This project will take a while to complete....I don't have all the pieces yet, and there are some details to work out. These initial photos will give you a good idea of what's going on. The case is a full tower Lian Li, sold at Best Buy under the name Rocketfish. It's being closed out, and available as I write this for about $50 at many stores. It seems to be plenty big for this new (to me) design, involving two boards in one computer.

The airflow will all be from below, using three fans......two 120mm fans under each CPU heatsink, and a third 80mm fan under a yet to be constructed copper HD cage. The swing-out board will be a low-power setup, using a PICO PSU. The main board will be in the standard position....with a std PSU mounted internally on top, further back in the case, , blowing toward the upper vent holes. This case was designed to have a PSU on the bottom. Since that lower section of the case is now a filtered intake chamber, this setup will be different from the norm. As I get further along, I'll post more photos. They're all under 50K since I'm on dial-up.











Stay tuned...more to come.

One more to 5000!

Always interesting to see where you go with these projects.

Very interesting project, looks like you got an elegant layout too.

Any thoughts on how to connect them to the monitor? DVI kvms are still horrid, maybe use an analog connection for one of the boards?

As always.... the Devil is in the details. I might use a standard KVM, since I have several. The option is to use vga from the second board, and DVI from the main board. The monitor I'll use has both inputs, and automatically selects whichever it receives first. I don't mind using a different mouse for each board, but I only want one keyboard.

I don't expect to run both boards at the same time, but the cooling setup should handle that possibility without problems. All the fans pulling air from the filter chamber need to be running at the same time....but at different rpms is ok. If one is turned off, there would be a recirculation problem, without resorting to a one-way flap door of some sort or maybe divided intake paths......no way I'm going that route unless there's a noise issue.

No doubt I'll have some unforeseen problems.... but that makes it fun.

This post is indeed #5000 for me.....indicating I have too much time on my hands, or I talk too much.

Everything fits.....so far. I tried to measure closely before I mounted anything, but with close clearances, you never know for sure until you tighten everything down. I'm using two low-profile heatsinks.....a Thermalright XP-120, and a Scythe Shuriken. When the second board is swung inward, the two heatsinks are off-set by a few inches, so the airflow should be ok even if both boards are running at the same time.

The two intake fans are also off-set to maximize the clear space area above each board. The bottom tray to which the fans are mounted is a piece of relatively stiff outdoor carpet. It's sitting on the aluminum "L" brackets I added to the case. The fans are zip-tied to the carpet.... no vibrations.

There's enough clearance for a full-sized video card on the main board, and a low profile video card on the second board. The video cards to be used are still under consideration.

more to come.....

WOW ... awesome idea & thread, looking forward to the end result.

Thanks.....I've had this idea in my head for a while, but never found the right-sized case for such a project. I had been considering making the whole thing out of wood and steel panels, but this Lian Li case will be much lighter. Not to mention cheaper, and easier to construct.

I suspect there are other dual MB computers out there.....and Thermaltake makes a high-dollar case just for the setup. But I can't recall ever seeing one made out of a standard case.

Fan control......somewhat complicated. Neither board has a built-in fan control. But I've tested each board separately so I have a good idea what's necessary. The bottom intake fans need to run and be separately controlled whichever board is running....so. I have an AC power strip with a remote controller. Each board will plug into this strip, along with a dedicated 12V power supply for the intake fans. So in order to get either computer to turn on, the power strip will need to be turned on (which will turn on the intake fans). Then I'll need to turn on whichever board I want to use. Fairly simple.

The intake fans will have their own controller....An M-Cube MiniNG. This has two separate channels on auto-control. Whichever board is in use will cause it's own intake fan to speed up when it gets hot....the other intake fan will idle.

The swing-out board will run it's Scythe fan at a fixed voltage (about 8V), set by a manual controller. The main board will have it's Thermalright heatsink fan, controlled by a Noisemagic NTM3. The PSU for the main board will be stock....a PSU with a bottom intake located at the top of the case, directly above the secondary board. The HD for the main board will have a dedicated fan controlled manually. The secondary board will use a laptop drive, so no additional cooling is necessary.

That's it.....no exhaust fans, other than the PSU fan which will be back inside the case, simply blowing at the exhaust vents. Sounds complicated, but I have all the pieces already, and have tested everything separately. So I should be ok here. The next set of photos should make everything clearer.

Great project
I had similar thoughts, but thought it would require case to have at lest ~25 cm.
Looking forward how it turns out

Somewhere on Xtreme Systems there's Lian Li V2000 with uATX board in the bottom.

I fired up the main board (temporary hook-up) to check the fan setup and the filter chamber for noise. Checks ok. The fans for now are the Lian Li fans that came with the case. They're not bad at all at lower voltages.

