DIY case designed, renders

[niels007]

Hey SPCR readers,

This is my 'final' design for the new case. Made out of 15mm MDF, I would line the inside with aluminium foil (EM shielding, hopefully..) and put knobbly damping foam wherever it fits. I haven't modelled these in 3D yet.

Thoughts:
------------
Harddisks can operate safely in each 90degree position. Hanging them vertically next to the mainboard in an (nearly) enclosed section of the case would isolate its noise as good as possible, and the 30cm U profile should take care of temperatures.

Not sure on what is best, the powersupply can either be ducted to the side panel for direct 'cold air intake' or, without the duct, extract hot case air out instead. With cardboard I can cover the rear ventilation holes to see which combination works best.

I have no room for a huge case, it measures (h x w x d) 565x240x430mm like a bigger tower but certainly not huge.

I managed to use an old case for donor parts, notably the mainboard back plate and 5.25/3.5'' combo bay

The pics:











All comments are welcome!

[Gorsnak]

I wouldn't worry about the aluminum foil. My wood case leaks less RFI than my lcd monitor does, according to the transistor radio test. Make sure you have some sort of isolation mounts for those fans on the side, or you're likely to run into resonance issues.

[hvengel]

I wouldn't worry about the aluminum foil. My wood case leaks less RFI than my lcd monitor does, according to the transistor radio test. Make sure you have some sort of isolation mounts for those fans on the side, or you're likely to run into resonance issues.

The transistor radio test (TRT) may not tell you very much about the RFI coming from a computer. Any RFI from a computer is likely at frequencies that are on the order of gigahertz or several hundred megahertz and a typical transistor radio operates at a few hundred KHz for the AM band and around 100 MHz for FM. TRT is a good test for broadband RFI sources like power lines or your monitor but PCs are more likely to a create RFI at a limited number of mostly higher frequencies (not broadband).

The other reason you want RFI shielding is to keep RF out of your PC. May not be needed in most locations but you never know someone down the street may be a ham radio operator that occationally puts out his legal 1.5 KW max transmit power or there may be a AM or FM transmitter a mile down the road that puts out 50KW and this could really do a number on your PC if it is not shielded. If this happens it will be your fault not the owner of the transmitter.

[BillyBuerger]

First I would watch HD temps. Even with the U-channel, it's not going to get much airflow and could still eventually heat up.

Second, I'm wonder where all the hot air is going to go? If you duct the PSU, you've got all intake fans and not much room for hot air to get out. If you don't duct the PSU, the PSU would be exhausting all the hot air which would cause it to speed up. I see you put a bunch of small holes along the top. I just don't think that would let enough air out.

Last, you've got a direct path for the noise from the fans. I assume you'll have the case on the floor or under a desk or somthing. So it shouldn't be a real big deal. But just something to consider.

I'm working on a custom wood case design myself. What did you use to model that? Good Luck!

[Gorsnak]

The transistor radio test (TRT) may not tell you very much about the RFI coming from a computer. Any RFI from a computer is likely at frequencies that are on the order of gigahertz or several hundred megahertz and a typical transistor radio operates at a few hundred KHz for the AM band and around 100 MHz for FM. TRT is a good test for broadband RFI sources like power lines or your monitor but PCs are more likely to a create RFI at a limited number of mostly higher frequencies (not broadband).

The other reason you want RFI shielding is to keep RF out of your PC. May not be needed in most locations but you never know someone down the street may be a ham radio operator that occationally puts out his legal 1.5 KW max transmit power or there may be a AM or FM transmitter a mile down the road that puts out 50KW and this could really do a number on your PC if it is not shielded. If this happens it will be your fault not the owner of the transmitter.

Well, it doesn't interfere with my cordless phone or my cell, either, even when they're held right next to the box. At that point it becomes a question of why you should care if there's some radio noise coming off the computer. As to keeping RF out of the computer, you're going to have to explain to me just how RF is going to "do a number" on a PC. Thousands of people run their cases with doors off/open, which is just exactly the same degree of shielding as a wood case, and I've never heard of trouble arising from this.

