Aluminium RF/EMI shielding?

[thejoggingmat]

Can aluminium sheets/foil be used for RF/EMI shielding?

Basically, I need to design an interface for a proprietary module that receives signals sent over the phone line and outputs a RS232 stream. I need to integrate an AC 12+5+3.3v SMPS, a uC (likely arduino so it can be rapidly troubleshot in the field) and a wifi/ethernet module in one enclosure (likely laser cut acrylic). As only less than 5 of them will be fabricated, I will need inexpensive parts that can be bought in small quantities in the local corner shop/hardware store.

I have read multiple conflicting posts over the web that Aluminium can be used to shield components against RF/EMI, with nothing really definitive. A random YT video even 'proves it' with an oscilloscope.

Copper/steel etc sheets are out as I cannot purchase them in small quantities. Only aluminium sheets or foil is available.

If aluminium cannot be used, what other possible alternatives can I use that can be found commonly?

[calexanian]

RF shielding yes, EMI, not really so much. Typically you want Mu metal for that, or steel at the very least.

[dom0]

RF shielding yes, EMI, not really so much. Typically you want Mu metal for that, or steel at the very least.

EMI ~ electromagnetic fields ~ RF. Mu metal is the only really effective way of shielding magnetic fields (noted lack of "electro-"). Good conductors like aluminium and copper are very effective against electric and electromagnetic fields (hence most RF assembly enclosures are die-cast aluminium), but barely make a dent in a magnetic field.

Most instrument enclosures are made from aluminium sheet and die-casts since quite some time, it has been popular since at least the 50s.

[thejoggingmat]

RF shielding yes, EMI, not really so much. Typically you want Mu metal for that, or steel at the very least.

EMI ~ electromagnetic fields ~ RF. Mu metal is the only really effective way of shielding magnetic fields (noted lack of "electro-"). Good conductors like aluminium and copper are very effective against electric and electromagnetic fields (hence most RF assembly enclosures are die-cast aluminium), but barely make a dent in a magnetic field.

Most instrument enclosures are made from aluminium sheet and die-casts since quite some time, it has been popular since at least the 50s.

I think I do have some ancient PC cases lying around; if I could fish one out that has a thin sheetmetal body would it be suitable?
 They attract magnets if I could recall correctly. That's about the closest to any metal instrument cases as I can dismantle (everything's ghastly plastic!).

Aluminium comes with its own problems: Almost impossible to solder anything on it. I'd want to connect the metal shields to ground just to be safe. Would bolting tightly with nuts/bolts & washers be good enough to conduct to ground?

[Ian.M]

Bolting *ANYTHING* to aluminium and expecting good conductivity is problematic.  For long term reliability you need bright zinc plated hardware, freshly cleaned aluminium, and specialist electrical jointing compound. 

Contrary to popular belief, aluminium can be soldered quite easily.  You need to abrade it through a pool of liquid flux to break the oxide layer on the surface and exclude oxygen to prevent it re-forming, then take a hot high power iron and form a small pool of solder on the surface.  Scratch with the iron tip till the solder starts to 'take' and work outwards till the desired area is tinned, then wipe off excess solder and solder the wire or part on to the tinned area normally.  It took me about two minutes in total to solder a wire to a piece of kitchen foil, using a pure Rosin in IPA flux pen, 60/40 solder, a fibreglass pencil and a 100W soldering gun.  Some people prefer to solder it under a heat resistant oil. e.g.
He'd have had an easier time of it if he'd abraded it through the pool of oil as bare Aluminium is so reactive that even a few seconds in air will reform a nano-scale oxide film, but that would have gunked up his fiberglass pencil.  I don't mind a little thin rosin flux on the one I keep for PCB work, but really wouldn't want anything oily on it.   Self-adhesive aluminium tape would be easier to work with than kitchen foil, because its thicker and stays where its put, but the ordinary stuff has non-conductive adhesive so you'd have to solder all the joints which would be a right PITA.

My recommendation is to make your own steel screening cans similar to the ones you see in commercially produced equipment:

*  Cut up a large rectangular tin can for a source of thin tin plated steel you can bend to shape and solder to.  (Check it *is* steel with a magnet first!)

* Use paintstripper to get the markings off the outside.

