shield and emi filter

[coppercone2]

if you have a circuit board part with a shield over it, like a section, and its fed in through a shield, say a RC network, where should it be? If its just say a MHz filter on a op amp (something MHz). Do you want the filter right outside of the shield? inside of the shield? capacitor behind shield and resistor in front of shield?

how about power rails? Lets say you have a decoupling capacitor right near the op amp. Could you just feed the power under the shield, or should you have a little capacitor at the perimeter of the shield? Say 1/10th of the decoupling capacitor value maybe?

also, does anyone make through hole shield clips? they are all SMT. I can stake them with eyelets but its alot of work. it's not like its not gonna work.

[jonpaul]

Coppercone:  Rebonjour, Great question:

what power/freq range?  prod or proto?

In our FM tuner prod,   used ready made shield cans like these
https://www.harwin.com/products/S01-806005KIT/

Protos  were  made our own in the shop from copper sheet or tin can stock.

Power was via feedthru  RF bypass filters or caps like the old US mfg C-D, Centraab, Sprague. 


Signals were on the PCB or fed thru the shield via plain feedthru  terminals.

No issues at 100 MHz (FM band ) as all RF/IF was balanced.

Just my experience...1973

Your thoughts appreciated!

Have an Absolutely fantastic day,

Jon

PIX: Feedthru caps/filters
My 1973 Sequerra FM Tuner 100 MHz counter (TTL/MECL) with shield/filters/beads/feedthru

[coppercone2]

i mean if a network of chips was shielded on the pcb inside of a enclosure. no real feed throughs. and with a shield that is clipped on

[jonpaul]

Unsure your question.

any pix/schematic?

j

[Vovk_Z]

Do you want the filter right outside of the shield? inside of the shield? capacitor behind shield and resistor in front of shield?

I guess inside. You don't want high frequency signal to go out of the shield.

[MathWizard]

I put an amplifier or low MHz oscillator circuit in tuna can once, with the cover taped on, and I don't remember didn't noticing  any difference on my good scope. I know inside a perfect sphere of spread out charge on a conductor, the E field will sum up to zero. I can't remember if it exactly cancels in non-sphere's. And magnetics do something different and are way harder to shield.

So if you sealed up the can tight, even with solder, what would that be like ?

If you cut apart tin cans, into little rectangles, is there any reason you can't use that to make shielding cans ? If someone had the tools, they could cut/shape SMD cans too. If I knew where my tin-snips were, I might try it today.

[coppercone2]

I saw a massive increase in resolution on half a shield before on a ground plane pcb. Copper sheet

Jon Paul I mean if you had a pcb that has say a gain stage on it with an op amp. Then you just shield that chip and it’s resistors. Caps, etc. so traces go under the shield. Then that connects to other circuit stuff not under the shield

I decided not to do it right now because i don’t want to put rivets for the shield holders but in general I would like to know how to do it for a pcb

[MarkT]

I know inside a perfect sphere of spread out charge on a conductor, the E field will sum up to zero. I can't remember if it exactly cancels in non-sphere's.

If talking about solid conductors, the interior field is always at zero when no current is flowing.  And no the surface charge is not necessarily symmetrical on a sphere as it depends on the external electric fields.  The surface tangential field is always zero.

With hollow conductors the charge on the inside of a metal container also distributes itself so there's no tangential field at the interior surface.  The internal field depends on the internal charges, but always intersects the container at right angles.

You can only impose a particular distribution of charge with an insulator like an electret, as the charges are fixed in place in an insulator.

The charges/fields on the outside and on the inside are un-correlated with a conducting container (in electrostatics).  Skin depth is important in the real world though, so at low frequencies fields leak through (and there are currents within the conductor wall).

[PeteH]

Companies like tech etch can provide custom shields which have through hold or smt mounting options. Shields are often custom, so for COTS stuff SMT is probably the most common to achieve sufficient volumes to stock.

[coppercone2]

omfg pay someone to bend a sheet metal for me?

said while holding pressure on laceration

[coppercone2]

bought some EMI clips for experiments. first i will see how keystone eyelets do for fastening them to random PCB. it seems plausible to erect shields.. a nice combo to that bending machine I bought a while back for my vise.