Overclock.net banner
1 - 20 of 38 Posts

148 Posts
Discussion Starter · #1 ·
The Pride of Hiigara
* * * * *


  • Introduction
  • Base Layer
  • Design Layer
  • Electronics Layer
  • Peripherals
* * * * *

Greetings folks! I'm not much of a regular here, but it has been suggested to me multiple times that I throw up my build log on OCN for more exposure to everyone. So, a bit of background info about myself and the project:

Around mid last year, I got a new job that I consider one of the best jobs to work as. Easy work, good pay, fly around the world, company paid accommodation and utilities...nothing to complain about. So with that, I decided that I should use this good fortune of mine for something positive, and long story short, I came up with this idea. Free electricity, plus powerful air conditioner, equals the perfect setting for an insanely powerful [email protected]/BOINC system.

However, I did not want to go for some boring old case mod. Needed to do something much more original. Besides, the hardware I would use simply wouldn't fit in even the largest case. So, after some inspiration from an absolutely beautiful wall mounted computer, I decided this is the direction the project will head in.

The idea for the design took some thought, but in the end, when all things fail, my love for the game Homeworld always pulls me through, and just like that, I decided to use the Hiigaran symbol as my design, though upside down, as I prefer the look that way.

Confused yet? Righty, let me clarify. This wall mounted system is divided into three parts that I shall call layers. The first layer is the base. The base serves as a standoff between the wall and everything else. Its primary purpose is to protect the wall and hide cables. Otherwise I would be drilling a massive number of holes into the wall, and many cables would be visible. This layer is about as big as an entire side of my room, so it practically covers one whole wall. Rather than being mounted to the wall behind it, it is mounted to the sides, resting atop four drawer slides on each side, which are secured to the wall with some heavy duty toggle bolts. It spans the entire height of the room and the bottom rests on the ground, held up by caster wheels. The material is 18mm plywood, so my estimates for the weight would be in the 100 kilo range.

The second layer is the design layer. This would be the symbol I mentioned earlier, and made out of 10mm clear acrylic. 32 pieces in total, with 45 degree angle edges, painted front, and LED backlights to create a glowing effect around the edges. They will be mounted to the base using long bolts with standoffs to create a gap between the two layers.

The final layer is the electronics layer, and this will be the computer hardware itself, plus any other electronics such as the lighting. It will consist of a motherboard capable of supporting many video cards, four power supplies, and water cooling. The hardware will be arranged aesthetically on various parts of the design layer, and as such, I will be using risers to connect the video cards up to the mobo. However, due to the positioning, I will be using USB 3.0 cables mapped to PCI-e x1. While this would be detrimental to gaming performance, it should not affect performance in Folding/BOINC. Correct me if I am wrong on that, as the build relies on this to be true.

I haven't sorted out the details for the water cooling system yet, but I do plan to use squirrel cage fans, and a lot of tubing, preferably copper for additional heat dissipation. I predict that it will be a nightmare to remove those initial air bubbles once water starts pumping through. I'll likely install valves at multiple points to remove them.

Other points of interest:

- This project has already spanned over several months. With me out of the country for two thirds of the month, and trying to find time to work on things when my two roommates are out, things tend to progress slowly. Only recently did I finish the base layer.

- Four platinum efficiency 1600 watt PSUs are planned for usage. To prevent the tripping of circuit breakers, power for each unit comes from a socket hooked up to a different circuit. I am still looking out for power supplies that deliver a higher output while maintaining high efficiency though. I might consider getting eight power supplies and running them at 75% capacity instead of four at 100%, but this will depend on what is available once I am ready to purchase everything. Still a few months away with that.

- A master control panel will be built on the side of the base layer. It will contain voltmeters and ammeters for each circuit, along with guarded switches to cut power to the circuits, light control for the design layer LEDs, a starter button that activates each power supply through self-built circuitry, and a water temperature gauge.

