Slashdot Mirror
Ask Slashdot: Wooden Chasis and EMF
Red Leader
asks: "Hi. I'm writing in the hope that some electrical
people will be able to help me out with the nitty gritty
aspects of shielding a computer case. You see, I'm building
myself a computer case out of wood. I have already built
one, but it was merely a wood case (replacing the plastic)
which overlayed a standard metal chassis. Well, that was
too heavy, and I also want to use my own layout for the
'guts' of the box. So aside from heat, noise and grounding
(which i've pretty much figured out) - I'm worried about
the electromagnetic interference aspect of this new
machine. It's most likely going to be a dual Celeron based
Ultra-Wide SCSI2 system - which I think (!?!) will
generate quite a large electromagnetic field. I've been
thinking of shielding the case by lining it with this with
copper
fabric or something similar - but don't know how it
sizes up to a solid metal case. Any help would be greatly
appreciated - and I'll be sure to post pics of the case
(well it's actually more than that) in progress and when
completed.
Besides shielding your box, it's nice to put ferrite RF chokes on the cables, especially the video cable but preferrably all cables running out of the computer. The cables work as antennas, potentially leading some of the RF noise out of your shielded box and emitting it. Some cables have the choke molded in, if yours doesn't, get the clip-on kind from the Radio Shack. The more turns of wire through it, the better it works.
I also use shielded CAT-5 wire on my 100-base-T network, and I use that same wire for my DSL connection, from the point where it enters the house.
Thanks
Bruce Perens K6BP
Bruce Perens.
Oh. Right. You mean "potato chip bags." Heh. I forgot they say things weird on the other side of the big pond.
I seem to remember that a 100 MHz signal has a wavelength of 3m (c/100Mhz). Likewise, a 500 MHz signal has a wavelength of 3/5m = 60cm. Even at 1GHz, the wavelength is 10cm. I'd say you're fairly safe using a mesh (at least for now).
Okay, I've read everyone's comments, and I think that the copper mesh should provide sufficient shielding. I don't know how overboard on that I want to go, considering what else I'm trying to do...
As I said, I hope to put up pictures of it in progress, and when complete. I just need to get to a scanner...
There was one comment about heat, and one about the movement of wood. To those guys: I've got that pretty well covered.
As of my current design, insulating against sound is my primary concern. I'm going to line the entire case with a really dense (1 lb. per sq. foot) dB blocking fabric, as well as baffle any ventilation openings. The sound concern is directly related to my ventilation solution. I've long since decided that a filtered positive pressure system is the only way to go. The implementation on the first case wasn't so hot - but I've learned from my mistakes. Both cases are supplied with cool air from a large 90mm AC fan (80? don't remember) of all metal construction. It's not a normal case fan, but rather specifically a high power 'pushing against resistance' fan. The only problem is that it's pretty damn loud, hence the complicated sound protection.
As for the wood movement: Absolutely. I really enjoy joinery and fine woodworking, and realize that one can't make anything without keeping all the grain essentially parallel. For those who don't know, wood expands and contracts with the weather by varying amounts across its grain (perpendicular to the stripe pattern). If you were to make a joint in which the grains were not expanding and contracting in the same directions, then your joint will literally tear itself apart with time. Then what you made will crack, and or, fall apart. Case in point - on the old case I have a mahogany and maple board (grain running up/down; two strips of mahogany with a central 'racing stripe' of maple) working as the face of the box. In it is a slot for the CD tray. I took off the plastic tray's end cap, and mounted a pice of ebonized maple on it. That piece has grain running left/right (naturally), and hence doesn't significantly expand left/right. The face does move, though, and in the winter I was surprised to see that the extra room I had given it on either side wasn't enough, and that the CD tray wouldn't close unless I took the end cap off!
I can go more into depth, and talk about this for many kilobytes, but I'll save that for the web site where I'm going to put all this up for everyone.
Thanks to everyone for their help and comments (sorry, aluminum foil people, I'm going with a little more style than that... this *IS* a piece of furniture, after all). If you have any questions, feel free to talk to me. I love bouncing things off people, and I've given this project a LOT of thought and planning.
Thanks!
Unlike acoustic noise, it's a lot of work for people affected by radio interference to figure out first of all that it is happening at all (rather than that their radio or digital wireless system is broken) and where it it is coming from.
