Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.
Actually, I also not exactly know this name, "shielding via". But the thing I can comes out is doing cascaded vias along with shielding can on the PCB to isolate
Top(component) layer copper noise transportation.
Oftenly, radiation source comes from metalic part of your product, so top layer noise running could be a RF radiation source.
There can also be a thought that as this is an abrupt change in the direction of current flow along the trace, that a via can give of or be susceptible to EMI.
Putting groundplane around a susceptible tracks via's can allegedly protect it slightly better than no shielding.
But then EMC control is all black magic & I just made the above up..... but that may be a reason for it
Shieding vias are used between two ground planes to create a vertical shield structure. It is sort of a fence that can SOMETIMES reduce corsstalk and/or EMI by blocking propagation or coupling of a signal trace located between the two planes. Because of the resemblance to a fence, you will sometimes see them referred to as "picket fence vias". They are sometimes also referred to as "stitching vias".
To visualize the structure, picture a signal layer in PCB between two ground plane layers. If I place a row of vias parallel to a signal trace, and connect the vias to the two ground planes, I have created a sort of fence with the vertical barrels of via copper alongside the signal trace. If the vias have the proper spacing, they may help block coupling into and out of the signal trace. Take a look at the paper at: http://www.sigcon.com/Pubs/news/2_13.htm .
There is a need to exercise caution when choosing to use shielding (fence) vias at very high frequencies. If the spacing is improperly chosen, you can create a resonant loop structure that will both distort the signal on the trace by absorbing energy, and re-radiate rather than shield. For that reason, any shielding done with vias should be modeled carefully with 3D modeling software. Improperly used, shielding (fence) vias can make crosstalk, and EMI, worse rather than better. Proper layout of the board, with close attention to signal return paths, usually makes consideration of fence vias unnecessary.
Appreciated your points. For your description,
"the spacing is improperly chosen, you can create a resonant loop structure that will both distort the signal on the trace by absorbing energy, and re-radiate rather than shield. For that reason, any shielding done with vias should be modeled carefully with 3D modeling software. Improperly used, shielding (fence) vias can make crosstalk, and EMI, worse rather than better. Proper layout of the board, with close attention to signal return paths, usually makes consideration of fence vias unnecessary"
Can you pls say more clearly how to choose the spacing rule and why the loop would absorb energy and re-radiate the energy, and paper or comments of your favor is appreciated.
I probably should have said "exciting" instead of "absorbing" in my explanation to make it clear. A resonant loop formed by improperly spaced shielding vias will be excited by a nearby field at resonant frequency. When the signal level changes on the nearby signal path, the energy of the resonant loop will be returned as its inductive fields collapse. Both the mechanism of excitation (which couples energy out of the signal path into the resonant shield loop), and the deterioration of the resonant loop fields when excitation ends (which couples the loop energy back to the signal trace) will modify the signal seen on the trace being shielded. Edges can be delayed, jitter can be introduced, etc.
House_cat - thanks for pointing that out, I've been arguing exactly those potential problems with customers for years now - often including test data and published theory, more often than not using shield lines and/or stitched vias at sub-microwave frequencies, will cause more problems than they solve - a while back, one design house in particular was even advocating plating the vertical edges of boards to create "a Faraday Cage", all you have to do is visualize an electromagnetic field, as you learned in physics to know that this is useless, quite often inadequate power distribution creates a ground plane that radiates the sum of the simultaneous switching energy. always use more copper than adequate for power distribution, with an equal amount for return.
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.