[niels007]

About shielding, I don't really worry about it. As mentioned there are loads of people who run without side panels, and many plexiglass cases are sold commercially, certainly not having any shielding!

About the design, I said 'final' i.e. changes could be applied depending on the comments in this thread

Gorsnak:
I'm using the same sort of approach now with side panel mounted fans. I didn't say that they're isolated with four bits of foam on the edges, so yep I'm definitely isolating them. (Mind you, nexus 120mm fans don't really vibrate )

BillyBuerger:
About the HD temps, its something of an unknown really! You could well be right, in which case there is room for a 92mm fan in the bottom. I never had direct airflow over my harddisks and temperatures rarely exeed 40degrees c. Fingers crossed, but thanks for the headsup: I do have the 92mm fan backup plan.

About where the hot air is going, that again is unknown.. Well almost! My current setup is similair except for the PSU duct. I can still get away with a very slow PSU fan and since this is the cheapest 300W qtec, I would say a new 80% efficiency Seasonic or Fortron wouldn't have to spin up that much extracting the hot air out of the case. You did make me think again though, about having another 120mm exhaust somewhere, at least as an option.

The PSU duct is more of an experiment. At the back there is a 150x90mm opening next to the psu, slightly blocked by the round duct. Just above the mainboard there is a 100x40mm hole at the back. I don't know how clever air is in finding its way out. Theoretically I'd say the case is under slight pressure so holes in the rear should still be 'found'. The holes in the top of the side panels are only a small area compared to the gaps in the back.

I realise the direct air path for noise of the fans. Even in my crappy case that I have now, I don't have to spin them above 500rpm at which point they are inaudible from underneath my desk.

I wanted to use Solidworks for modelling as I really want to get the hang of that, but I made this in Autocad which I've used for such a long time that I'm just quick and efficient with it, despite the program really not being that good in 3D.

Thanks for the comments, keep em comming!

Niels

PS:
I thought about sliding the PSU over to the other side. No duct would be required if I want it to take 'out' of the case, and I could flip it 180degrees to make it extract case air. Downside is that all the wires will be up by 130mm and in either case the wires aren't neatly dealt with at the mainboard side where they're easier to manage.. This would leave ample space for a 120mm exhaust next to the PSU which isn't obstructed by a duct..

[Gorsnak]

(Mind you, nexus 120mm fans don't really vibrate )

That's what I thought, too. But if I crank my Nexus exhaust fan up to 12v, the whole case hums, and that's with some (obviously inadequate) isolation. My solid pine board construction is going to be more resonant than your MDF, but not by all that much. I ended up soft-mounting virtually everything with moving parts.

[IsaacKuo]

I'm also concerned about the hard drive airflow. As I see it, you've got two fans feeding air into the case and creating positive pressure. One blows air into the case through the CPU heatsink, and the other blows air into the case through the GPU heatsink. The result is that CPU and GPU temperatures should be good, but everything else starts off with warm air.

This air gets further warmed up by the chipset and other motherboard electronics...then it goes and exists out of exhaust holes in the top edge as well as through the vertical hard drive chamber.

My main concern is that the hard drive will be cooled by unhealthily warm air. Hard drives don't generate much heat so they don't need much air, but they do need cool air. This is the opposite of a CPU--a CPU generates a lot of heat, so it needs a lot of airflow, but they can run at relatively high temperatures so it's okay if that air is somewhat warm.

For this reason, I think it makes more sense to put the CPU at the "end of the line" rather than the "front of the line". IMHO, you should consider flipping the fans so that the side fans are exhausts rather than intakes.

BTW, if you're hoping for a good decrease in hard drive noise from placing it within that vertical "semi-enclosure", prepare for disappointment. It will have some effect, but it takes a ten-fold decrease in sound energy to merely halve the perceived sound level. It doesn't take much of an escape hole to let a lot of sound leak out.

[JohnnieStalker]

If this happens it will be your fault not the owner of the ham transmitter.

were closer than you think

[cansan]

Looking good, I hope you can build it without too much hassle. (Eagerly awaiting the finished pics.)

Why don't you try the power supply at the bottom back part of the case? The fan facing downward (to outside air) and the exhaust facing the back. In terms of heat, it could be almost like having the psu outside the case. It would reduce the height of the case as well.

Another random idea:
For the two side fans... IF you have passive heatsinks on CPU GPU: Instead of having them directly open to the outside, maybe have them on the partition (Between the mb and hd) blowing to the back and then have holes at the bottom and sides of the hd compartment to supply fresh air. The back wall would also need holes for this to work.

Good luck, please post pics when you're done!

[Elixer]

What about reversing all three fans? This would directly remove the heat from your 3 main heat sources, your cpu, gpu, and power supply. This way your hard drive would be getting cool air instead of warm air as well.

[niels007]

Thanks for the thoughts and comments guys!

Elixer and IsaacKuo

What you both think will happen is that the rising air will go down through the harddisk bay as partly drawn here:



And instead of the 'boom' the air would escape at the bottom? Can you be sure that this will happen in an ever so slightly pressurized case? I was pondering on this and hoped that the 'energy' from the air heated by the HD would still go up and escape through the case ventilation holes.

Originally I had this in mind:



which definitely solves the problem but also emits HD noise through the front of the case.. If you all say that the air wants to go down through the HD 'chamber' in a slightly pressurized case, I will go for the second option.

Currently, my harddisk is suspended in a 'normal' case and there are two 120mm fans in de side panel. I do hear some HD noise escaping but it is so little that even option 2 would probably be better having a much smaller ventilation hole area, and the source (HD) being 150mm down below.

Edit:
about reversing the fans. I have two small problems with that. Firstly turbulent air on the 'blow' side is slightly more effective than the laminar 'suck in' side. Secondly, a fan is louder on the 'blow' side for the same reason. Even if there would be no difference in efficiency, which could well be true, I would be exposed to the noisy side of 2 fans.

[niels007]

Here is the other option I had in mind, basically more conventional with an extra case fan AND the psu extracting the air.



The PSU is still reversable so it can take in ambient air and not case air.. which would leave the case slightly pressurized, which means the HD question remains..

It doesn't matter what you have to design, there are always too many variables and options!

[IsaacKuo]

Yes, the "hot air rises" effect is really weak, compared to the amount of pressure even an undervolted silent fan creates. Against two fans? No contest, I think. (I'm assuming the PSU is ducted in its own little airflow world, and thus out of the picture.)

Things would be different if you had a carefully balanced neutral pressure setup going on. With a push-pull setup with a balance between intake fans and exhaust fans, the internal pressure would be neutral. In that case, heat from your hard drive would indeed create an extremely small amount of hot air driven airflow. This will be an extremely small amount of airflow! A hard drive doesn't generate much heat, so it can't power much airflow. I would NOT bet on this airflow being sufficient to cool itself.

As for your concerns about placing exhaust holes in the front--don't be. I'll bet you won't be able to hear the difference. Noise is really skilled at finding its way out of exhaust holes even if you think the pathway is very indirect. Basically, I suggest you mock up something where the only thing you put in it is your hard drive. Use an old AT power supply and remove or block the fan so your only noise source is the hard drive itself. IMHO, that's the only way you'll find out what will actually work in terms of reducing hard drive noise. (Even then, you might later on be shocked by loud seeks.)

Anyway, I think your case design is basically very good and will allow you to experiment with a lot of simple modifications to see what actually works out in practice. Theory is theory, but there's nothing like actually fiddling with the physical reality to figure out what works. Examples of things to fiddle with:

1. Try flipping the CPU/GPU fans. Theoretically, this should very slightly increase CPU/GPU temperatures while rather greatly decreasing HD temperature--a good tradeoff. In practice...well, you'll get to choose either way and see what's better!

