When we design electronics, there are a few things we shouldn't skip. Often, CEOs and CTOs pressure us to design things quickly. But what is the cheapest option? Should we spend one or two extra days analyzing the board or should we run it a second time?
My goal today is to discuss some design analysis I perform during the layout of a DCDC converter. These are general considerations that can be applied to all designs, not just DCDC converters. EMC appears to be failing primarily due to a lack of current loops that were overlooked by the designer. In this post, I will explain the design considerations I make to maximize my chances of passing EMC on the first try.
First of all, what is the estimated amount of radiation released by a looped current?
In any book on electromagnetic compatibility (EMC), we are going to find the formula relating the electric field strength of a loop antenna. :
E(x) = 13.2 * (10^-15) *(f^2) *A *I/x
A = area of the loop
I = current
x = distance from the loop
It is true that most books stop here, but most of them should go ahead of this discussion and explain how these concepts relate to the pass/fail in the real-world context.
Therefore, I use a spreadsheet with my limits (which I receive from the testing house), and then I estimate the electric field at a distance of 3 meters using the spreadsheet.
As a result, I end up with a graph that looks like the one I have just made in the drawings, and I am very pleased with it.
In most cases, estimating the current I is a difficult task; however, we have tools like LTSpice, spice 3f5, etc., and so we should not have any problem assessing the RMS value for the current using those tools.
In the process of performing this analysis, I sometimes discover that I am unable to pass the EMC test. Then what is the best thing for me to do?
We have a lot of options at our disposal. It is easiest to use capacitors in antiparallel, which is the simplest solution.
This is the kind of case wherein I create a second current loop that runs in a 180-degree phase with the first loop that will tend to eliminate the problem. In order to achieve this, it is easiest to place a capacitor "antiparallel" to the main capacitor, and this, in turn, will create a current loop that will tend to create a magnetic field in opposition to the first magnetic field that was created.
I hope you enjoyed it.
Bye now!
Comentarios