My DIY Isolation Box design

Posted: Sun Oct 05, 2008 9:25 am

MadMax - Yeah, my design is really just based on my own personal brand of common sense. I have no idea if it'll work. I did research on the internet for other people's previous efforts and tried to learn from their mistakes, but in general I'm just guessing what might work. If experts can chime in I'd be a lot better off.

Currently my computers have these intake fans:
One 120mm fan (Computer 1)
Two smaller fans--not sure size (computer 2)
Two very tiny fan (external hard drives)

And the exhaust fans are:
Two 120mm fans (computer 1 & 2)
Two typical power supply fans (I'm guessing 70mm?) (computer 1 & 2)
Two tiny fans (external hard drives)

I have two other computers that I rarely ever use--they're more like legacy machines, so I don't consider them part of my normal setup.

Brien wrote:

I would build a quiet computer OR get a rack mount purposed machine.

That's not a possible choice. Even if I replaced the entire cooling system of my computers, I still have the hard drive noise to deal with, and when I shop for hard drives I cannot tell if a drive will be quiet or loud. I also have upwards of 10 hard drives in my computers--it's impossible to build quiet systems with that many hard drives.

Getting a new computer is also not an option. I had just put together a new quad-core machine and it'll be a long time before I add new machines to my setup.

Posted: Sun Oct 05, 2008 11:19 am

"but in general I'm just guessing what might work."

I here ya.

Posted: Sun Oct 05, 2008 7:01 pm

Over-engineered? No! We all know the value of starting with a quiet recording environment, and silencing a computer without killing it with heat takes some doing.

Beautiful design and great dedication and thought.

I build custom computers for a living and recording is a hobby. I too prefer to quiet the computer rather than put it in a different room, which is inconvenient to say the least.

Before I get into your design, take a look at the source. Depending on the types of computers you have, there are ways to quiet them. Are you willing to modify the computers a bit? How much are you willing to spend?

I don't know what kind of computers you have, but let's say you have a modern system that adheres to common internal physical layout standards.

· The PSU (power supply) tends to be noisy. There are some nice quiet PSUs out there, my favorite of which is the Enermax Modu82 625w. It has enough power to run a system with demanding audio software. You can find prices on Google or Newegg. Other almost-as-good options are the Corsair 520HX/620HX and the lower-powered Seasonic S12.

· Hard drives can be replaced with quieter models. Good choices are the Samsung F1 750GB and 1TB models—lots of space, inexpensive price per gigabyte, and quiet. They can also be decoupled from the computer case simply by suspending them with elastic bands inside the computer's optical drive bays. Additionally, there are products that completely enclose hard drives and then fit into the optical drive bay, such as SmartDrive and NoVibes.

· Fans can be replaced with quieter ones. Here are some wonderful 120mm fans that run at 28dba and are very inexpensive: http://www.jab-tech.com/product.php?productid=3851 These would be excellent for your isolation box as well. 28dba itself does not sound very impressive, but it is the broadband noise characteristics of these fans that makes me prefer them. That is one thing that can't easily be quantified on a spec sheet. They also put out, in my testing, a little more than their rated 47CFM of airflow.

· CPU (processor) HSF (HeatSink Fan): most modern CPUs come with a loud HSF unit for cooling. An excellent replacement that is compatible with modern AMD and Intel CPUs is the Xigmatek S1283. For best results, replace the 120mm fan included with it with one of the 120mm fans I linked to above.

· Video cards also tend to be quiet loud. Your system may have integrated graphics, in which case that's not an issue. If you do have a graphics card, however, the Arctic Cooling Accelero or ACCELS1 will be a good choice, depending on what kind of card you have.

· If you want to go more in-depth, you can replace the entire computer case. For this, look to Antec's P182 case. It's designed to be quiet.

Making some or all of these changes to your computer will ensure a quieter source to silence in the first place, as well as better cooling for the system. I do all my own testing of components and have chosen these for my own line of quiet computers. Why not share what I've learned? Smile

As for your isolation box. First off, I want to know what program you used to make that. Very Happy

I'm designing one myself, but with some 2D photo editing software.

