Skin Effect Current and Fields in a PCB Explained

[@vatsan2483]

The Dan Beeker quote is actually legendary, the full quote is like this: "Building have walls and halls, people travel in the halls and not the walls, circuits have traces and spaces, energy and signals travel in spaces and not the traces!!" This is actually given by EM legend Ralph Morrison, author of bestselling EM books!! Mentor for many!!

[@dgyldmark]

What causes the electrone path in first place? How do they know witch way is the shortest? What a gold teacher you are. Thanks for your uploads! /Dan sweden

[@Bob-zg2zf]

outstanding! thank you, Dr. Peterson!

[@AltiumAcademy]

Glad it was helpful!

[@chromatec4311]

Thanks for the teaser - now please explain what are the implications and design choices required to minimise skin losses?

[@janstoeb]

I think you cannot really minimize losses overall, but you should account for it and might be able to safe some copper, if you are at midrange frequencies (frequencies before your transmission line is electrically short) by running thin parallel traces instead of one wide one. For copper the skindepth is 6.5 um at 100 MHz.

[@jimjjewett]

When you said "half" the energy is not from the directly facing side, was that just a very rough estimate (because the top is equally large, and the sides are small), or is there some sort of measurement about which portion of the energy travels/is concentrated on the portions closest to ground vs the more distant portions?

[@Zachariah-Peterson]

The statement "half the energy" is not to be take literally. As far as a measurement, I can imagine some measurement with a loop antenna (near field probe), which is basically measuring the magnetic flux passing through some area (the cross-section of the antenna), and that can be used to calculate the energy passing through that loop. From calculation, the method to get the intensity distribution in space would be to calculate the Poynting vector as a function of angle (in a cylindrical coordinate system, the trace runs along the axis). Then you could calculate the power flowing through some cross-sectional area using a surface integral. Or you could just plug in the angle around that trace and you would have your answer.

[@jimjjewett]

​@@Zachariah-Peterson OK, that makes sense until "plug in the angle" (except that I'm not sure how to do it -- but am aware that it is too complicated for quick explanation) ... I would expect the energy to be more dense in the places where it can go straight to a close ground plane than in the places where it has to curve around. I don't know whether it is (for example) 2/3 on the bottom and 1/3 on the sides and top, or whether the "skin" penetrates farther on the side towards the ground plane, and wondered if that were a solved problem.

[@melissareitzloff1637]

Because of currents valocity the trace format is porportional

[@melissareitzloff1637]

How does frequency and it's conductivity work in the brain as for mood transmission

[@87Spectr]

Thank you very much! Very useful

[@AltiumAcademy]

Glad it was helpful!

[@janstoeb]

I think two things are getting mixed up here. First off all, the conduction current density is not 0 in PEEC for all frequencies, since at DC it is distributed uniformly. Secondly, the electric field lines you are presenting here are the rotation free part of the electric field, which is produced by charges and not currents. Strictly speaking, the skin effect is a phenomena appearing first in the quasimagnetostatic case, so is purely magnetic and has nothing to do with charges, electricfield lines between the two conductors of a transmission line and capacitance. In my eyes this video really misses the topic/title. The video should have explained skin effect, proximity effect and skindepth based on an intuitive view of self and mutual induction inside of conductors. I am genuinley suprised by the content of video

[@Zachariah-Peterson]

Hi Half, thanks for watching. The point was not to derive the skin effect current or skin depth based on self/mutual induction. The point was to explain why the electric field is distributed around microstrip/stripline transmission lines even when skin effect is present as it is shown in some diagrams or simulation programs, which as I'm sure you know has everything to do with the charge/current distribution around the trace as well as the rate of change in the fields. Yes the E-field shown is the non-rotational part, but when a digital signal is traveling along the line, there will be a non-rotational component in the region where the fields are not changing and that will produce the field distribution shown here. If you like to analyze this in the frequency domain, then we would say that the signal's bandwidth is insufficient to excite any higher order modes, and thus the field is quasi-static and nearly-TEM, so we can analyze this in a 2D cross-section and get reasonably accurate results. This is the entire basis for 2D electromagnetic transmission line field solvers. Yes, we are skipping some steps as you have referenced but the picture of the electric field as shown in the video is correct and can be seen in any simulation program. There is sometimes a misconception that the fields and thus the skin effect only exists along that bottom surface between the trace and the plane, but that is not the case if you look at this kind of quasi-static situation and it should not be the case based on the current distribution.