5
\$\begingroup\$

This question is really aimed at computer cases, but I suppose it also applies to other PCB cases. Most of the Computer cases I come across are aluminium, or else some other metal; As I understand it this is because of the following properties:

1) thermal, a metal case will nicely dissipate heat applied to any point, and conduct it to the cooler areas (outside). Also, doesn't burn (though many cases have plastic on them too)

2) RF, a metal case will shield from outside interference, and also keep inside EMF from interfering with other things.

3) strength, to keep heavy equipment in-place and protect them.

4) earthed, add a level of electrical protection by having an earthed metal case around components.

My question it this: why can't other materials be used if these are the only concerns? Other materials may be cheaper, easier to get or work with, lighter and importantly cheaper.

For example, Wood has good properties for #3, it can be strong, and servers don't need to be portable much, so it only needs to be as strong as not falling apart when just standing there. #4 isn't really needed, #1 is important, but most computers use fans and heatsinks (rarely the case itself is used as a heatsink) already, so with well planned airflow the cases needn't also be thermally conductive (maybe thermal sensors may be more important though); which leaves #2.

Now, would it be possible to use some kind of metallic tape on the insides of the case? I've seen thick RF tape, but that stuff's expensive. As far as I know, aluminium (baking) foil is actually mainly plastic with a thin layer of aluminium. I've seen cheap aluminium tape (for the roof) but I assume that's much like the baking foil; Just how thick does the layer need to be, and can it be done cheaply? Could I even use something like chicken wire?

\$\endgroup\$
5
\$\begingroup\$

Regarding the thickness of metal required for shielding: the conductor used as a shield has to be thicker than the "skin depth" of the metal at the frequencies you're trying to block out. According to a textbook[1] I have here, 1 micron is enough in practice.

[1]: Engineering Electromagnetics, Inan and Inan, 1999, p. 322.

\$\endgroup\$
  • \$\begingroup\$ Nice answer, thanks, although what are the frequencies that is based on? I wonder if there is any way to test this? \$\endgroup\$ – CH Apr 24 '10 at 22:03
  • \$\begingroup\$ You can calculate skin depth here: microwaves101.com/encyclopedia/calsdepth.cfm The depth decreases as frequency goes up. One implication of this, if I understand it correctly, is that low-frequency noise, like 60 Hz hum, is difficult to block. \$\endgroup\$ – pingswept Apr 24 '10 at 22:30
  • \$\begingroup\$ So according to that calculator, to block 60hz a single sheet of aluminum baking foil isn't thick enough correct? \$\endgroup\$ – Earlz Apr 25 '10 at 2:09
  • 1
    \$\begingroup\$ Another good question, what kind of frequencies do we need to block? That is, What kind of radiation might there be externally that may be harmful to the computer, and what frequencies will the computer itself generate that need to be prevented from escaping? \$\endgroup\$ – CH Apr 26 '10 at 15:22
  • \$\begingroup\$ Using the values 280.2 uΩ*cm for resistivity and 1.00002 for rel. permeability (for Aluminium) with the calculator posted above, I get: 1Hz -> 0.843 m, 60Hz -> 10.876 cm, 20KHz -> 5.957 mm, 1MHz -> 0.843 mm, 1Ghz -> 26.64 um, 2.4GHz -> 17.2 um, 5.8GHz -> 11.06 um, 10GHz -> 8.42 um. Although, someone please check my working, I can't be sure of any of the above. – CH 2 mins ago \$\endgroup\$ – CH Apr 26 '10 at 15:23
2
\$\begingroup\$

The biggest reason is mass-manufacturability. It's so cheap and easy to stamp aluminum enclosures once you've done the tooling, while wood requires milling each piece. But if you're doing something one-off or for aesthetic reasons, the reproduction cost shouldn't be a concern.

For the other reasons:

  1. Thermal - you're right.

  2. RF - chicken wire would allow high frequencies through. If the wavelength of the EM wave is smaller than the opening in the metal, it can get through. Aluminum foil doesn't, AFAIK, have a plastic layer. It should be good enough. But it tears easily, so you have to be pretty vigilant to get good coverage. Also, you need some spring mesh or something around any flanges or lids etc to ensure the electrical connection carries around the corner. Note contact should be continuous all the way around any opening. In theory, the thickness of the metal shouldn't impact the shielding (as long as the metal doesn't tear) but I don't know enough RF engineering to really assert this.

  3. Strength - if your box sits on the floor, wood should do fine. If it mounts in a rack, you'll have to decide what to do about the mounting flange. The shear strength of wood is not as good as metal, so you might have to fabricate the flange out of metal, and use a lot of bolts to connect it to the wood. Wood is heavier for a given volume so the flange will have to carry more load.

  4. Earth - I'm not too sure about this one. I guess lightening strikes are the biggest reason you need to worry about this, and I'm not sure how much protection the metal case would give you anyway in that situation. Anyway, your RF cage needs to be connected to ground so maybe you'll just end up with some charred aluminum foil to clean up. If someone pokes something metal through the case and makes contact with the power supply, thicker metal on the RF cage would help.

\$\endgroup\$
  • \$\begingroup\$ But how thick does the metal need to be? and Would the roof tape be adequate? I was just gonna use screws to ground the foil, and to make connecting across corners. I'm just surprised this kind of this isn't on the net already (while the mineral-oil fish-tank PC is), So I was afraid there was a good reason that I'd missed; come to think of it, what RF protection does the fish-tank PC get? \$\endgroup\$ – CH Apr 23 '10 at 1:45
  • \$\begingroup\$ In regard to earthing, I meant if there was an exposed contact or something; If you push a screwdriver in there it would probably short with the earthed case first. \$\endgroup\$ – CH Apr 23 '10 at 1:46
  • \$\begingroup\$ I updated my answer, but as you'll see I don't really know the answers to your additional points. And I know nothing about the fish-tank :) \$\endgroup\$ – mtrw Apr 23 '10 at 6:33
2
\$\begingroup\$

As far as I know, stamped sheet aluminum is the cheapest way to mass-produce metal cases.

If you want a pretty wood case, I see 4 options:

  • no shielding at all. The electronics in a wood case has almost exactly the same EMI/EMC as if it were sitting out in the open air, or in a glass fish-tank of oil. My understanding is that all PC motherboards emit so much RF that such a case would fail FCC testing, making it illegal to sell in most countries.
  • put a metal box inside the wood case -- perhaps by covering the metal case with veneer.
  • cover the worst RF emitters with tiny metal boxes attached to the PCB. The metal box should be attached to the ground plane of the PCB, preferably at many points all the way around the perimeter. An isolated metal box often makes emitted RF interference worse.
  • spray the inside of the non-metal case with a conductive paint, and somehow attach it to ground.

The last two options (small metal boxes, conductive paint on the case, or often both in combination) are extremely common in many consumer electronic devices that come in a plastic case such as cell phones, televisions, radios, etc.

I suspect that metal roofing tape or aluminum foil would work just as well as or better than the conductive spray paint, but I've never seen it in production. (I'm guessing that it has a higher labor cost to install properly than the conductive paint).

A brief google search gives me (in no particular order):

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.