- Due to space limitations in my room, I plan to modify my bed by having five monitors in an enclosure that is at the foot of my bed. During use, I pull the enclosure up, then push it towards me. This enclosure will serve as a desk with all necessary peripherals, and will be lifted using linear actuators.
* * * * *

Base Layer:


The base layer began with these. Eight drawer slides, and a bunch of toggle bolts. The width of the project will span the area between a small protrusion on the left of my room that is designed to hide my curtain slides, and the wall on the right, above my doorway. This is where the slides are located. Clarification below:


Four drawer slides to each side, held securely with the toggle bolts. You're probably wondering why I need slides for this. Given the sheer size of this project, and the need to hide cables and such at the back of the base layer, I need some way to be able to access all of that. I went through several design ideas, including a swing-out design where the entire assembly can swing open like a cupboard door on heavy duty hinges. I decided against that for several reasons.


Here I have removed the inner section of the drawer slides, so that the side of the base layer can be attached to it. The side is 25cm wide, and as such, the gap between the wall and front part of the base layer will be 25cm.

I will refer to the side of the base layer as the perimeter, and the front of the base layer as the front face.


The right side of the perimeter is complete and mounted. It consists of three planks of plywood joined together with three flat braces. I could have used one long piece, but I had a few limitations here. The first is that my supplier only has lengths of up to 8 ft, and the height is a little over that. The second is that 8ft is the largest size I could fit in my tower's service elevator (I live on the 28th floor). The third is that I always try and use my supplies efficiently, so whatever resources I have, I use.

As a side note, I tested the holding capacity of the drawer slides and toggle bolts by fully extending the slides and right side perimeter assembly, grabbing a hold of each side of the wood, and hanging off it. I'm a mere 50 kilos, so this tells me that the left and right sides should be able to easily hold 50 kilos each, just on their own.

I've also made sure to keep the ceiling and floor clearance tight. There is no more than a few mm gap at the bottom, and the gap up top was surprisingly too small to be measured with my ruler, meaning it is less than half a mm. No scrape marks on the ceiling either, so I'm really impressed with that.


With the left and right perimeters complete, all that's left to complete this part is the bottom perimeter, which will join it all up, and align the left and right sides straight. The construction was exactly the same as the left and right, except with the addition of five caster wheels, secured with large bolts.


Underside. The wheels come with brakes to keep the assembly secured when extended. The brakes aren't used when the assembly is pushed closed against the wall, as it is secured through other means. Plus, the wheels get hidden once the front face goes on, so the brakes are only accessible from the back.

I will refer the act of pulling the entire assembly away from the wall as 'opening' the assembly, and pushing it back against the wall as 'closing' it.



Aligning everything, and making sure it is level. Also, C-clamps are amazingly versatile tools. Here, they serve as handles, which make maneuvering the workpiece an absolute breeze.


The first board for the front face. These are secured to the perimeter via 15cm long wood screws at 18cm intervals. Wasn't exactly easy doing this. One of those boards probably weighs almost as much as I do, and trying to keep it upright while measuring, marking, and drilling is challenging, to say the least. On top of that, I needed to keep the boards up off the ground as well, since I needed that small clearance to avoid the wood snagging against the ground when pulling it in or out. a combination of flat braces for clearance, and strategically placed furniture served as adequate tools to accomplish this.


5 out of 6 sections of the front face mounted here. The C-clamps were just there to keep the adjacent sections aligned until I could fit some flat braces along the edges.


A behind-the-scenes look at the bracing. You'll also notice some bars of T-section aluminium at the top. Some of the wood was warped, so they were not perfectly straight. Plywood is quite vulnerable to this. The majority of these issues can be fixed by applying moisture to the concave-in side. Spray of water, wet cloth wipe, whatever. The goal is to reintroduce moisture to this side because it lost more moisture than the other side, causing it to shrink more than the other side. Regardless, these support pieces fixed the remaining warp.


Due to the location of the wheels, plus the assembly's center of gravity, there is a slight tendency to lean forward. To stop this while the assembly is closed, I came up with a very simple solution: A latch on each side.

These latches aren't meant to be used perpendicular to each other, though, so I took the metal piece that the large bit latches in to, and hammered the curvy section until it formed a C shape. This prevents the latch from slipping out.


Now if you are electrically and mechanically inclined, before this photo you might have been wondering how would I be safely attaching power and network cables to something that can move. The above image is your answer. A cable track. If I did not have this, and decided instead to run the cables from the wall, directly to the assembly, any attempt to open the assembly would result in the cables being ripped from their mountings, resulting in cable, wall, hardware and/or assembly damage.


Two layers of primer, three layers of matte finish paint, and seven months since the project started...I have a big freaking white box on wheels, stuck to my wall.

