Re: gEDA-user: Embedded Power Pins

[Karel Kulhavy <clock@xxxxxxxxxxxxx>]

On Sun, Mar 05, 2006 at 02:03:23AM +0100, kmk wrote:
> Stuart Brorson wrote:
> 
> > FWIW, I wrote a blurb about this question for the wiki a couple of
> > weeks ago.  Here it is: 
> > 
> > http://geda.seul.org/dokuwiki/doku.php?id=geda:faq-gschem#what_should_i_do_about_power_pins_on_my_symbolsmake_them_visible_explicit_or_invisible_implicit
> 
> Well, I agree only half way with you.
> It is true, that power pins in analog circuits should be visible on the
> schematic. But having them on attached to the triangular shape is not

What is analog and what is digital circuit? Is Ronja Nebulus analog or
digital circuit? http://ronja.twibright.com/schematics/nebulus.png

I hold an opinion that Ronja Nebulus shows there is no sharp border
between analog and digital and therefore there shouldn't be separation
to analog and digital regarding drawing powers in the schematic. I
suggest the one with explicit powers be used because I once needed to
power crystal oscillator over an RC filter and it made problems because
it shorted the RC over.

CL<
> correct either. Imagine a quad opamp: If you use symbols with visible
> power pins, you get eight visible power pins.
> 
> My preferred solution would be a special slot just for the power
> connections. This is how we dealt with this issue in Protel99 and it has
> its advantages for circuits that contain lots of opamps. The power
> connections plus the necessary caps don't clutter the feedback loops. So
> the circuit is easier to read. Still the power pins are explicitly drawn
> and can be tied to some exotic net. I usually draw all the power symbols
> with their caps in some quite corner of the paper.
> The only drawback is that the supply caps are not in the neighborhood of
> the corresponding triangular  shape. But with quadamps the triangular
> shapes that belong to one component are scattered over the schematic anyway.
> 
> Nevertheless, I don't see how I could implement this technique with
> gschem. There seems to be no way to let a gschem symbol contain slots
> with different pinout. The same problem arises if a component has slots
> that perform different functions. This is the case for some analog
> multipliers that contain an additional opamp with no internal connection
> to the multiplier. Did I miss something obvious?
> 
> ---<(kaimartin)>---
> -- 
> Kai-Martin Knaak
> kmk@xxxxxxxxxxxxxxx
> Blog: http://lilalaser.dyndns.org/blog