* If you want to retain good solderability, you need to wash thoroughly, and immediately hot air dry and wipe with thin liquid Rosin flux after stripping.   

*Food cans often have a protective lacquer on the inside which is a PITA to get off so is best avoided.  Thinners and solvents cans are free of lacquer and don't need cleaning out so are the easiest to work with.

* You can get good sharp bends if you cut a wood block the size shield you want, thicker than the depth, and clamp the cut out shape and the block in a soft-jaw vice then bend each side in turn over the edge of the block. 
* Mock up the design of the screening can you need using thin card to mke a cutting template as its a PITA to solder on separate corner flanges or mounting legs later!

Another alternative is to solder up screening boxes out of bare copper clad PCB.  Its excellent for HF work but is far worse than tin plated steel at low frequencies.

See: https://en.wikipedia.org/wiki/Skin_effect#Material_effect_on_skin_depth
For screening you want the skin depth to be as little as possible with a highly conductive surface, so tin over copper plated mild steel is about the best possible cheap screening for all-round use, with plain tin plated mild steel coming a close second.

[T3sl4co1l]

Yes, bolted connections are fine.

Aluminum of useful structural thickness (0.5 mm or more) is fine shielding for most frequencies and sources you'll have problems with (i.e., >= 150kHz).  If you know you have a much more difficult noise environment (e.g., it has to operate inside a huge induction coil, for some bizarre reason..) you might need more shielding (ferrite or mu metal, heavy steel, etc.).

Even aluminum foil is perfectly effective, when the interference has a high impedance (more electro- than magneto-).  Coils and high currents generate magnetic fields, while proximity to wires and high voltages generate electric fields.  YMMV.

Mechanical bonding (screws, spring stock, metal gasketing, etc.) is fine.

More important is that your input/output signals have their grounds connected very nearby (if not directly, then with capacitors), otherwise RFI conducted along the wires will be led into the enclosure.  Less of a problem if you have good filtering inside, but more is better.

BTW, connecting to POTS requires suitable isolation and protection -- the line can have transients of some kV on it during extreme conditions (nearby lightning strikes, for example).  Your circuit needs to handle these safely -- even if it doesn't survive them, it needs to not catch fire and explode.  Interface circuits usually use an isolation transformer (all the important telephone stuff, i.e. ringing and voice, is AC signals) with a relay to handle DC (on/off hook signal) and GDTs (gas discharge tubes) from tip to ring and from each to GND (which has to be earth GND, which means you need a GND coming from your AC line input, so use a three wire cord).  I don't know exactly what type of GDT is required, but you should be able to find out by searching for appnotes on the subject.

Tim

[Ian.M]

Experimentally, I  tried tinning the rim of a rectangular aluminium sardine tin by the same abrade through the flux method.  The 1" length I was working on tinned quite nicely.   You could use a double sided PCB with a wide ground track topside that follows its rim outline, well stitched to a bottom layer ground plane by frequent vias, and simply solder the sardine tin to the perimeter track.  Signals in and out would go through a via near the perimeter to the bottom and cross the perimeter on a track there. The section to be screened must be done as a single sided + solid ground plane layout as any bottom side tracks break your shielding.   If you need a smaller can, eat anchovy fillets!   

[f5r5e5d]

Tin plated steel is still common in canned goods and biscuit tins, hardware stores may have galvanized steel sheet and often both Al and Cu roof flashing sheet rolls, hobby stores have brass, copper sheet

and I would like to add my usual warning that most hobbyists have no business worrying about Mu Metal shielding - I designed Strain Gage amplifiers with V gains up to 8000, looked at the results with 16 bit converters - and never used Mu metal - just low carbon steel enclosures

[thejoggingmat]

BTW, connecting to POTS requires suitable isolation and protection -- the line can have transients of some kV on it during extreme conditions (nearby lightning strikes, for example).  Your circuit needs to handle these safely -- even if it doesn't survive them, it needs to not catch fire and explode.  Interface circuits usually use an isolation transformer (all the important telephone stuff, i.e. ringing and voice, is AC signals) with a relay to handle DC (on/off hook signal) and GDTs (gas discharge tubes) from tip to ring and from each to GND (which has to be earth GND, which means you need a GND coming from your AC line input, so use a three wire cord).  I don't know exactly what type of GDT is required, but you should be able to find out by searching for appnotes on the subject.
Tim

I took a third look at the board of the module, and it seems there is absolutely no protection between the POTS inputs and any outputs. What's worse (if my above deduction is correct), the board is meant to connect directly into a PCI/PCIE/USB socket AND is used rather widely in security centres. On top of that, its a US$2000 piece of card.


