[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]

Re: gEDA-user: pcb question : same structure several times on board



On Mon, May 09, 2005 at 04:48:23PM +0200, Carlos Nieves ?nega wrote:
> El lun, 09-05-2005 a las 13:40 +0200, Leva escribi?:
> > On Mon, 09 May 2005 13:21:50 +0200
> > Bernhard Kraemer <Bernhard.Kramer@xxxxxxxxxxxxxxxxx> wrote:
> > 
> > > Hello,
> > > 
> > > I am creating a pcb board on which I have four times the same structure 
> > > (four voltage stabilisators together with their resistors and 
> > > capacitors). It would be very fine if I could copy one structure in a 
> > > way that all the three other structures, together with their wiring, 
> > > will look equal and straight. If a tool like this exists, I'd like to 
> > > use it for another circuit where a couple of ever-the-same filters 
> > > appear. Does a tool like this exist ?
> > 
> > Bonjour,
> > 
> > Well.... copy to the buffer, and then past buffer to board. This will mess up
> > the refdes'.
> > 
> > I was trying to create such things, but if you make very crowded boards, your
> > power supply will always be different too. 
> 
> Or not... it depends on the design. However, you can copy the layout
> without the traces or nets you don't want to be copied, and then paste
> the structure of those nets.
> 
> I guess this is not yet implemented, but how about using a hierarchical
> blocks for that?. I mean:
> 	- Draw the basic structure in a single schematic.
> 	- Draw a new schematic with as many instances of the structure  as you
> need (each as a block).
> 	- Now go to pcb, and do the layout of one structure.
> 	- Copy it to a buffer.
> 	- A new operation "Paste instance from buffer" is defined, which takes
> the structure from the buffer, finds another instance of the block
> (based on schematic blocks and subcircuits), and changes the refdes' to
> the new one. It can also ask the user which one should be used.
> 
> Something like this is implemented in high-level programs, and it's very
> useful. The problem is that there should be a tight integration between
> gschem and pcb, or at least pcb should know the block hierarchy of a
> component/block within the whole design....
 
This is something I'd really like to see in both gschem/gnetlist and in
pcb.

In gschem, I think all the needed stuff is there.

gnetlist really could use some better hierarchy support.  As near as I can
tell, gnetlist really only supports putting out a flattened netlist.  I'd
love to see it learn to preserve hierarchy in the output.

For pcb, here's a big reason I'd like to see hierarchy support.  Suppose
you have a something like a combline filter.  The layout is now truely
a critical part of the circuit.  It would be nice to lay that out as a
subcircuit and be able to move the filter as a block around.  

-Dan

--