2. Closing the top edge exhaust holes vs closing them up. The really aren't necessary, IMO. In fact, I'd try out the case first without even drilling them out. By eliminating these holes, there's a little bit less place for sound to escape (not that this will make much difference, but it might help). More significantly, eliminating those exhaust holes forces all air past the hard drive. That'll help hard drive cooling regardless of the direction of the fans.

3. PSU duct? Or not? Again, you've got your choice. Without the duct, you can also decide whether or not you want to flip the PSU fan.

Also, you could expriment with eliminating some fans altogether. For example, if you lose the PSU duct, you can run all airflow with the PSU fan pulling air out of the rear. This sets up negative pressure in the case, naturally sucking air in through the CPU and GPU ducts, as well as sucking air past the hard drive.

[IsaacKuo]

About your theory about the "noisy side" of the fan...I'll bet that the fans will be a lot quieter blowing air out of the case rather than into those ducts.

Those ducts are conically restricting the airflow, which the turbulent flow from a fan's exhaust will not like one bit. The exhaust airflow is somewhat helical, and your conical ducts will force this helical airflow into a narrowing screw. My gut tells me that can't be good, and it'll make turbulent noise as well as create restrictive backpressure.

In contrast, with the fans flipped to blow air outside of the case, the airflow within the ducts will be smooth, with no helical component. The helical component of the airflow will be unrestricted in the outside air, and will generate no noisy turbulence.

But like I said, there's no substitute for expermenting with it in reality!

[hvengel]

snip
As to keeping RF out of the computer, you're going to have to explain to me just how RF is going to "do a number" on a PC. Thousands of people run their cases with doors off/open, which is just exactly the same degree of shielding as a wood case, and I've never heard of trouble arising from this.

There are all kinds of things that can happen to all kinds of electronics when there are powerful near by transmitters. The vast majority of people do not live near a powerful transmitter and therefore do not experience any problems.

But consider this. One of the ham bands is right at 2.4 GHz. If the ham living across the street from you fires up a kilowatt transmitter on this band and you are running a 2.4GHz processor in an unshielded case do think that perhaps this could cause a problem? There is another ham band at 220 MHz. If you have overclocked your system by pumping up the memory bus to 220MHz don't you think that a strong RF signal on this band could cause a problem.

The point is that this kind of thing is rare as the right set of things has to happen for it to be a problem. But it can and does happen even if you don't personaly know anyone that this has happened too. However I do know many ham radio operators and many of them have had computer problems because of strong RF even in cases where the computer appeared to be properly shielded.

Hal AC6VZ

[Gorsnak]

One of the ham bands is right at 2.4 GHz.

Seriously? That's a cordless phone frequency. Actually, there's a ton of consumer wireless stuff that uses 2.4GHz. Seems like a recipe for interference to me. Anyways, the gist of your anecdotes here seems to be that in the vast majority of cases there's no reason to be concerned about EMI shielding, and in cases where it is a concern, the shielding provided by a metal case might not be sufficient anyways. I don't see that this gives anyone a reason to concern themselves regarding EMI when building a non-metallic case, unless they actually live across the street from one of your hypothetical ham operators.

[niels007]

I've read somewhere, probably SPCR or the E-group, that a processor running at 2.4ghz has little or nothing to do with RF of the same frequency.

Anyway, if there is a RF / shielding discussion, please start a new topic and keep this one for snazzy wooden cases. (and its design flaws )

[JohnnieStalker]

AC6VZ
HAL V ENGEL
3690 COWELL RD
CONCORD CA 94518

I'll park in front with my high power ham mobile and lets see if I get into his computer

[Tzupy]

niels007, I had many thoughts about a similar configuration, but couldn't find time and energy to do the drawing, so: congratulations!
My concept requires two 120 mm fans in the back (one ducted to the CPU heatsink), the PSU moved about 8-10 cm to the front and turned 90 degrees (exhausts air through the right case panel), and the GFX cooler similar to the AC Silencer, but with vertical fins, ducted to the fan on the left side panel.