In this case I doubt my circuits will need much protection since apparently the card (and its predecessors) has been used since the DOS days by my client without a single hitch/surge. However, since I am covering all my bases, how may I protect my uC circuit from the module? The only thing that they likely will have in common is the RS232 link from the module and the power supplies.

[thejoggingmat]

Tin plated steel is still common in canned goods and biscuit tins, hardware stores may have galvanized steel sheet and often both Al and Cu roof flashing sheet rolls, hobby stores have brass, copper sheet

and I would like to add my usual warning that most hobbyists have no business worrying about Mu Metal shielding - I designed Strain Gage amplifiers with V gains up to 8000, looked at the results with 16 bit converters - and never used Mu metal - just low carbon steel enclosures

I might try . Unfortunately I've rummaged around my house and the only tincan that attracts a magnet is a RS branded 5L PCB degreaser, barely used. Even my biscuit tin I have fails the magnet test.

I might try heading down to a warehousey-type hardware store soon and have a look at the Al/Cu sheets.

[T3sl4co1l]

Ah, that looks safe -- possibly.  See the path from input jacks, with capacitors, fuse, filters, uh, rectifier bridge I think, ferrite beads, transistor with some heatsink copper, and wide SOIC chip?

I'm guessing that chip is an integrated POTS-to-digital converter and isolator.  So that's where your isolation comes from, which is safe.

Ed: in fact if I'm reading that right, they are indeed http://www.ixysic.com/home/pdfs.nsf/www/CPC5622.pdf/$file/CPC5622.pdf
Tim

[dom0]

And near the fuses there are two devices marked SPn => surge protector?

[helius]

In addition to foil you might consider painting or spraying the acrylic with a conductive paint. The key phrase here is "print", as in nickel print or silver-coated copper print.

[thejoggingmat]

Ah, that looks safe -- possibly.  See the path from input jacks, with capacitors, fuse, filters, uh, rectifier bridge I think, ferrite beads, transistor with some heatsink copper, and wide SOIC chip?

I'm guessing that chip is an integrated POTS-to-digital converter and isolator.  So that's where your isolation comes from, which is safe.

Ed: in fact if I'm reading that right, they are indeed http://www.ixysic.com/home/pdfs.nsf/www/CPC5622.pdf/$file/CPC5622.pdf
Tim

Well, I sure hope I am safe then.
Much more concerned about the welfare of the card than the welfare of my circuit 
I always assumed the long array of filters & caps were only to filter/protect the power/gnd line, leaving the data/voice lines exposed. Used to see POTS phones being damaged due to surges on the voice lines rather than from power.

And near the fuses there are two devices marked SPn => surge protector?

I don't really know... The last time I had a chance to take a quick look at the card, I remembered it being unmarked.
Can't really fool/play/experiment around too much as it is a $2000 piece of hardware that I can't afford, and currently in use..

[thejoggingmat]

In addition to foil you might consider painting or spraying the acrylic with a conductive paint. The key phrase here is "print", as in nickel print or silver-coated copper print.

That could be possible, but how do I conduct that to ground?
I can't ensure good conductivity with nuts/bolts without risking acrylic cracking, neither can I solder anything on it. Or maybe I could just leave it floating?

[helius]

That could be possible, but how do I conduct that to ground?
I can't ensure good conductivity with nuts/bolts without risking acrylic cracking, neither can I solder anything on it. Or maybe I could just leave it floating?

You're correct, the most common (pun) way to do it is using spring clips. Or just bend a strip of copper sheet into a D shape. Also "EMI gaskets" with fingers are sometimes used when equipment emits at high frequencies.
If you use a floating shield it will "shadow" RF energy around the device but if there are holes they can be a path for ingress. When the shield is grounded it acts more like an antenna to divert RF energy away to ground and holes are less an issue.