But it also signified the completion of the base layer.


An inherent flaw in the design of the base layer, which I never even considered, was a significant tendency for the unit to lean forward. This wasn't immediately apparent upon completion of the base layer, but as the design layer progressed, I started noticing this more and more. Here's the issue:

Because the front face is where the majority of the weight is located, the center of gravity is very far forward on the longitudinal axis. Well past the mid-point, where the wheels are positioned. As a result, the heavier the unit gets, say, from adding the acrylic pieces, or the hardware, the worse this lean becomes.

I came up with a simple solution. If it leans forward, I counteract it with an opposing force. While I can't create this backwards force directly, I can indirectly do so through the use of anchors, pulleys, a steel cable, and a weight.

If you look at the image above, you can see one of the two anchors. It's the grey rectangle located near the top of the perimeter. This is just a couple of pieces of metal screwed into the perimeter, holding the cable in place by clamping down on it. One end of the cable starts here, looping around the anchor to achieve a better grip, then threads through a nearby pulley. The cable then travels across to the weight that it supports, then threads through the second pulley on the other side, and ends at the other anchor.

For those curious, the weight that I am using is the mirror that was hanging from this wall. Also, you can ignore the piece of wood underneath the visible pulley. I placed it there to prevent the pulley from scratching the all as it rotates, but later found out that it remains clear of the wall when a weight is on the cable.


This is a closeup of the pulleys. It consists of a large diameter M12 washer with a bolt securing the actual pulley. A 6mm hole was then drilled above the center to thread a toggle bolt through.
* * * * *

Design Layer:


Before even starting to work on the base layer, the very first step was to try and visualise how the finished product would look. To help with this, I took a vector of the design and resized it to the dimensions of the area I wanted to fill on my wall, keeping the ratios the same. Once I resized the vector based on those measurements, I headed over to a print shop, got it printed, and hung it up on my wall. Honestly, I loved looking at it, even in paper form.

Unfortunately it needed to be taken down so that the base layer could be built, so I pulled it down. Its future use would be more important, so each individual shape was cut out.



What I did here was arrange the shapes to find the arrangement that requires the least amount of acrylic boards to order, so that wastage is minimised. The knives at opposite corners show the dimensions for 4x8ft.



Before I can cut out the acrylic into the right shapes, I first need to cut out a template. As I mentioned earlier on, the acrylic will have 45 degree angled edges, which will be made using a router. However, I lack superhuman steady hands, so slapping the paper designs directly to the acrylic will not help.

To do this, I first stuck the paper onto some 6mm plywood. This was then roughly cut out with a jig-saw, and sanded down so that the shape of the wood perfectly matched the shape of the paper stuck onto it. From there, I can discard the paper, clamp the wood to the underside of the acrylic, adjust my plunge router to the correct depth so that the pilot bearing on the chamfer bit is aligned with the wood, then simply push against the edge of the wood and work the router along this shape. This will get a near perfect replica of the wooden shape on the acrylic. The major advantage to this method is that wood is very easy to shape, compared to acrylic, and allows a higher tolerance to error. Plus, if I make a mistake, wood is very cheap. Acrylic is not.


Finished the first shape. This is the small circle that will be in the center near the bottom. The edges are faded because I had to peel the masking tape off, so that I could see the final millimetre of wood that needed to be sanded off. As for the thumbtacks, well that's because once the wood was sanded down to the edge of the paper, the tape was no longer securing the paper to the wood, so I needed another way to keep it all steady.


Same piece with the paper removed. I drew some perpendicular lines and measured their lengths to determine how accurate my work was. I was off by a millimetre in one area, which honestly is impressive for me, considering I've never done something like this before. Bench-mounting a Dremel is amazing!

Anyway, this small piece took me two hours to do, though mostly because the sanding drum on my Dremel was well past its usefulness. The drum I had used was rather smooth after extensive use on previous parts of this project, and past projects. Replacement drums naturally sped up the process with the rest of the shapes.


Someone wanted a photo of this particular template, so I obliged. This is the template for the majority of the wing section. To save on how much wood I'd need to order, I split this up into three pieces. The first is what you see above, and the other two sections are the left and right bits that I will join up once all are finished. This piece was also the third largest.


Several exploding sanding drums later, the large circle is complete. Praise the sun!


And the final template is complete! I had been putting this one off because I knew it would be the hardest to make, and man was I right! Due to the length of this thing, maneuvering it was a nightmare. I was unable to feed the piece through the sanding drum at a smooth and constant rate, and it showed, with uneven edges. The entire process took about two hours to get to the final product, which also involved manually sanding the edge to get the curvature right. There are still some imperfections, but I consider them insignificant. One would have to look very carefully to see it up close.

Considering how many months the template creation process took, it's only fitting that I seem overly happy in the photo!


First acrylic piece complete! Despite the edges looking uneven, the acrylic is actually quite smooth. That's just the protective coverings making it look that way. These will remain on until it is time to mount and paint the acrylic.

It was nice to see the shape created quickly, compared to the templates. I'm also quite proud of myself, since this is the first time I've ever used a router, or made something like this. Of course, I wasn't too happy with the cleanup (1, 2), which took about one and a half hours of vacuuming.


All but one of the feather pieces complete, plus the small circle.

One of the pieces are missing due to a mistake I made while routing the piece. The template underneath it slipped out of place, and I couldn't fix it due to lack of material on that particular piece. There's plenty of scrap acrylic that can be used to make another one, but this will be done later, in case any more mistakes are made with future pieces. That way, I can redo them all at the same time, making the process a little faster overall.


As a result of minimizing the amount of wood I needed to order for my templates, the wing template was made in three pieces. These were then joined back together using super glue to bond the edges. L-braces were added for a little extra support while the template is moved around. These braces will need to be removed when the template is needed for routing.


After tracing the template out on the acrylic board twice, I made rough cuts with my jig-saw and got two duplicate wing pieces. I simply flipped one of the two around, so that I'd get a mirrored piece, for a left and right wing. Just to satisfy my ever increasing excitement, I decided to arrange some of the pieces as shown above.



All acrylic pieces routed. The last of it was the large circular section, the three pieces making up the small circular section, the two wings, and that feather that I messed up earlier on.


In preparation for mounting the acrylic, I needed to determine suitable points on each piece for the mounting screws, which will be M6 rods with tapered phillips heads. Countersinking will take place later on, once all pieces have been drilled.


A visit to the print shop later, and I had another printout of the design. After some painfully meticulous aligning of the five rolls of paper, which involved the use of three chairs stacked precariously in a pyramid arrangement that I nearly fell off twice, I had the design attached to the base layer.

From here, I simply grabbed each of the feather pieces, aligned the acrylic to its respective shape on the paper, and drilled. I started with the small pieces first, since they were the only ones I could do on my own. The larger pieces will be two man operations.

At this point, no countersinking has been done for the screws. The goal is to first get everything mounted, and then countersink.


After getting the positioning correct, the acrylic was taken down, along with the paper. Masking tape was then attached to the edges of the acrylic, so that the top edge of the tape would align with the front edge of the acrylic. The excess tape was secured underneath the acrylic.


Placing cling wrap on the floor with sticky tape, I made myself an area to paint on, and proceeded to apply two coats of white primer, and three coats of night blue paint. This was the required number of coats needed to prevent any light from the LED strips from shining through the front.

Once the paint had time to dry and cure, I took my sanding machine and sanded the front edge to easily remove any paint that had spilled over onto the masking tape. By doing this, I could then take the masking tape and peel it off easily, without worrying about damaging the painted front. After removing the tape, I manually sanded away the last remaining bits of paint that had made it past the masking tape in a couple of areas. My method of sanding was to use those little nail files you get in the hotel vanity kits which I swipe for absolutely no reason every time I'm on a layover. Told myself they might come in handy one day, and I guess I was right.

During this final sanding, I realised that I sanded away a little bit of the blue paint right at the edge, revealing the primer underneath. Turns out this resulted in giving each acrylic piece a very thin white outline, which I liked the look of, so I continued with it.


Finally, after about a year and two months since the project began, I have completed the design layer!
* * * * *

Electronics Layer:


Originally I had created an external box that could hinge downwards to access the wiring from behind, but I decided to completely scrap that idea and integrate everything directly into the perimeter on the right side. This was done for a few reasons. The box stuck out, so now things look much better aesthetically. Second, the fact that the box would need to hinge down means that the cables would move about, possibly causing a loose connection somewhere, since there is no room for any cable tracks. At least, any small enough for me to find in stores here.

