Send comments to an expert

Board level shielding

You are here: Product design techniques > PCB layout > Board level shielding >

 

 

 

 

 

 

The on-board shield

Rather than shielding an entire product in a single enclosure, you can place a shield just around a particularly sensitive or noisy section of the circuit board, as suggested in the section on partitioning. The major function of such an on-board shield is to minimize capacitive coupling to other parts of the circuit, including interface circuits which couple directly to the external environment. If used in this way, it is important that it is connected to a point of low RF noise voltage as otherwise the greater surface area and hence larger capacitance can actually increase coupling to other parts. It will shield against radiated fields provided that it forms a good Faraday cage with the PCB ground plane, i.e. there is near-continuous contact across the joint and there are no appreciable apertures or seams in the shield.

Connection to the ground plane must in any case have the lowest possible inductance so that the increased capacitance from the shielded parts to the shield itself does not result in resonances with the grounding inductance. If these resonances are at too low a frequency (2nH and 12pF resonates at 1GHz, for instance) there is the danger that the shield will be an efficient, unintentional radiator. Therefore the question "how widely spaced should the ground contact points be?" should be re-phrased as "how low an inductance do I need to minimise the resonant effects?". The more points the better, and a continuous run of contact (which is normally quite possible with a surface-mounted screen to a continuous land strip on the board surface) is best, not to say essential for GHz shielding. This means that the PCB must be designed from the start with a region or regions designated for shielding and laid out with the suitable surface lands for this purpose.

Tracks which exit from the shielded area may have to be filtered or decoupled to prevent noise from bypassing the shield. This highlights the importance of partitioning the PCB so that there are the fewest interfaces across from the shielded to the un-shielded area, and so that those which do pass across can be filtered at the least cost and with the lowest impact on the circuit operation.

The shield can to a first order be represented by an equivalent circuit which includes the capacitance from the noise source to the shield, the capacitance from the shield to the victim – for instance with wireless on board, the receiving antenna – and the inductance of the shield connection to the ground reference. This neglects any resonant effects across the shield if it is not electrically small, i.e. substantially less than /4 at the relevant frequency. It also assumes that the ground reference between the source and the victim has negligible impedance, i.e. they share the same ground plane.

With this simple equivalent circuit, the transfer function from the noise source to the victim can be calculated, either estimating or using a field solver to compute the three coupling components. Doing this almost invariably shows that the most critical component is the inductance of the ground connection LG. This must be minimised by making a continuous bond around the shield perimeter to a surface land which has multiple vias into the ground plane. Anything less will compromise the shield's performance.

Construction

An on-board shield can be constructed from thin pressed or electroformed sheet metal such as tinplate, and soldered onto the ground plane or connected via finger strip. The actual metal that is used is not too important since, as a capacitive screen, it is the surface conductivity rather than the thickness which is important. Multiple compartments can be made in it; it can be fabricated either –

as a wall (or walls) which is surface mounted to the board, with a separate clip-on lid, or

as a single folded part which is pressed onto clips surface mounted onto the board, or

as a single folded part which is soldered with through pins or SM feet onto the board (this won't allow re-work underneath the screen!)

Several manufacturers offer standard sizes, and will also do custom sizes as necessary. Other techniques such as conductively coated plastic parts and form-in-place gaskets also are available, the choice depending more on mechanical, thermal and cost/volume consideration than on electrical performance.