I also am going to try a new HD cooling setup, which might be able to run without an extra fan. It's a Logisys HD cooler box.....with the top left open. There's a square hole cut in the bottom carpet/tray. The lower fin section of the HD cooler extends down into the intake chamber, so it will get airflow, as will the top of the drive from the right intake fan. I'm not sure if this will be enough airflow to avoid a dedicated HD fan....I'll see. There's enough room in that position to add a second HD if necessary.

Next on the list.....mount the laptop HD for the second MB.

Why two machines in the same box? (just curious)

Very interesting project... I'm looking forward to seeing how this turns out.

Just a quick question about the mechanical construction - is the "secondary" board simply swinging on the hinges? I'm guessing that it 's stopped from swinging into the case cavity by the travel on the hinges, but wouldn't it be better to put a proper stopper on the top and bottom?

What are the general specs of the two systems?

As to why? Many answers....mostly because I like to try out new ideas, and this is something really different. I already had most of the pieces, so when this case deal popped up, I went for it.

The case will handle any size main-board.....but I needed a new home for an Intel D865GLC, with a P4-3.4EE, easily the most powerful setup I ever owned. The swing-out tray will handle most MATX or smaller boards. When I find the right board (for me), I'll replace the Intel/P4-2.4 that's in the photo. The only limitation I can see....the position relationship between the two CPU heat-sinks. They have to be able to be positioned with some side-by-side clearance when the second board is swung inward. The Scythe HSF seems to be ideal for this setup.....it's super low-profile, and comes with mount hardware to fit almost anything. In fact...it will run without it's own fan, just using the airflow from the lower intake fan. I tried this out already, but I don't know if this would be the best setup. (more experiments on this later).

The swing-out tray will have a positive stop of some kind. It will now swing inward until it hits the other board. I suspect when the VGA cable and the other cable bundle is hooked up, there won't be much movement possible, but I'll all add real stop of some sort......one of the many details to work out.

Last night I made a suspended laptop drive mount for the second board. It is located directly under the 5.25" cage, and is close enough to the board to only require a very short cable. But there are other places to mount a drive for this second board (it doesn't have SATA ports however). Before I concern myself too much with this second board, I want to get the whole setup running.

Next up....construct a PSU mount for the mainboard. Might be tricky.

I decided to use my X800AIW card (my hottest-running card), with it's custom fanless heatsink. This involved repositioning the heatsink....it used to extend past the end of the card. Not much trouble. In this position the card should run super-cool. The heatsink is actually 3" above the lower intake fan, although from this perspective it looks closer.

So.....AFAIC, this setup is capable of running a hot video card, maybe better than a standard setup. The heated air from this heatsink should go straight up toward the upper exhaust vents, mostly by-passing the CPU heatsink and fan. A deflector panel could make it even better. Remains to be tested....

Further along.....I mounted a small MATX Fortron PSU on the top, blowing toward the vents. This exhaust mostly by-passes everything else. It's lower intake fan is directly over the right intake fan, so the temp of the intake air is close to ambient. Normally this Fortron will not be running when the second board is running (it'll be on a PICO).

Then I added a small duct to the Thermalright XP-120. This duct draws it's intake from the airflow from the right lower intake fan. And the duct prevents most of the recirculation problem that's common to "blow-down" heatsinks. An additional benefit....the heated air from the video card and it's heatsink, by-passes the intake to the Thermalright. And....since the duct is made from soft materials, it softens the noise from the Scythe fan.

I added sensors to various places (DD5 temp sensor panel). So now I'm running the main computer with the ability to monitor every important temperature (12 points total). Looks good so far. The CPU idles at 35C @about 800rpms, with the two intake fans @5V. There's a NoiseMagic NTM3 controller on the CPU fan.....this increases the fan speed gradually as the temp rises. At max CPU usage....50C @ about 1250rpms at a 24C ambient.

More testing/tweaking to come..

I spent the last few weeks tweaking the main board setup....the swing-out board is not running yet. I made a few changes. I now am running two drives for the main board, relocated into the intake chamber, below the pull-out filter. There are two intakes....one in the front, and one in the rear. This pulls air over each drive....they run cool, about 6C over ambient.





I now have three intake fans.....the front fan will only run if the swing-out board is being used. This simplifies the fan control a bunch.



I finished the upper exhaust vent.....which is now a louver plate, with a noise deflector. This setup is amazingly effective at blocking any noise from going forward.





I have added sound dampening to the interior, including roofer's tape on the side panels, some carpet panels, paxmate, safety-mat foam, and various foam weather-stripping. I lined the inside of the front door with a carpet panel. This door is similar to a P-180, except that it won't warp. I reversed the opening direction. This is typical Lian-Li.....high quality, nice hinge setup, and positive stops.