The layout is simple enough. Four voltmeters on the top row, and four ammeters on the bottom row. After cutting the holes and sanding them, I inserted the gauges by pushing them in, then securing them in place with screws. Four holes were then drilled underneath for the master circuit power switches.

A couple of mistakes you might notice here. The first is that black bit between the third voltmeter and ammeter. This was a result of me sanding away too much in a moment of inattentiveness. Nothing that can't be fixed at a later stage though. The second thing you'll notice are the bits of white here and there on the gauges. I had a very stupid idea of trying to paint over everything to fill up the gaps between the gauges and make the black housing less apparent. That was a massive waste of an entire day. Removing all that paint was a nightmare. I don't even know what I was thinking when I tried that.


Halfway done. This was also something else that I decided to redo. My original intention was to run cables into my room which connect to external wall-mounted outlets, and then have another set of cables plug in to these outlets, and go to the control panel. The original intention of these outlets was to have a more convenient place to disconnect power from, rather than going all over the apartment to each room delivering power. Instead, I have decided to run the cables directly from the rooms, to the panel. It's simpler, more reliable, and there are fewer chances of running into issues of poor contact where one connection meets another.

Above, only two of the circuits have been completed. The one that is powered from my room, and the longest one which comes from the living room. I've done away with the crude wood clamps that held the cables in place, and instead, I have hot glued them. I managed to get the cables mostly straight, except for an area I had to work around due to the acrylic mounting screws being in the way.

At the panel, the power cables attach to the terminal blocks and then split to different areas. The live cable splits and goes to one leg of an ammeter and voltmeter. A single neutral cable returns from the voltmeter. At this point, since neither the lighting nor the hardware is done, the other end of the ammeter has nowhere to go. Later, the live wire continues on the other leg of the ammeter to the switch, which then goes to the PSU.

A third dedicated terminal block is located underneath the first two, which will be used exclusively for grounding. Naturally, the switches will be grounded, as will several other parts of the system. This area will also be lined with metal for additional grounding safety, as well as acting as a base to create an enclosed metal box. The enclosure will provide fire safety, and act as a barrier to ensure I don't accidentally touch live components while working behind there, in the event I forget to shut off power. I have 2mm of 4 x 8 ft steel sheet metal which I will use.

As a side note, I severely underestimated the weight of that sheet. And that noise when I dragged it across the floor! I'm glad my floor doesn't scratch easily!


Time to set up the LED lights! I decided to start with the 26 feather pieces first. I cut my LED strips to length in either one, two, or three sets of three LEDs, depending on the length of the acrylic pieces.

Since each strip has four contacts for positive, and negative red, green and blue, I needed four sets of cables per strip. 26 strips came out to 104 hand lengths of cables. These were then soldered on after removing the waterproofing with my flat precision knife. However, because the cables were too wide (couldn't find thinner), I had to solder the wires onto the contacts by alternating between soldering onto the front for the first, then underneath the strip for the second, then back up top for the third, and bottom again for the fourth. I then applied hot glue to both insulate, and to help immobilize the area joined.

I have only soldered things once or twice before today, and the results were horrific, so I was pretty much starting with no experience on something so tiny and fiddly that I found myself losing my patience many times. It took me an hour to get four cables onto one strip when I did my first five or ten strips. By the end, I found myself able to complete four sets of strips (16 wires total) in 15 minutes.

Once I had finished soldering all the strips, I secured them to the backs of the acrylic pieces with glue. Originally I used hot glue, but it didn't bond well to the acrylic. So I purchased something a little more potent: Loctite 495. Tested it on scrap acrylic first with a faulty LED strip. Didn't believe it when I was told I only needed literally one drop and it bonds in a couple of seconds. I couldn't get the strip off the acrylic!

Once bonded, an 8mm hole was drilled into the base to thread the cables through to the other end.


Wings were next. Made the appropriate measurements and cut the strips as required. Here I secured them first to the wing, and soldered the cables on using only the contacts on the back of the strips. In this case, I needed to alternate the contacts by having two wires on one of the four sets of contacts, then soldering the other two wires to the next set of contacts.

