Protecting high impedance sections of PCB with metallic Sheets rather than metallic molds
in [Design]
|
From: whit3rd on 19 Apr 2010 06:00 On Apr 18, 11:02 pm, PureSine <Green.Tech.Co...(a)gmail.com> wrote: , > In many sensitive devices/ instruments which have a sensitive high > impedance section like a high impedance amplifier(10^12 Ohm or even > higher) or other high impedance signals, I've seen they are protected by > a metallic mold(mostly copper) A bit of foil or sheet metal would work, of course, but it can also vibrate; that would make the stray capacitance into a microphone. The sensitive parts of my old 10**-18 A electrometer are in a cast iron housing. Vibration control is why.
From: Bill Sloman on 19 Apr 2010 06:21 On Apr 19, 8:02 am, PureSine <Green.Tech.Co...(a)gmail.com> wrote: > Hello, > In many sensitive devices/ instruments which have a sensitive high > impedance section like a high impedance amplifier(10^12 Ohm or even > higher) or other high impedance signals, I've seen they are protected by > a metallic mold(mostly copper) that sits on top of sensitive parts and > tracks and is usually screwed to the PCB. Such metallic molds which are > specific to the shape of PCB seems very expensive. I wonder those of you > that are familiar with such protection do you think it is necessary ? > They are for protection against EMI but for instruments that have no > signal/Clock faster than a few hundreds of MHz, the EMI wavelength > should be fairly smaller than 30 Cm, So effective it would be to instead > of designing a board specific metallic mold for each high impedance > section, Just cutting copper or aluminum sheets approximately to the > size of sensitive area and then mount them on PCB using 5mm metallic > Spacers that are grounded. Well there is a 5mm gap but this method is > much cheaper and simpler and to the extent of theory that I know it > should stop all the EMIs that their wave length is higher than a few Cm > and it is translated to about 10 GHz. > Even harmonics of a few hundred MHz clock are very weak at such frequency.. > > What do you think ? Search on slot antennas. Narrow gaps in shielding can form slot antennas at specific wavelengths which can provide better coupling to the sensitive elements than you'd get if there wasn't any shielding at all. I've not seen this happen but the risk is mentioned in most discussions of shield design. -- Bill Sloman, Nijmegen
From: Jan Panteltje on 20 Apr 2010 05:51 On a sunny day (Mon, 19 Apr 2010 16:11:36 -0700) it happened John Larkin <jjlarkin(a)highNOTlandTHIStechnologyPART.com> wrote in <udops55kcmk8clmifa8c9dnh2bbep838tv(a)4ax.com>: >We make irregularly shaped covers out of sheet metal, bent into sorta >boxes, sometimes with mounting flanges that can be bolted hard against >the pc board. In colors! > >http://www.highlandtechnology.com/DSS/V470DS.html >http://www.highlandtechnology.com/DSS/V450DS.html DOOO theeey alzo come wizz nozee piccsss?
From: MooseFET on 19 Apr 2010 10:05 On Apr 18, 11:02 pm, PureSine <Green.Tech.Co...(a)gmail.com> wrote: > Hello, > In many sensitive devices/ instruments which have a sensitive high > impedance section like a high impedance amplifier(10^12 Ohm or even > higher) or other high impedance signals, I've seen they are protected by > a metallic mold(mostly copper) that sits on top of sensitive parts and > tracks and is usually screwed to the PCB. Such metallic molds which are > specific to the shape of PCB seems very expensive. I wonder those of you > that are familiar with such protection do you think it is necessary ? > They are for protection against EMI but for instruments that have no > signal/Clock faster than a few hundreds of MHz, the EMI wavelength > should be fairly smaller than 30 Cm, So effective it would be to instead > of designing a board specific metallic mold for each high impedance > section, Just cutting copper or aluminum sheets approximately to the > size of sensitive area and then mount them on PCB using 5mm metallic > Spacers that are grounded. Well there is a 5mm gap but this method is > much cheaper and simpler and to the extent of theory that I know it > should stop all the EMIs that their wave length is higher than a few Cm > and it is translated to about 10 GHz. > Even harmonics of a few hundred MHz clock are very weak at such frequency.. > > What do you think ? I have used the sheet metal plate over the sensitive stuff method. Here is what I did: 1 The shield needs to connect to the "shield ground" trace that runs around the electronics in question. 2 The connection points need to be outside the area of the the sensitive stuff. 3 The sheet metal needs to extend out so that it nearly reaches the metal chassis. 4 All connections must be good connections or not at all. Poor connections can work as radio detectors.
From: mickgeyver on 19 Apr 2010 16:22
On Apr 19, 2:02 am, PureSine <Green.Tech.Co...(a)gmail.com> wrote: > Hello, > In many sensitive devices/ instruments which have a sensitive high > impedance section like a high impedance amplifier(10^12 Ohm or even > higher) or other high impedance signals, I've seen they are protected by > a metallic mold(mostly copper) that sits on top of sensitive parts and > tracks and is usually screwed to the PCB. Such metallic molds which are > specific to the shape of PCB seems very expensive. I wonder those of you > that are familiar with such protection do you think it is necessary ? > They are for protection against EMI but for instruments that have no > signal/Clock faster than a few hundreds of MHz, the EMI wavelength > should be fairly smaller than 30 Cm, So effective it would be to instead > of designing a board specific metallic mold for each high impedance > section, Just cutting copper or aluminum sheets approximately to the > size of sensitive area and then mount them on PCB using 5mm metallic > Spacers that are grounded. Well there is a 5mm gap but this method is > much cheaper and simpler and to the extent of theory that I know it > should stop all the EMIs that their wave length is higher than a few Cm > and it is translated to about 10 GHz. > Even harmonics of a few hundred MHz clock are very weak at such frequency.. > > What do you think ? > > Regards. Why not just make them out of a few scrap pieces of PCB soldered together? Shear narrow strips for the sides, cut a piece for the top and solder the seams. al |