My testing shows slightly negative pressure inside a contained system to be ideal for cooling. Equal intake and exhaust pressure takes second place, and in last is positive pressure. Positive pressure makes air swirl a bit stagnantly inside the contained system, equal pressure is neutral, and slightly negative pressure helps ensure that air goes where you want it to, rather than seeking alternative exits. I do mean "slight," however: extreme negative pressure complicates things in a bad way.

You've got the right idea with the fan ductwork to manage acoustic noise. I have some suggestions. First of all, with modern bearing designs, fans themselves don't make much noise; generally, air turbulence is the most prevalent sound. To make discernible noise, air has to hit something relatively solid—like the walls of your ductwork. If each of three fans has its own passage, the total duct surface area that the air hits is greater than if there is one larger passage. So having the three fans side by side, blowing down one larger duct, would create less noise from turbulence.

But then, a large duct may carry a lot of noise from inside the box. So you might want two levels of ducting: the final large duct, and a "pre-duct" that runs parallel to it (perhaps underneath it, with a 180 degree bend joining them) beginning at a small, high airflow velocity opening to the computer chamber, and widening to meet the final large duct at said bend. Turbulent air noise at this opening will be greater than at the final duct opening; lining the duct with sound-absorbing material should kill
this noise, especially at the 180 degree bend. It's the venturi effect (Google if unfamiliar) used to kill sound. Smile

The advantage of having a small opening to the computer chamber won't matter much if sound is bouncing around in the computer chamber so much that all sounds eventually hit and travel through the small hole anyway. Lining the computer chamber with sound absorbing material as you have already done in your design will be essential; otherwise, you'll just have created a nice ported subwoofer box, and guess what ... your computer is the subwoofer. Razz

You can apply the same ideas from the isolation box's exhaust system to reduce sucking noises from its intake system.

You may want to replace some doors with solid walls. Less convenient, but also less chance that there will be a tiny whistling leak somewhere in your little slight-vacuum box.

So that's acoustic noise. As for as structure-borne noise goes, you have a good idea with the absorbers under the floor of the computer chamber. However, I would take out all the wood supports between the "false floor" and the rest of the box, and replace the absorbers with polymeric [or similar] foam. Put your absorbers to good use lining all surfaces (ceiling, floor, and walls—except intake, exhaust, and cable holes, and the floor beneath the computer[s]) in the computer chamber, and sandwich the polymeric foam between said ceiling/floor/walls and the rest of the structure. Basically then, your entire inner computer chamber will be decoupled from the isolation box.

And you've just killed most airborne and all discernable structure-borne noise while ensuring good airflow. =)

I'm also toying with an idea for the ceiling ... but it might be.... well, never mind lol, I'm debating whether or not to try it.

By the way, once you go to these levels, you might consider throwing your computer cases away entirely and just mounting your computer parts right to the inner chamber of your isolation box. You might be surprised by how complicated it isn't.

Hope this helps! I'd love to go through and make sure this all makes sense and there are no typos, but ... no thanks. Razz

Props to anyone who actually reads all of it.

Posted: Sun Oct 05, 2008 8:47 pm

TIGR wrote:

As for your isolation box. First off, I want to know what program you used to make that. Very Happy

I'm designing one myself, but with some 2D photo editing software.

Try google sketch-up its fairly easy to use and free.

Posted: Sun Oct 05, 2008 9:45 pm

I would place fans inside the unit and use baffling to prevent the noise from exiting the box. So instead of having a straight air tunnel, use acoustic baffling staggered along the tunnel to prevent reflections and make sure the fan keeps a positive air pressure to the box. On the air's exit use baffling again but no fan. That would require only one large fan (bigger than 120mm).

And reading the post above, if positive pressure causes stagnant zones inside the box, you can use something to "break" the air. I guess the concept would be similar to a diffuser.

Posted: Sun Oct 05, 2008 10:18 pm

My apologies, it isn't over engineered. Engineering take thought, maths and the practical application of physics to create something useful and effective without being extravagant.

This isn't over-engineered. Just over cooked.

I read all your post Tigr, only noticed one extra dot, no biggy. However this "By the way, once you go to these levels, you might consider throwing your computer cases away entirely and just mounting your computer parts right to the inner chamber of your isolation box. You might be surprised by how complicated it isn't." is as cool as ice.
I've seen it done across a wall, as in all pieces of hardware mounted to the wall with cabling making cool patterns. Functional art.