The strips are all wired up to each other. There is no beginning and ending strip. Because the LED components are all in parallel, I can essentially create a ring with this circuit, which I decided to do for lower resistance, and just in case the clockwise path is damaged for whatever reason, the power can travel anticlockwise instead. Basically, there are two possible paths to each light.

You may notice a single cable coming from the strips to one of four mounting screws. This is my method of getting the power to go through to the other side. As I expect to be pulling down the wings when the video cards are installed on them, having the power cables threaded through as I have with the feather pieces would end up being a more time consuming task. Instead, I have essentially turned the wings into giant plugs. No need to disconnect additional wires from the back. I simply undo the nuts that secure the screws in place, and the cables come off with it. Plus, it's one less hole I need to drill.

Also, you can probably guess which set of four cables I tried soldering on first. Compared with the other neater ones, these four stick out like a sore thumb!


I can breathe a little easier now, having completed the LEDs. Strips are all attached, and soldered. The large circle was the last one to complete.


Wiring has been put in place behind the base for each acrylic piece. However, only the 26 feather pieces are lit up here as the wiring at this stage is incomplete. To give you an understanding of how the wiring works at the back:

As there are so many LED strips, there are a lot of wires. To help cut down on the amount of wire used, I have devised a solution that serves to combine similar wires. That is, one to combine all the + cables, one to combine all the - blue, one for - green, and one for - red. This way, instead of hundreds of cables all going towards the power supply, only four cables will go there. For lack of a better term, I'm calling these wire combining areas 'hubs'.

There are three secondary hubs; one for the left set of feathers, one for the right, and one for everything else. These three hubs are connected to the primary hub, which in turn connects to the power supply. So power comes from the supply, to the primary hub, splits to the secondaries, and splits again to each LED strip. These hubs are nothing more than four grounding bars. I figured they would be more practical than terminal blocks, since everything connected to the bars would be in electrical contact with each other.

With these hubs arranged as they are, I have the advantage of adding resistors to control each hub as a whole. That way, I am able to fine-tune the brightness, and fiddle with the light distribution later on.

With the LEDs all wired up now, this is the perfect time to show the complexity of the wiring, and why the hubs are so crucial:


This is the primary hub. At this point, there is no power supply. That will be added later. Seems quite simple enough at this stage, though. Only three sets of cables. Each of these three cables lead to the three secondary hubs.


This is the center secondary hub. One of those sets of cables comes from the primary. Everything else heads to the LEDs. Three of them power different sections on the bottom circular segment, two power the wings, two power the small and large circles, and one powers the top circular segment. So it's getting a little more complex now. I've saved the best for last though...


Amateur cable porn! I should work on making those things straighter, and cut back on the hot glue! This is what the left and right secondary hubs look like.

Hopefully this shows how essential the hubs are. Without them, there would be a hideous mess. That, and I'd probably use at least three times the amount of cables.


I give you the final product of the LEDs from behind. Well, if you ignore the missing power supply and LED controller, which may or may not have blown up from an excessive current draw...
* * * * *


Building this beast is all well and good, but I'll obviously need some way to interact with it, and headless is certainly not the way to go! While I have yet to decide what specific parts to use for my peripherals, the general idea is to have five monitors, a keyboard and mouse, and surround sound.

The issue here is where to put all that. My room is quite small. I'd say about 15 x 15 ft. As you can see in my previous photos, I have a large wardrobe to the right side of the room (which is primarily filled with tools, rather than clothes), a small desk to the left, and my bed in the center. The desk is too small, and I'm not content with anything less than three monitors, so instead, I've come up with a different idea: Build everything in to my bed.

I got this idea from those fancy beds that have TVs built into a little storage area at the foot of the bed. When not in use, all the peripherals are stored there, so that it keeps the room more open and tidy. During use, it would rise up with the help of linear actuators, to a height that would be comfortable with me in the bed, and my back resting against the head of the bed.

Of course, this doesn't help with everything at the foot of the bed. I lack the long arms required to reach that far! So further modifications to my bed would be made, by introducing rails to each side, connected to each other underneath the mattress, and mounting the peripherals, movable housing, and actuators onto these rails. The rails would then allow me to slide everything back and forth to a position most comfortable to me. At this stage, I am considering manually sliding it back and forth, rather than electrically, but I may review this later on. My design ideas are still only drafts.


