Page 2 of 2

Re: farvorite ways to start a pcb design?

Posted: 24 Jan 2011, 10:40
by lolbou
RnFR wrote:is that bad practice?
This I don't really know (no expert either! :wink: ), but I'd rather avoid it. It might be more important in higher gain and higher power stuff (like amps)...
RnFR wrote:the ground is running everywhere the pour is, it's not like it's running down wires. :?:
It is to me... Look at the righthand side 10k resistor shown on your PCB above: it's connected to ground. Now look at the 1M one, lefthand side. It's grounded too. But there are actually two path to go from one resistor to the other (on top of your ground loop, or on the bottom). Not really star-groundesque, right? Opening it somehow would solve this. I remember Francisco Pena from tonepad doing so...

Re: farvorite ways to start a pcb design?

Posted: 24 Jan 2011, 11:33
by Greg
lolbou wrote:
RnFR wrote:is that bad practice?
This I don't really know (no expert either! :wink: ), but I'd rather avoid it. It might be more important in higher gain and higher power stuff (like amps)...
RnFR wrote:the ground is running everywhere the pour is, it's not like it's running down wires. :?:
It is to me... Look at the righthand side 10k resistor shown on your PCB above: it's connected to ground. Now look at the 1M one, lefthand side. It's grounded too. But there are actually two path to go from one resistor to the other (on top of your ground loop, or on the bottom). Not really star-groundesque, right? Opening it somehow would solve this. I remember Francisco Pena from tonepad doing so...
I think the danger of grounding problems and ground loops in effects is sometimes overstated.. I certainly tie component ground points to the ground plane all the time.

For there to be a ground loop issue there must be a difference in potential between the grounding points.
I have to question whether there'd be any significant potential difference on the ground plane of a single and relatively small pcb..

As far as Star Grounding goes, I've always followed the advice of R.G.... this :
"In PCB based effects, you generally don't have to resort to star grounding on the effect (as I mention in my book on PCB layout)."
and this:
"... stompboxes are so small and have such small currents that
(a) the area of a ground loop is too small to pick up much radiation and
(b) the currents are so small compared to the conductivity of the grounds that you have little trouble."


I hope R.G. doesn't mind me quoting him on this.. or taking one comment out of a more complete explanation.
If he does, maybe he could leap in and explain it all better.
:thumbsup

Re: farvorite ways to start a pcb design?

Posted: 24 Jan 2011, 17:10
by Dirk_Hendrik
ppluis0 wrote:I agree with Dirk about the fact to keep signal tracks as short and direct as possible.
Eeeh, I did not exactly say that ;)
Keep high impedance lines as short as possble. Like the 2 resistors going to the inverting pin of an inverting opamp aplifier. When a high impedance pin is connnected to a low impedance source they can be longer if that makes the PCB layout easier.

Re: farvorite ways to start a pcb design?

Posted: 24 Jan 2011, 17:55
by snmavronis
Doesn't in theory making PCB traces wider (when possible) lower the impendance as well? I'm not even sure if it matters that much on these low voltage stompbox circuits.

Re: farvorite ways to start a pcb design?

Posted: 25 Jan 2011, 00:55
by Duckman
A very interesting article below
Point 12 talk about how wide traces reduces inductance and 15 about how corners increases it. Point 17 is about component orientation.

http://www.analog.com/library/analogdia ... ounce.html

Re: farvorite ways to start a pcb design?

Posted: 25 Jan 2011, 01:22
by snmavronis
Ah, I always get impedance and inductance confused.

Re: farvorite ways to start a pcb design?

Posted: 25 Jan 2011, 04:41
by Duckman