[sdiy] bypassing caps

[Jim Patchell]

    You can never have too many bypass caps....

    If you want to see how I bypassed...never mind, you can't actually see, I
just looked at the photo and realized you would never be able to find the 4
ASM-1 type oscilators on the board, but here is the link anyway
:http://www.silcom.com/~patchell/smb1/smb.html

    What I try to do is have one bypass cap for every power supply pin on every
IC.  This can be a lot of bypass caps.  I just recently finished a layout for a
vocal filter, and there are 100 (exactly) bypass capacitors on that board.  On
the smb board above, I was not quite able to get the ideal of one cap per pin,
because of limited board space, but there are still about 150 bypass caps on
it.  One of the things I worried about when I did the smb board, with 4 VCO's
was unwanted soft syncing between the oscilators.  I am able to adjust the
oscilators to within 0.1 Hz without any apparent interaction.

    So, my general rule is I use a 0.1uF ceramic capacitor on every (or as many
as posible) power supply pin.  Except for CA3280's.  They have 2 V+ pins, the
part still only gets one bypass cap...

    Another important thing is grounding.  Bypassing won't help much if you
don't have good grounds.  Keeping the path for the bypass cap as sort as posible
is important as well.  Generally, you will have to caps per IC, one for V+ and
one for V-.  Both of these will go to ground.  Keeping the distance between
those two ground connections short is important.  Of course, you have to work
within the constraints of what is posible on a board layout.

    -Jim

Nils Pipenbrinck wrote:

> Hi again.
>
> Since I have to re-design my asm-1 vco pcb anyways I thought about improving
> it a little bit. I want to add more bypass caps. Right now I have two 10µF
> caps from the rails to ground directly beneath the power connector. I also
> routed the rails and the ground seperately for each "part" of the circuit
> and connected them at the power supply connector. I think that's the best
> way to make the supply lines as clean as possible.
>
> Now I think about adding small bypass caps at the power supplies of the
> opamps and especially at the LM311 voltage comparator. Maybe that's not
> nessesary, but at least it won't hurt, right? Is a 10nF ceramic a good
> choice here?
>
> Also I wonder about bypassing in general: I always bypass from rail to
> ground, but never from rail to rail. That's the way it is done in all
> circuits I've seen. Does a rail to rail bypass cap has no effect or is not
> important (or does it even hurt stability?)
>
> Many thanks,
>     Nils Pipenbrinck