Re: gEDA-user: basic anti-EMI design q

[Steve Meier <smeier@xxxxxxxxxxxxxxxxxxx>]

Hmmm, not likely... the two grounds should be very very close potential.
so we are expecting very little flow through that power inductor.
Normally I am trying to isolate my sensitive analog signals from the
digital signal processing. So it is the noise on the digital ground that
I am trying to keep off the analog ground. Sure if some one was to
suddenly charge a ground plane up wit a power supply you might get some
interesting effects but I think the inductor will be the least of your
problems.

Besides there is no capacitor other then the overlaping of the power and
ground planes. Paracitic capacitance so to speak.

Lets assume that some one put a strong electrical field into DJ's ribon
cable. Let us also assume that that ribbon cable is only connected to
one of the grounds. And that that ground gets pulled up with respect to
the other ground. Now we will have a DC current running through a power
inductor. But it won't instantaniously allow the maximum current needed
to bring the two grounds back to the same potental.

Realisticaly, the issue I deal with is that the digital electronics has
one or more clocks humming away. And I want to make sure that the noise
those clocks put on the imperfect ground plane is attenuated as small as
possible from the perspective of the a ton of high frequency analog
gain. Hence seperate ground planes.  Now when ones board has multiple
analog chanells and wishes to avoide cross talk between them one can
break up the ground planes further. Each time tieing them together
through an inductor. And if one starts seeing weird things you can take
the inducts off or short them. I can not stress enough that when
debugging complex boards the ability to pop an inductor off and turn off
the power or isolate one section from another can be very handy.

Steve Meier


Phil Taylor wrote:
>Steve Meier <smeier@xxxxxxxxxxxxxxxxxxx> wrote:
>
>  
>>but.... an inductor along with the capacitance between the planes 
>>    
>
>you didn't mention the cap ... and without that part, the coil connecting
>different grounds could just turn out disastrously! 
>
>phil
>
>
>
>