Posted: Sun Oct 05, 2008 11:02 pm

I added airflow arrows to the side view image, since it seems some people are having a hard time reading the airflow direction and fan positions.

TIGR wrote:

Beautiful design and great dedication and thought.

Thank you. It took a long time to figure out the design I want, and drawing it took a while too. I drew it all by hand in Photoshop with a Wacom tablet (I've been working as a graphics professional for many years now). I've also used 3D software such as Wings3D and Sketchup before to visualize designs, but I wanted to do it in 2D this time just to see if it's any faster/easier. I think next time I'll do it in 3D since I have a habit of wanting to merge Photoshop layers since I hate huge stacks of layers, and once you merge layers you lose the flexibility that Photoshop gives you in the first place.

Quote:

Before I get into your design, take a look at the source. Depending on the types of computers you have, there are ways to quiet them. Are you willing to modify the computers a bit? How much are you willing to spend?

I've tried that route but in China it's nearly impossible to find specific models of products that are even a little bit esoteric in nature. I've tried and the store owners always just end up laughing and saying "We don't ever order stuff like that since people here don't know or care about such things." I couldn't even find temperatur sensing 120mm quiet fans here. If I ask for special order they will quote me some outrageous price. I could try to order them from the States or elsewhere but the import tax and shipping will be a lot. I can also trying bringing them into China in my luggage during my trips back to the States, but timing then becomes an issue, since we'd only travel back to the States about once or twice a year.

Quote:

Making some or all of these changes to your computer will ensure a quieter source to silence in the first place, as well as better cooling for the system. I do all my own testing of components and have chosen these for my own line of quiet computers. Why not share what I've learned? Smile

As much as I'd love to replace all the parts, it'll just end up costing too much. Just replacing all the hard drives alone will kill me.

Quote:

You've got the right idea with the fan ductwork to manage acoustic noise. I have some suggestions. First of all, with modern bearing designs, fans themselves don't make much noise; generally, air turbulence is the most prevalent sound. To make discernible noise, air has to hit something relatively solid—like the walls of your ductwork. If each of three fans has its own passage, the total duct surface area that the air hits is greater than if there is one larger passage. So having the three fans side by side, blowing down one larger duct, would create less noise from turbulence.

But then, a large duct may carry a lot of noise from inside the box. So you might want two levels of ducting: the final large duct, and a "pre-duct" that runs parallel to it (perhaps underneath it, with a 180 degree bend joining them) beginning at a small, high airflow velocity opening to the computer chamber, and widening to meet the final large duct at said bend. Turbulent air noise at this opening will be greater than at the final duct opening; lining the duct with sound-absorbing material should kill
this noise, especially at the 180 degree bend. It's the venturi effect (Google if unfamiliar) used to kill sound. Smile

Gotcha. So if I merge the outside ducts into one big opening, but keep the inner ducts separate like I have now, and merge them somewhere (probably before they reach the half point of the full travel length). I'll have to think a bit about how I'd do it. I was under the impression that the harder the bend in the air path the more turbulence you get due to more clashing of air against walls. That's why I used slight slants in the air duct so that the fans do not line up with the intake/exhaust--but are not so dramatically bent in the pathway to cause more clashing.

Quote:

The advantage of having a small opening to the computer chamber won't matter much if sound is bouncing around in the computer chamber so much that all sounds eventually hit and travel through the small hole anyway. Lining the computer chamber with sound absorbing material as you have already done in your design will be essential; otherwise, you'll just have created a nice ported subwoofer box, and guess what ... your computer is the subwoofer. Razz

You can apply the same ideas from the isolation box's exhaust system to reduce sucking noises from its intake system.

You may want to replace some doors with solid walls. Less convenient, but also less chance that there will be a tiny whistling leak somewhere in your little slight-vacuum box.

I would prefer to have the doors since it'll safe me lots of headaches down the line when I need to do anything to the computers. I'll make sure the rubber seals are as tight as I can make them, and I'll do double seals if necessary.

Quote:

So that's acoustic noise. As for as structure-borne noise goes, you have a good idea with the absorbers under the floor of the computer chamber. However, I would take out all the wood supports between the "false floor" and the rest of the box, and replace the absorbers with polymeric [or similar] foam. Put your absorbers to good use lining all surfaces (ceiling, floor, and walls—except intake, exhaust, and cable holes, and the floor beneath the computer[s]) in the computer chamber, and sandwich the polymeric foam between said ceiling/floor/walls and the rest of the structure. Basically then, your entire inner computer chamber will be decoupled from the isolation box.