Causing interference like that is also against FCC regulations, and people can track you down and have you fined.
Unless you live on a desert island, if you run your computer open, sooner or later you are going to cause other people a lot of unnecessary headaches. So, please shield your computer.
I think the question shouldn't be whether you are likely to get caught, but whether you are causing someone else trouble. If you run your computer unshielded, there is a good chance that you are.
I would like a "music server" in my house. That is, connecting an old computer with no peripherals to a music system. I would control it by telnet or a web interface. Obviously, good sound quality is a high priority, and that might very well mean shielding.
Actually, though, what I would really like is an all-digital sound card, with any analog processing completely cut out of the loop. Ideally, it would output a digital signal that a good home audio system will decode, such as Dolby Digital, if that is possible. Or maybe the DAT format.
Anyone know how, with digital signals, shielding, or some combination of the two, you could make the best, cleanest audio signal possible? This wouldn't just be for MP3s, but also higher quality stuff. Any advice on this greatly appreciated.
Vidi, Vici, Veni
Well, the only one I know of (that's not a professional card) is the SB Live, full edition. With it's digital I/O card and the extra optical digital card, you can input digital, process the signal digitally, and output digital. It can even mix digital audio, internal digital CD, digital MIDI, one analog source, digital SP/DIF, and optical digital all at the same time. I use my computer as a piece of my stereo system, with the internal digital CD, two digital in's (MiniDisc and an external digital out CD player,) and MP3; all being processed for better sound (if you're running [god forbid] Windows, try using the "Concert Hall" environment for classical music.)
And, of course, it's 48Khz digital SP/DIF out is great for recording to MD or DAT. (Or going to a high-end reciever)
Another non-functioning site was "uncertainty.microsoft.com."
The purpose of that site was not known.
You might just be forgetting Maxwell's time-dependant equations, then (specifically $\rot{\vec B}-\frac1c\pder{\vec E}t=\frac{4\pi}c\vec J$). As someone said further up, the problem with meshes is that they effectivly act as high-pass filters, quietly letting through any radiation whose wavelength is less than $\approx1/\sqrt{5}$ of the mesh length. You can check this easily: look through it :) .
I find that most of the vox-populi notions about Faraday cages come from radio people, where you only have to worry about very low frequency (long wavelength) radiation.
Hey, where is the `submit as TeX' button?
Wire mesh is just fine for basic emf and rf shielding...
It should be roughly contiguous, although, that is hardly worth pulling out a micrometer of magnifying glass.
Just make sure each seperate piece of mesh when assembled makes good electrical contact with the frame? (if you have some sort of framework on this) or other conductive bus. To test if this is roughly adequate,
1. get a multimeter (measures voltage/resistance and usually current, sometimes mislabeled a voltmeter)
2. set it to measure resistance or continuity.
3. touch 1 probe to one portion of mesh, the second to a point opposite...as in as far as you can get inside the case from point 1
4. look at the resistance readout.
5. It should read 0 ohms (if yours has a floating needle, it may be a bit off, not to worry) If it reads anything > 1 ohm, you don't have a significantly conductive material to be useful as a shield/ground...
(For those who will argue here...I just finished measuring the foil from a pack of marlboro light 100's that came out to 200 mOhms , or 0.2 Ohms, in case you cared.)
Now for the final key point. Ensure that the power supply is properly grounded, your shield is properly grounded, and that your new case is plugged into an adequately grounded 3 prong receptacle.
For clarification...if you don't plug it into a grounded receptacle, what you've just configured is essentially a floating common attached nicely to a power cord which sould make a passable antenna for rebroadcasting all the stuff you just collected.
If you need to add a better ground to your house...i.e. you get lines in the tv when you run the vacuum...talk to an electrician unless you are comfortable in a circuit breaker box.
You can get a 3-4 foot copper clad steel rod, with copper wire, to act as a grounding rod. have the electrician hook up the copper wire to the ground bar in your circuit breaker box, then you attach it with a clamp to the grounding rod.
Note..you're going to need a sledge hammer...preferably a big one...
Apply the sledge hammer to thetop of the grounding rod until only about 3-4 inches sticks out of the ground....
now attach that wire, and your house ground should be much better.
Whew...wrote a lot more than I planned...but you get the idea.