[klunkerbus]

For my attempts at shielding, I usually use K&S Engineering tin sheet.  Inexpensive, thin enough to cut easily, and solders very easy.  Many hobby stores and specialty hardware stores around here have display racks from K&S.  I'll sometimes cut a piece to fit inside the bottom and top of a plastic case, or form a shield that gets soldered to the circuit board.  The last piece I bought was marked as K&S 5254, with dimensions of 0.008 in x 4 in x 8 in. 

[Ian.M]

Connecting conductive paint shields is fairly easy if you have copper foil tape with a conductive adhesive.  Simply stick a square of tape to the dried shield and solder to that. You'll also see shield paint over a  soldered foil connection, but that relies on the paint film not cracking at the foil edges so is only done by cheapskates who cant afford conductive adhesive foil.

Another extremely common method of connecting shielding is with a small brass or copper plated compression coil spring directly soldered to the PCB.   However its a coil so is poor at high frequencies, so leaf spring fingers or a block of metal cloth over foam EMI gasket retained by conductive tape is a better choice.

[Dape]

From my experience (when I was doing EMI and RFI shielding), simply shoot a mail, question to some guys who are doing it for living. We are potential clients, so they will always do their best to help us. Simple as that. These are one of them, they are EMI/RFI shields manufacturers, so find out from them exactly what are you looking for.
This is the link:
http://masach.partcommunity.com/

[ptricks]

Since you are building only 5 and want shielding that can be bought locally, have you considered using electrical boxes used for AC work ? They are heavy gauge metal, come in all sorts of sizes and shapes with removable covers and fittings and can be bought everywhere for just a few dollars.
Square 2 gang metal boxes and the screw on cover are one of my favorite containers for projects. They are also very durable for portable items and can be mounted as well for installed projects .  If you need to close off the tiny holes in the boxes you can solder them closed.

[ChunkyPastaSauce]

Can aluminium sheets/foil be used for RF/EMI shielding?

Copper/steel etc sheets are out as I cannot purchase them in small quantities. Only aluminium sheets or foil is available.

If aluminium cannot be used, what other possible alternatives can I use that can be found commonly?

Aluminum foil definitely works in many cases except for the low frequency stuff. But one thing to watch out for is galvanic corrosion or anything alkali. Also when making mechanical contacts, you typical want a sliding contact, rather than just simply compressing the material; the sliding removes oxides from the surface so you get a good contact. Also I am guessing you have access to PCB panels.... those can be used for some shielding as well.

Also for steel shields, since you are have trouble sourcing, you might want check out a HVAC distributor. Steel sheets used for duct work. Maybe not as good as true mu metal, but way way better than alum/cop for low freq stuff

[Stray Electron]

RF shielding yes, EMI, not really so much. Typically you want Mu metal for that, or steel at the very least.

EMI ~ electromagnetic fields ~ RF. Mu metal is the only really effective way of shielding magnetic fields (noted lack of "electro-"). Good conductors like aluminium and copper are very effective against electric and electromagnetic fields (hence most RF assembly enclosures are die-cast aluminium), but barely make a dent in a magnetic field.

Most instrument enclosures are made from aluminium sheet and die-casts since quite some time, it has been popular since at least the 50s.

I think I do have some ancient PC cases lying around; if I could fish one out that has a thin sheetmetal body would it be suitable?

   I didn't read the entire thread so maybe already pointed this out; but you can find Mu metal in old CRT type TV sets. Mu metal was used in the shields around the CRT and a large TV has quite large sheets of it.  Other devices with CRTs probably also have Mu metal in them but I'm not 100% certain of that.

[T3sl4co1l]

1930s electrostatic sets, perhaps, but I've only ever seen steel in anything 50s onward.  Even the top luxury sets (e.g., RCA CT-100, the first or second commercial color set) didn't need the stuff.

I've seen shielding metal around the deflection yoke, but that's probably all steel too (or ferrite, for the ones where the core is a band over both saddle-shaped coils, rather than winding the vertical coil around the core halves).

Scopes used it fairly frequently, if in part just to say they could.  Tektronix's strict geometry specs probably needed it, but most didn't need or care.  Anyway, a modest thickness of steel (good clean soft mild steel, AISI 1020 or lower carbon, preferably annealed) does just as well, and is cheaper; the only reason you'd care for mu metal is because of size or weight (less material for same performance) or improved performance (for same thickness).

Tim