Folding Only)
3X DDR3 for each system
8x Aerocool 140mm Case Fans
1x Metric Crapload of Lego Bricks (Technically it was about 2,000pcs)
Through careful timing of Newegg sales, along with promotional codes and rebates I was able to get all the computer parts for right around $1,800. As far as Legos, I already had a lot of the black pieces I would need for this build, and I purchased the others I thought I would need in addition.
I eagerly awaited all of the goodies in the mail from Newegg. On the day the first batch of parts arrived I quickly tore into them and started tinkering.
My first step was bench testing all the parts. After that I started working on modifying the harness of the power supply to fit all 3 systems. After a little bit of work I finally had everything plugged in and ready to be test fired. I hit the button and all 3 systems lit up! I used my test stand to then test each system to ensure everything was running as it should be.
Once the testing was done I quickly moved my focus to building the case. I had a veritable mountain of Legos before me to work with. I slowly started to piece together the design that I had envisioned in my head. I had planned to incorporate some clear Lexan windows into the case. I purchased some Lexan and cut it down to what I thought was the appropriate size. The next step was mounting of the motherboards. I wanted to stick with a basic design philosophy; loading in a downward direction only. The outside walls of the Lego case actually support the load and weight of the components. Trying to hang anything of a significant weight from Legos will pull them apart. This is why the weight must always be pushing them together.
In order to accomplish this I used a couple of thin pieces of aluminum bar, cut them to size, and drilled and tapped them to accept the motherboard screw pattern. These aluminum bars have the motherboards attached to them with regular PC case standoffs. The bars span the case and rest on each of the Lego walls and are encapsulated by Legos. This arrangement uses the weight of the components to apply a compressive force on the Lego walls and ensures that everything is stable. There are 4 of these aluminum bars. The first set at the middle section of the case supports the lowermost motherboard which hangs upside-down, and also the motherboard that sits directly on top of them right side up. The second set of bars sits across the top section of the case and supports the upper-most motherboard that is hanging upside-down. This arrangement of inverting the two motherboards allowed for me to pack a lot of components in a very small space.
Another thing I carefully considered was the overall airflow of the case and the layout of the components. I wanted a short, direct path of airflow from the front case fan directly into the CPU fan/cooler. Behind that, another open section leading to the exhaust case fan. Each CPU/cooler has its own intake and exhaust fan directly in front and behind. The power supply also has this fan arrangement. The space between the motherboards was designed to allow for airflow over both the top and bottom surfaces of the motherboard to ensure maximum air cooling of the PCB and components attached to it.
Another thing I took into account was air pressure. Cases that have a lot of large air spaces, and voids tend to have low pressure over the components they are supposed to be cooling. Air takes the path of least resistance, which means given the option of flowing through a heatsink or around it, air will flow around it. I attempted to avoid creating paths where air could flow through dead space without cooling anything. This is part of the reason that the components are spaced so close together. I also made sure to buy case fans that had a higher pressure rating to make sure I had adequate pressure to correspond with the airflow.
I attempted to do my best to cover and hide wires. This was both from a standpoint of appearance and also for avoiding possible interference to airflow. Many sections have wire hidden or concealed under Lego panels to provide a cleaner look.
After getting the majority of the case structure done I moved onto wiring. To say I had ten pounds of wires in a two pound basket is an understatement. It was tedious work ensuring all the wires were out of the way of the airflow paths and components, especially with having the wiring of all 3 systems crammed in such a small area.
After getting everything setup, I worked on installing Windows 7 on the first system. The install went very quickly and before I knew it I was in Windows configuring the SSD RAID 0 setup. I then moved onto the other 2 systems and did the Windows installs and configurations on them also.
Once I had all 3 systems up and running I went to work on overclocking. The new UEFI BIOS was a bit unfamiliar at first, but after some tinkering I got the hang of it. I played around a little with the settings and soon enough I was staring back at a 4.7Ghz number for each of the CPUs. I setup each system running an instance of Prime95 and let the machine go overnight to test my stability. When I returned the next morning I was happily greeted by all 3 machines still running without errors and with temps right at the 60-65 degree mark.
Seeing that the overclocked systems had all performed without error, I pulled up the World Community Grid/BOINC program on each system and started crunching. After a couple of days it looked like my average points per day was about 43,000 to 47,000 points per system. With all 3 systems crunching as a team this gives me a per day average of around 135,000 points. Given that my old system used to average about 10,000 or so points a day I would say I’m very happy with these numbers. I’ve managed to increase my folding/crunching performance by a factor of about 13 while only increasing my power requirements by about double.
Since my UPS has an LCD readout that displays wattage consumption I used it to compare the differences in power between my old system and the new folding farm. Not exactly super-duper accurate, but close enough for comparison sake.