And you've just killed most airborne and all discernable structure-borne noise while ensuring good airflow. =)

Originally my plan was to use a metal wire stand of some kind for the computers to sit on, so the floor absorbers can do more work by being exposed. I then thought maybe just using a wooden one wouldn't change things too much since I'm more concerned about low frequency and surely the low frequencies will be able to cut right through such a thin wooden board anyway? I could of course make the wooden board into wooden strips spaced apart to expose the floor absorbers--that'll be easy to do. I just thought later when I want to slid/rotate the computers while doing work on them, the gapped floor will make it hard since the computer's legs will get caught on them.

Quote:

By the way, once you go to these levels, you might consider throwing your computer cases away entirely and just mounting your computer parts right to the inner chamber of your isolation box. You might be surprised by how complicated it isn't.

As crazy as I seem to be, I don't think I've gone off that deep yet... Very Happy

Thanks for the suggestions--much appreciated.

BTW, those that suggest I replace the case, hard drives, or cooling system--it'll will be very problematic because the convenient shopping that we do in the west is far harder to accomplish in China. They don't carry any products that even slightly esoteric in nature--it's all very common and mainstream products all the way. I can't even find quiet 120mm temperature sensing fans here--just normal ones.

Posted: Tue Oct 07, 2008 5:08 am

I did a quick edit to show another possible design for the intake/exhaust tunnels. The yellow is the new design and the blue is the imagined airflow. I'm guessing this might be quieter than the previous design since the air tunnels become progressively wider thus less friction, but still contains one soft angled bounce to slow down/diffuse the air stream. Does that make sense?

Posted: Tue Oct 07, 2008 8:51 am

Hey Lunatique, I just have a minute here to submit an idea:

You and I don't want to move our computers into another room ... but why not dump noise into another room/outside? Long flexible ducts could take the noise wherever you route them. Aperture noise wouldn't matter with them out of the room so you could use some powerful fans for incredible cooling. In some climate zones, running the ducts outside in winter could provide ... quite incredible cooling. While long, this ductwork could be devoid of sharp bends that impede airflow.

Just an idea that hit me. Smile

I'll post again later when I have more time.

Posted: Fri Oct 10, 2008 8:35 pm

Good idea making it mobile, That way you can rearrange if need be. Still I do not see why you need so many fans when one would provide all the air you need. Think of a bathroom or kitchen exhaust fan, some of them are designed to run very quietly, and should be suitable for the box. An idea to reduce noise is to widen the intake/exhaust ports and create a venturi to lower the air pressure that creates turbulence while entering and leaving the box. It has to do with air impedance matching, so look it up. Combine that with some mid/high frequency baffling and you have a silent design. Since some of the guys here know more about this than I do, I would take their advice and design the flow for negative pressure.

Posted: Fri Oct 10, 2008 11:17 pm

Just a coupla' odds and ends...

TIGR has a valid point to consider about the extra walls.

Each additional surface in the enclosure adds a coupla' percentage points of db from the air resistance/turbulence.

Logic tell me (be afraid.. be VERY afraid) that following the principle that adding surfaces will add air noise, I would suspect that the entrance and exit baffles should be up for removal.

There's also a few things that most everyone is overlooking here... and that is the fact that there are hundreds of fans on the market, with all kinds of fan ratings... rpm, blade angle, diameter of the fan, diameter of the motor hub, and general velocity of air (as rated in cfm or cmm).

Again, you want to move the air as either a 1::1 or as a negative pressure at the input to pull air through the enclosure. Calculate the total cfm ratings of the existing fans in your computers.

That rating, PLUS a couple of percentage points is what you want to pull through the box... IMHO, camsr's mistakenly assuming some rather broad statements of fact that are just that... assumptions. Yes, it could take one fan (I all but guarantee it'd noisy as hell) or it might take 12. No one can actually know how many fans until you calculate the required airflow and you have found fans to move that volume of air at the rate that makes sense.

I'm just hopin' that Rod will jump in here and verify the calculation for determining the actual air column.

This topic sponsored by: