[sdiy] Analog & Digital Ground

Seb Francis seb at burnit.co.uk
Fri Apr 21 02:09:37 CEST 2006


Colin Hinz wrote:
> On Wed, 12 Apr 2006, Seb Francis wrote:
>
>   
>> Well, I've just laid out the digital part of the PCB, and I've left
>> spaces for ferrite beads on both the digital power connection and
>> ground.  So when it's up and running I will get my scope out and do some
>> *real life testing* of the difference with and without the ground bead!
>>     
>
> Bear in mind that high-speed digital circuits will exhibit flakiness when
> confronted with an unstable ground. A big lump of inductance in the
> ground path may qualify as an 'unstable ground', depending on the switching
> speed of the ICs. How are you going to test for this, in the digital domain?
>
> Ferrite beads on grounds may be OK for specialized applications such as
> a separate ground pin for a PLL circuit within a big digital IC. But for
> powering all of your digital stuff.....my experience suggests that this
> will cause you grief.
>
> Remember, a microcontroller is not made of 4000-series metal-gate CMOS....
>
>   

Well there are quite a few people who will be building the Box-O-Trix, 
so they all have a choice whether to leave out "FB2" (the ferrite bead 
on the digital ground).

Personally I'm going to try it with it in first.  There is local 
decoupling for the PIC power so any short current spikes should be able 
to come from here rather than from the PSU.

If you don't want to try it .. just put in a link instead of FB2.  In 
any case it was necessary to have some component separating the analog 
and digital grounds otherwise my PCB routing software would have joined 
the digital ground trace to the analog ground plane - I definitely 
wanted to at least route the digital ground all the way back to the PSU 
before it connected to anything analog.

Seb




More information about the Synth-diy mailing list