This is the foot of my bed. Yes, my pillows are there, I know. My bed is rotated to make room for my workbench. Anyway, at the foot of my bed, there are these four large hand screws, likely for mounting things at the foot. Using these, I took the measurements of both the screw holes, as well as the cross-section of the entire bed, then cut it out of my last board of plywood. I mounted it as-is, and left it there, so that I can come back to it later. I did this primarily so that I can get some of my construction material out of the way.

The purpose of this piece of wood is simply to serve as a hard surface between the bed, and the movable housing. I wouldn't want the bedding to get caught on the bottom of the housing as it closes down. Later this will look a little more elegant. I plan to stain the wood and route the edges.


This is the backboard for the movable housing. The monitors will be mounted to this. Ideally I want to have five monitors, but as you can see from just this one monitor, it looks like the most I could fit is three. I still believe I can fit an extra two, provided that I can use some monitor mounts with three degrees of yaw freedom that would allow me to create a semi-circle arrangement when in use, and fold the monitors into a flatter, more compact state when not in use.

148 Posts
Discussion Starter · #3 ·
Hello again! Finally home, and made the relevant edits. I've included all the important bits in this edit up to what stage I'm currently at with the project, but for simplicity, a couple of photos aren't in chronological order, and I've only included details of the important and most relevant things.

148 Posts
Discussion Starter · #5 ·
The concern I have is primarily with ensuring the addition of the hardware does not ruin the look of the design. I'm probably going to stick paper templates of the hardware and water cooling components onto the design when complete, so I can decide certain things, such as whether I want the water tubing visible, or hidden behind the whole thing. It would be nice to have it visible, but for the design to remain impressive and aesthetically pleasing, I think minimalism may be required. Either way, will experiment when the time comes

148 Posts
Discussion Starter · #6 ·
Small update. Found some time to continue with my work, so I finished cutting out all the shapes on the first board of wood, and decided to try sanding down the excess wood to replicate the paper shape perfectly. Turned out quite well, even if I did spend two hours of almost non-stop work on it. I've edited the first post to include the photos.

I have a feeling the templates are going to take a looooooong time to complete, what with the little free time I have for this.

148 Posts
Discussion Starter · #8 ·
Been a while since I returned here. Work is progressing slowly but surely. I've added a photo showing the second largest piece that I've finished cutting and sanding. Needless to say, it took a long time to complete, between by job, and being restricted to making noise only when my roommates are out. The good news is that there are only 7 more pieces left to complete! The bad news is that one of those remaining pieces is the large circle. Considering how long it took me to prefect the edges of the small circle, I'm really not looking forward to the large one.

148 Posts
Discussion Starter · #10 ·
Hey everyone, I'm still alive and things are still progressing. Just reaaaaally slowly. I'm currently in the process of sanding down the final template. Hopefully my roommates will be away tomorrow and I can finally finish this long and painful process. I can't wait to rip into the acrylic with the router! Hopefully that will be a much faster process. In the meantime, got two additional photos added.

148 Posts
Discussion Starter · #12 ·
More updates. I've also redone the first post to make things a little more organised. I now have a contents list for each major part of the project. Everything under each section is in chronological order, but the major parts themselves aren't necessarily in that same order (if that makes sense).

I am currently working on the acrylic now, using the templates to guide my router. I had very little time to make noise before my roommates returned, so by the time I had completed the final template, I only had mere minutes to make a very small start on the acrylic. As such, I managed to successfully complete about 15 cm of the edge of the large circle. The edge is good, and the shape appears to be a near perfect replica of the template, but I'd need to make more progress to see what sort of quality I should expect of the final product.

148 Posts
Discussion Starter · #13 ·
Design layer section updated. First acrylic piece cut and routed.

Using what I've learned from this first attempt, I should be able to make cutting and routing the rest of the pieces more efficient, so hopefully I can make even more progress.

Casual Watercooler
352 Posts
I applaud your ingenuity with the design of the voltage control box, but making that box out of wood is going to be a huge fire hazard. I would recommend you look into getting a pre-fabricated control box from likes of companies like Hoffman. Something like a NEMA-4X rated flat enclosure with latches. This particular enclosure is UL-508A listed, which is the electrical safety standard.

