Remember, this thread is about ESD! What, if the enclosure 2 is hit by an ESD event?? Then, the enclosure will be thousands volts higher than the internal electronics!! Only the cap between signal ground and enclosure, or the direct connection of course, can then force the enclosure on the same potential as the internal electronics. Only then the enclosure can act as a shield.

What you propose is in total contradiction to common design practice. Open any device you want and you will always see a direct or cap connection from signal ground to enclosure. I have seen audio mixers, where even at each pot looking through the enclosure such a cap was added. But what you propose will miserably fail every CE test.

Again, I'm NOT talking about any exotic construction, but about what point 7 of mentioned FAQ suggests, namely:



This statement just means, that as standard design practice signal ground and enclosure ground should be connected via a ferrite. And that's just wrong! Show me any personal computer where this rule is followed. You will not find any, just because it's totally stupid to destroy the shielding ability of an enclosure by inserting a ferrite! This COULD make sense (if at all!), when someone has made a certain shielding mistake. Then, adding this damned ferrite can give him the illusion to have cured the mistake. But what he actually does is trying to cure a mistake by making another, which will not work at all in general.

A shield must always be at the same potential as the internal electronics, at least for high frequencies, where the stray capacitance from enclosure to signal traces begin to become harmful and can allow to inject unsane interefence direcly into the circuit. Only when the enclosure is on the same potential as the internal electronics, concretely spoken signal ground, then this stray capacitance cannot inject noise!!! So, it's extremenly important to always have the enclosure on the same potential as internal electronics, at least for RF!!!!!!!

What you propose, mutates the enclosure from a shield to an interfering antenna! You just couple the noise directly into the circuit, instead of shielding the circuit by the enclosure!

There are many other reasons, why the enclosure must be connected to signal ground for RF. I will add here one more:

Each circuit containing HF emits electrical fields, which will interact with the enclosure. You can quantify this in terms of stray capacitance, means a stray capacitance is formed between the circuit and the enclosure. As consequence a current will flow in the enclosure (dielectric displacement current of course), which makes the enclosure noisy. Now a return current from enclosure back to circuit will flow, which can cause extreme interefence when not being returned on shortest path back to the origin of electrical fields, means the circuit.

This can cause cables leaving or entering the device to radiate so much, that the device will fail the CE radiation test. Not even mentioning the magnetic interference resulting from the huge increase of loop areas.

The consequence is, that in HF circuits there should not even only one connection from enclosure to signal ground but even many! Have a look how the mainboard of a personal computer is mounted to the enclosure wall...


Kai