AMD Phenom Quad Core System- (4 CPU Cores, 4 Threads)
Full Load- 350 Watts
Folding Farm Sub-System- (4 CPU Cores, 8 Threads)
Full Load- 270 Watts (Including all case and CPU cooler fans)
Entire Folding Farm- (12 CPU Cores, 24 Threads)
Full Load- 600 Watts
Instead of having 3 separate computers taking up my desk space I now have one system that functions as three. I sold off the two other computers I had to recoup some money from this build as well. In the end the most important thing to me though is that I feel like I’m doing more to help contribute to a good cause in humanitarian and medical research. I know it’s just one system, but every little bit counts in finding cures and solutions.
Folding Farm vs. Old Workstation PC
Crunching Points Per Day Average- 135,000
Power Consumption Full Load- 600 Watts (UPS Measurement)
Old WorkStation PC-
Crunching Points Per Day Average- 10,000
Power Consumption Full Load- 350 Watts (UPS Measurement)
I choose to use Windows 7 as the operating system for all 3 systems primarily because I already had copies I had bought and installed on the other computers. There was no added cost for me to keep using it. Additionally, I have a Windows Home Server that plays very nicely with all the other Windows 7 machines and wanted to keep it that way. The remote desktop function native in Windows 7 also makes it brain-dead easy to remote in from any other computer to keep up with the folding progress.
If you were starting off from scratch and the operating system cost was a factor you could very easily repeat this setup using Linux instead. This would save the cost of the operating system and give you a lot of the same functionality.
The entire build process had me looking for solutions to problems that arose during construction. As I look further in detail at certain areas I think there are changes I would make with future builds.
Add lower, middle and upper layer between sections about 1″ thick that would have openings for all the wiring to go into and be concealed. This would allow for an almost complete elimination of wiring to work around and organize.
Get more rounded Legos and other shaped pieces that would allow to create more aerodynamic surfaces for airflow in certain parts of the case.
One area of concern I had initially was the orientation of the CPU coolers. The reason being that the coolers I chose to use are a Heatpipe style cooler. This type of cooler uses tubes filled with a liquid that go through a phase change from a liquid to a vapor to release heat. The issue I thought I might have is that if you invert the cooler (by installing the motherboard upside down) that you would not allow the cooler to function properly because the liquid was moving in the heatpipes as a function of gravity.
I made some calls to Thermaltake which put my fears to rest. The cooler uses a capillary action inside the heatpipe in this model that allows the liquid to move back to the base of the cooler no matter what direction it’s mounted. Keep in mind that there are motherboards which use heatpipe coolers on the PCB directly and these may not have this same internal capillary function.
Z68 versus P67 Chipsets
When I started this build I had wanted nothing more than onboard graphics, which are native and built into the Sandy Bridge CPU architecture. However because of the way that Intel was offering chipsets at the time you only had two choices, you either got a H67 with onboard graphics capabilities and no overclocking ability, or you got a P67 with no onboard graphics but with overclocking ability. Because of this I was forced to go with a P67.
Fast forward to the present and Intel now offers the Z68 chipset which offers both onboard graphics and overclocking. This is definitely something I would have preferred since there would’ve been fewer components to worry about and, more importantly, less wattage required.
Testing out the multi-system wiring and the power supply I found that if you just used one power switch it would turn on all the systems and shut them all off at once. However, it would not turn on every portion of each system, just the main power and fans. Missing was the triggering to enable graphics and a few other things. I had to wire the switch to activate all three boards at once in order to get correct operation. You could accomplish the same thing by having three separate switches and turning them all on, but I went ahead and just wired it with one switch. Granted, this system runs 24/7 so it will be rare that I will ever be turning it off and on.
I’ve been experimenting with adding Lego pieces in various parts of the case to alter airflow paths and try and focus the air more on the things I want to cool and less on the dead space of the case. After trying a couple of different variations in air dams and directional vanes, I’ve noticed the temps move around quite a bit. So far I’ve managed an additional 2-3 degree drop by adjusting and optimizing airflow. I plan to continue using this system as a test-bed for further airflow and case development.
When I built the case I tried to keep the design fairly symmetrical in the upper level that houses the two motherboards. My goal was to be able to add another level at some point down the road and add an additional two folding only sub-systems. For this reason I intentionally chose a power supply that was larger than I needed. The power supply needs to operate somewhere between 50-90% of its peak in order to run at maximum effciency. I should have enough power in reserve to add more sub-systems down the road. I also drilled and tapped the upper most set of aluminum bars with the micro-atx motherboard layout on the other unused upper side. Hopefully when the time comes this will simplify adding another level on top.