148 Posts
Discussion Starter · #15 ·
Once I have the water temperature gauge and lighting controls added to the panel, I plan on lining the inside of the panel with sheet metal for that same reason (which will also make my wiring relevant to grounding much simpler). I haven't decided on the specifics such as sheet thickness or material yet, since I've got several other things to do in the meantime.

I would have loved to make a completely metal housing, but without welding tools, I'd be limited to sheet metal, L-braces and bolts, which would probably look horrible. At least with my hybrid method, I should have an acceptable middle ground.

One thing I am concerned about, though...At the moment, my housing is built in such a way that the strength at which the latch holds it against the mounting effectively causes it to be airtight. Even the holes drilled for the cables are exactly the right size for them to fit through without any gaps. Now, if it's airtight, that would snuff out any flame pretty quickly due to a lack of oxygen. However, without any ventilation, would there be any issues for four circuits with 13A max on each as far as heating during normal use is concerned?

Casual Watercooler
352 Posts
A quick way to test whether your make shift enclosure is air tight is to pressurize the box, and then spray the other side with a diluted Dawn mixture and check for bubbles. The enclosure I linked is made out of fiberglass so you would not have to do welding and the box is wall mountable. Of course you can customize the box as you see fit, but that would provide a turn-key solution (looks like you can get non-metallic hole seals with the box) that shouldn't get you in trouble with your tower management if they randomly check on you.

148 Posts
Discussion Starter · #17 ·
Alrighty, I'll consider it and see if there's something like that I can find here. For the time being, I'm going to leave it as is, since nothing is live just yet. Still have plenty of things to do before I need to worry about this. Thanks.

148 Posts
Discussion Starter · #18 ·
Design layer updated. Making progress on the acrylic pieces. I only have 4 more acrylic pieces to route out; the two wing bases, and the top and bottom circular sections.

Before I can proceed with the wing bases, I first need to join together the three pieces making up the template. Couldn't make the template in one piece, so it had to be segmented. Now it needs to be joined back together again with a combination of glue and flat braces. The same applies to the bottom circular section. The template is in five pieces. However, the finished acrylic product will have to be segmented, as I won't be able to fit the joined template onto one board of acrylic. Luckily, the top circular section can fit on one, so that can stay whole.

One challenge I believe I will face with the top circular section is making sure that the final joined piece is symmetrical. Even a small angular error in joining one piece to another can completely throw that symmetry out the window. Guess I'll cross that bridge when I get to it. For now, I will be working on joining the wing base template pieces together.

I'll update once more when all the acrylic pieces are complete. Until next time!

148 Posts
Discussion Starter · #19 ·
Hi, I'm still here! Took a while to find time where my free time did not coincide with my roomates', but I had to call sick for a flight which left me with the last two days on my own. So I decided I was going to finish that pesky acrylic once and for all! Design layer section updated.

Unfortunately, I made another mistake during the routing of the small circular section. I forgot that I was not supposed to route the edges where the three pieces would join to each other, so now the left and right pieces will need 1cm removed from the offending edges. Thankfully I didn't do the same thing to the center piece, so I'm hoping that those 2cm I remove will not noticeably affect the design. If they do, I'll make an elongated center section, since it's the smallest and easiest to recreate.

So next step is to print out the exact same design that I had when I first started the project. This will be attached to the base layer. I will select appropriate places to drill holes into the acrylic for the mounting bolts, then place the acrylic over it's relevant shape on the paper, so I can mark the holes on the paper and drill through the wood. Before mounting, I think I'm going to flame polish the edges. Will first check for propane torch prices though.

In other news, I plan to change the control panel on the side. Going to integrate everything into the perimeter, instead of having the box sticking out. Looks better, and despite sacrificing a bit of practicality, I doubt I'd need to access the innards that often. Plus, I can design a metal safety cover on the other side, which will contain all the wiring, and improve fire safety.

And I've been sitting in my room for over three hours now with my respirator on. I really went all-out with the router this time, and despite the ventilation, the acrylic fumes are still strong in here!

148 Posts
Discussion Starter · #20 ·
Design section updated.

I've mounted all the feather pieces. There may be aa delay with the next update though. I'll need to find someone to help me mount the larger pieces. I'll also need to purchase more M6 screws and connectors for my standoffs/mounting.

We're getting there, people!
1 - 20 of 38 Posts
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.