MOTM

hello again..
I just downloaded the express PCB layout software a few days ago.  Tons of fun!!!  Am I a freak!?

Anyway, I was wondering if there are any general rules or theories that are followed when laying out a circut.  Or if any of you had any basic tips.  Like trace proximity to certain components, interference reduction, component density, certain shapes (90 degree bends?) or diameter of trace to avoid.  What signals should be kept far from eachother, where a ground plane is important or when it is required..  With the vias and double sided boards, getting most analog circuts (that Ive tried so far) laid out is not that hard, but I can sit there and refine and shape the layout for hours.  Is there be a "best" or certain goal I should be trying to reach? Or is this only the domain of computer circut (Spice?) simulation?

thanks for any tips, I'm hooked on this little program!

~Steve M

Well, certainly you have *many* fine MOTM example pcb layouts to look at :)

I don't want to give away too many secrets, but here some general rules:

1) use 15mil trace to connect signals.
2) use 50mil trace for power & ground.
3) use pad sizes of 80mil, with 40mil holes
4) don't 90 degree bend the 15mil traces. 50mil is OK.
5) use 50mil pads with 28mil holes for vias.

Have fun! I find it relaxing (for easy stuff! The next 2 layouts are nightmares. For Maxim, I'm
starting a 6 layer, 10 x 10 board at 100Mhz speeds. 5ns SRAMS and a 248-pin Xilinx part)

Paul S.

> 
> Have fun! I find it relaxing (for easy stuff! The next 2 layouts are nightmares.
> For Maxim, I'm
> starting a 6 layer, 10 x 10 board at 100Mhz speeds. 5ns SRAMS and a 248-pin Xilinx
> part)

Not that I would want to do any of this myself, but one of the folks I worked with
(whom I found out worked with Paul eons ago, stealing all of their computer sim time
for antenna simulations) completed a 44 layer backplane.  The board is .3" thick and
19" by 30" or so.  The connectors aren't even soldered in, they are just "press fit"
but with ~1K pins each, they have enough surface to hold them in place just fine. 
Needless to say that the cost for a single board could buy a whole lotta MOTM!  Of
course, the cards with the 1500 pin ASICs that plug into the backplane are 22 layers
thick on their own!  Stuff you just don't want to do at home.  [Dear, could you
crank up the wave solder line and the electron microscope so I can put down a few
chips tonight?]  Oh, and if you have to fix up some of the surface mount resistors
and capacitors by hand, I've been told it is like "soldering fly shit".  OK, now you
might understand why I'm in the "assembled module minority"....

    Eric



__________________________________________________
Do You Yahoo!?
Make a great connection at Yahoo! Personals.
http://personals.yahoo.com

>
> Not that I would want to do any of this myself, but one of the folks I worked with
> (whom I found out worked with Paul eons ago, stealing all of their computer sim time
> for antenna simulations) completed a 44 layer backplane.  The board is .3" thick and
> 19" by 30" or so.  The connectors aren't even soldered in, they are just "press fit"
> but with ~1K pins each, they have enough surface to hold them in place just fine.

Did one of these for Fujitsu Telecom about 6 years ago. Those press-fit connectors were a BITCH!
All metric and such. Mine was an Enternet-based backplane, only 18 layers, about 9 x 24 inches. I
charged them $1200ea for the blank pc boards.

The nastiest pcb I ever did was when my pcb tools were in DOS on a 486. I did a 6 layer PC AT
card for a SDLC network. I was running 3 traces in-between IC pads. I worked like 6 18-hours days
in a row, over Thanksgiving.

Paul S.