0
\$\begingroup\$

What is more important for EMI shielding effectiveness with respect to guard traces and via fences, the width of the copper trace that circles around the area or the distance between the vias to ground that connects to the trace and the other ground planes? Are both equally important or is one important and the other not so much?

Any good web resources out there for this?

Some background threads: the thick copper trace

What is the purpose of holes on edge of the PCB?

TYIA

\$\endgroup\$
2
\$\begingroup\$

The short answer is that neither matter in most cases. The only good reason I have for doing such a "fence trace" is to prepare for maybe adding a shielding can later (after testing if it works okay without). The trace in itself does not help much - other than adding some distance (which you could do without the trace).

The longer answer requires you to first explain what you are trying to achieve, why you think this will work and how you think it will work.

In general: If you don't understand what you are doing - maybe don't do it :-) A lot of time is spent copying stuff we "think" is good and often creating more problems than we had in the first place.

\$\endgroup\$
  • 1
    \$\begingroup\$ was there an answer in there other than to say that if I don't understand it, don't do it? \$\endgroup\$ – broncoremy Aug 2 '15 at 17:17
  • \$\begingroup\$ I am not copying anything. My question was more theoretical than specific to any one PCB design. Many applications require a guard ring around the perimeter at the edge of the pcb for both EMI and safety. In that case, the good reason to put it on the PCB is because your customer demands it in their requirements. Normally one would put such a ring with a via fence to cover a spectrum of frequencies up to a maximum signal frequency you would see on the board. \$\endgroup\$ – broncoremy Aug 2 '15 at 17:27
  • \$\begingroup\$ A rule of thumb I have used over many years is the 1/10th or 1/20th of the wavelength rule for the amount of distance between the via pickets to contain both the maximum fundamental freq and the harmonics of consequence that come along with it. \$\endgroup\$ – broncoremy Aug 2 '15 at 17:29
  • \$\begingroup\$ I have recently come across some designs that have a very thick trace, a few with thicknesses over 400 mil on internal layers of the PCB. To me this extra trace thickness seems to have little to no purpose given that the via fence connected between the ground planes does most of the errant EM signal blockage. I tried to ask a general question so as not to give away my bias. \$\endgroup\$ – broncoremy Aug 2 '15 at 17:35
  • \$\begingroup\$ The answer in there is this - sorry if that was not clear: The trace does not help much and the vias do not help much. The can does the job - if you need it. The last comment was really just to instill some bravery in you :) Just because someone else does it does not mean it's good or works or is worth copying. If you can't see why it should work, well then maybe it doesn't :) There is always a good chance you are smarter than the guy who deigned the board you are looking at - especially if you do your homework. \$\endgroup\$ – Rolf Ostergaard Aug 3 '15 at 18:24
0
\$\begingroup\$

The thick trace is as the post said used for EMI or RFI shielding. They are also needed in RF boards to place an RF shield. http://media.digikey.com/Photos/Laird%20Tech%20Photos/BMI-S-202-C.jpg

The above is an image of an RF shield. Now, RF PCBS have very high frequency in both the transmitter and receiver section. These fast switching signals are a designers nightmare as they can produce noise on nearby tracks. Suppose we have long tracks near to RF antennas, the signals on the antenna will produce magnetic effects whose influence is felt on neighbouring traces and tracks. This will corrupt the data being sent on those tracks and thereby effect system performance. Hence forth most RFIC manufacturers recommend a 3R rule. It means other tracks should be at a distance of 3R from the antenna. R is the width of the antenna track. Also it is recommended to have such ground tracks on the periphery of the RF boards. These thick tracks are connected to ground planes. So any EMI or RFI generated is quickly grounded, thereby nullifying its effects.

\$\endgroup\$
  • \$\begingroup\$ Thank you for your response. As the signal edge rates have dropped I have moved away from 3R to more like 5 or 6 for proper separation distance. My latest design has separation of 5 for these high edge rate signals. I do still use 3 for other, more general signals, so I certainly agree with you there. \$\endgroup\$ – broncoremy Aug 2 '15 at 17:42
  • \$\begingroup\$ If u like it then kindly vote for my comment \$\endgroup\$ – Board-Man Aug 2 '15 at 17:45
  • \$\begingroup\$ But does the protection come from the thickness of the track which is planar to the PCB? I would contend that this thickness does almost nothing. The protection comes from the metal can if you on the surface of the PCB or the via fence if you are looking at internal layers, do you agree? \$\endgroup\$ – broncoremy Aug 2 '15 at 17:45
0
\$\begingroup\$

Yes. The track thickness is proportional to the size of the rf shield which will be soldered to your pcb.

\$\endgroup\$
  • \$\begingroup\$ agreed, but that is ONLY for mounting of the can to the board on external layers. The walls of the can does the work. On internal layers there is no metal can to mount - you have the via fence, so from my understanding, no reason to have a really thick trace. \$\endgroup\$ – broncoremy Aug 2 '15 at 17:51

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.