I've been trying out different fans....it seems this setup doesn't need fast intake fans. Right now the main setup runs on one 800rpm S-Flex, and one 1000rpm Enermax. These both run about 5-6V except when stress-testing.... at which point they both go to 12V. Still quiet.

This setup idles at 105W.....and can go to 290w running CPUBurn and ATI-tool. So it's no light-weight when giving off heat. I've also been tweaking the CPU fan ducting.....not satisfied yet. And the wiring is temporary....I'll make it neater when the second board is running.

More to come.....

Update.....I spent a bunch of time tweaking the XP-120 heat-sink I'm using on the main MB. I tried some tweaks to the heat-pipes, as well as numerous ducting changes. Not satisfied.....I can do better. So I just placed an order for a Scythe Zipang HSF.....from Coolerguys. I hope it will fit the board...hard to say. Since it's taller than the XP-120, it probably won't fit with the swing-out plate in the photos.....

No worry.....I ordered a smaller VIA board as the secondary MB. This will enable me to shorten the swing-tray by almost 2", which should give me enough clearance. This new board is rated about 12W.....low power, and fanless. Hope it works out.

Stay tuned....

Still waiting for the new pieces......but I think I've got the wiring straightened out for switching between the two computers. The connections for the swing-out board will route behind the board toward the front bezel. They go through the bezel, along the aluminum channels and out the rear of the case. For now they form a little dongel.....Vga, Cat5, USB, phone line, audio.







Then...in order to use a single Monitor, keyboard, mouse, audio, and external usb devices, I got this Belkin KVM switch. Despite the negatives in the Newegg reviews, it works fine (so far). There's a little lighted external switch, which changes between computers in about three seconds, I tried an external powered USB hub, and it switches that also. Amazingly, the computers do not have to re-detect the various devices when you switch. I hope it works this well when the new VIA board is running.

More to come.

I've been using one of those in a client's office for, oh, a year or so now. Not my first choice for a KVM, but they needed one immediately, and it was available locally. The KVM has worked flawlessly. Good luck with yours!

Major changes......The CPU cooler is now a Scythe Zipang. It works better than the XP-120 that was in there. The fan is an Aerocool 140x20mm, that maxes about 1250rpms. With an ambient of about 25C, it idles about 850 @ 36C, and maxes about 1050 @ 46C. Quiet enough....good temps. This heatsink can be mounted four different ways by off-setting the mount location on the heatsink. I have it as far upward as possible. This enabled me to use the swing mount for the second board as is, even though the Zipang fan is bigger.

The swing-out board is now a VIA CN board, running at 1000ghz. The laptop drive mounts directly above the pci slot.....easier wiring. This board runs with a 15W current draw....maxes at 20W. The center lower fan only runs when this board is on....it's a 800rpm S-Flex running at about 6V. The whole setup running on this board is completely silent at all times unless your ear happens to be right at the upper exhaust vent.

As I suspected the wiring is out of control. There's no practical way to avoid this with two boards. There has to be some slack in the swing-out harness, and it's entire wiring loops forward, down, and then back to the rear. Lots of wires. I did make sure as many wires as possible are out of the airflow, running on the inner sides of the case.

More to come...

It's coming along nicely.....with both boards running at once, very little heat is added to the setup. The temps of the main computer are affected very slightly...with about a 1C rise in CPU temps and a 20rpm increase in the CPU fan. The VIA board runs cooler when the other board is running because the intake fans are on. I eliminated the intake fan for the VIA board. It now has a little fan directly over the heatsink running at 6V.....that's it, using no intake or exhaust fan, just convection to get the heat out. (when the VIA board is running by itself)

What's a new computer without a little bling?





Switching between the two boards is still on the table. I'm going to avoid a KVM switch, because I want to use the DVI output from the main board. So.....I'll have two cables to the monitor, and a separate USB 2.0 switch for the keyboard and mouse. I can run two audio wires to the Bose speakers, since is has a built-in mixer for the two inputs. Should work ok....

Still under construction....

I love the front panel artwork. It's so Asian....like me!

Bluefront, very nice work! I do have some questions for you though. I just purchased this Rocketfish case for ~$50, and while I am generally pleased with the aesthetics and layout, I am a bit disappointed in the fit of the side panel and the slop of the front door in the closed position. Both lead to very pronounced rattle when any type of impact force is applied to the case. The side panel just doesn't fit very tightly, as I can tap around the edges and cause the panel to smack against the frame behind it (there is obviously excessive clearance). Likewise, the ball-bearing style locking mechanism on the front door allows the door to rattle freely against the case.

It seems that you have applied extensive damping and sound absorption treatment to the case; however I didn't see any specific mention of the problems I described in the previous paragraph. While I believe some visco-elastic damping applied to the center of the side panel may help reduce the amplitude of the resonant response, it does nothing to eliminate the source of the excitation--the excessive clearances caused by the poor side panel fit. I suspect there will still be some rattling, though somewhat muted. Did you take any specific actions to improve the side panel fit? In the past, I have rigged up cheaper cases with cotton squares and tape to eliminate rattles from poor side panel fit, but this less-than-elegant solution doesn't seem appropriate for such a nice case.

Similarly, how did you constrain the motion of the front door while in the closed position?

If you had any suggestions they would be greatly appreciated!

I haven't gone into specific detail about this case because there are few people on SPCR who have one. But I can answer questions....

As to the front door....I flipped it over as you can see. And I installed a section of 1/4" carpet to the inner section of the front panel. The fit of mine is perfect.....it doesn't move or rattle when closed. Perhaps you could tighten it slightly by installing some thin black electrical tape over the areas where the spring-balls depress. Or maybe add spacers to the spring setup....but I haven't had to do either. Maybe there is an adjustment for the whole spring-ball module?

Yeah the side panels are a loose fit, and made worse by their light weight. I lined the panels with roofer's tape (like Dynamat), and then added another layer of Spire sound-dampening dense foam (from Jabtech). This increased the weight of each panel by 50% (I actually measured). Then I installed strips of felt along the edges of the case where the panels should touch, but don't. You can see the strips in the third picture up.... What you can't see is the fit of the removable filter. It is slightly wider than the case, so when I install the side panel, the filter pushes up against the Spire foam, and acts like you're pushing your hand on the panel. That fixed the loose fit of that side panel......the right panel got similar treatment. It seems to be a tighter fit....doesn't require anything other than the felt strips.

The side panels are much more difficult to install now.....but don't make noise and don't leak air. More questions?

This picture shows the felt strips better on the far left side....it also shows the foam edge of the fan mount frame that also pushes against the foam of the side panel.



Oh I forgot....I also added a small foam square to the back edge of each spring-ball module on the front door. If you do this, be certain the foam is thin, open-cell. This puts a slight amt of outward pressure on the door, and stabilizes it slightly.

Thanks for the thorough response. That's a very good idea to use felt strips along the edges where the panels contact. Also, I will see if I can make some adjustment to the door on my case so that it's held shut firmly.

Do those felt strips have adhesive on them? If so, where can I get some of that stuff?

The felt you see in the picture isn't readily available. I got it through a Nissan car dealer.....it's used mostly in the dashboard area between plastic pieces that rub together, around the audio module where it touches the dash, etc. Yes it does have a sticky back. I can find the part # if you're interested. I comes in a relatively large roll that could line several computers....not cheap and special ordered.

But you can find sticky-back thin felt in craft stores. It comes in 9"x12" squares (Presto Felt). You can cut thin strips with a sharp blade....cheap. about $1 a sheet.

Thank you! I have one more question just for clarification. The small foam square that you applied to the back edge of each spring-ball module on the front door: Did you unscrew the spring-ball modules and place the foam between the door itself and the modules?

No the foam is on the outside of the module. It just gives the door a soft stop against the bezel. You could substitute small foam strips on the case side....instead of the felt. This stuff comes in rolls at the hardware stores. I think I've seen some about 1/8"x1/8"...... used to seal around windows and such. You need the open cell type that compresses to almost nothing. The gap you want to fill is not too great.....if you use stuff too thick, the panel won't go back on.

Thanks for the tip! I went to the craft store and picked up some supplies. I cut out tiny adhesive felt squares and stuck them on the spring-ball modules. The door now closes perfectly with no rattles. I also applied felt strips to make the side panel fit tighter. It took a bit of trial and error, but the panels don't rattle any more. Although it is more difficult to take them on and off.

Next on my to-do list is to move the door hinges to the opposite side as you did. I almost always have headphones and audio inputs going into the drive-bay panel of my Audigy 2 Platinum ZS, and it would be much more convenient if the door opened in the opposite direction. Is there anything I should keep in mind when I do this? It looks like I just need to unscrew a bunch of stuff from the door and the case and flip them around.

Yeah that's it......just be careful you don't loose any little pieces. The pivot bearings for the hinges are easy to drop. You do know the door will then be up-side-down? The lock won't work and the LEDs can't be seen with the door closed. That's it....a screw-driver is all that's necessary.

There's another thread about this case at EVGA Forums. Quite a few people are posting in this large thread..... some have water cooling. Nobody there (except me) seems to care about noise however.....

Thanks for the link. No, I didn't realize I'd have to install it upside down. I thought it could be installed right-side up by flipping everything but the door.