> wrote: >>=20 >> There's no change since forever for this, it's been the same all = along >> - you have to manually short the two grounds as the last step in >> layout, and ignore the DRC errors for those grounds after that. >>=20 >>=20 >>=20 >>=20 >> --=20 >> "Creativity is intelligence having fun." =97 Albert Einstein >=20 --Apple-Mail=_AD149F33-C66F-4870-879F-6D74A6FF6C60 Content-Transfer-Encoding: quoted-printable Content-Type: text/html; charset=windows-1252 I do the following:

1. Create the schematic without the AGnd and DGnd nets = connected. Route the board. When finished, DRC it.

2. Change the schematic to connect the nets together. = Usually I have a separate section of the schematic that makes this = explicit. Change the PCB to reflect this change.

3. = DRC the layout to verify that there are no new errors.

Step 1 ensures that the = layout is complete and that there are no inadvertent connections between = AGnd and DGnd. Step 2 adds the connection, and step 3 ensures that = the addition is correct and no inadvertent connections were made or = lost. For simplicity, do steps 2 and 3 only after the board is = otherwise complete. If you choose to edit the design at a later = date, remove the AGnd - DGnd connection on both schematic and board, DRC = to ensure correctness, edit the schematic, edit the PCB to match, DRC = the edited PCB, then do steps 2 and 3 above again.

A little tedious, but it = works fine for me. One caution: Put a note in the schematic = and in the area outside the cutting marks on the PCB to explain where = the connection is located on the PCB. That makes it much easier to = remove it during future edits. Otherwise, you _will_ forget where = it is and that makes finding supposed DRC error a real pain!

Best regards,

Peter

Peter Gregson, Ph.D., = F.E.C., P.Eng.
Engineer in Residence
Professor, E&CE

=93A good engineer is = someone who can do for 50 cents what any idiot can do for a dollar!=94 = =97S. Campbell

On Mar 11, 2016, at 3:50 PM, James Battat (jbattat AT wellesley DOT edu) [via geda-help AT delorie DOT com] <geda-help AT delorie DOT com> wrote:

I=92m also = hesitant to use the =93do this at the last minute=94 scheme.

A through hole jumper is = certainly an option, but I was hoping for a solution that didn=92t = require a physical component (e.g. a =93virtual=94 component in gschem + = pcb that allows you to legally connect the AGND and DGND nets.

Open to suggestions from other folks as = well (I imagine this is a very common need =97 single-point ground in a = mixed signal board).

Thanks,
James

On = Mar 11, 2016, at 2:34 PM, Mike Bushroe (mbushroe AT gmail DOT com) = [via geda-help AT delorie DOT com] <geda-help AT delorie DOT com> wrote:

What about using a through hole = jumper to perform the same function as the zero ohm SMD when the planes = are on the surface? Assuming you have some freedom to place the cross = connect anywhere along their boarder you should be able to find two = spots for the through holes. I don't know if that would change the cost = of manufacturing an otherwise all SMD board with only layer to layer = vias, but I would think the process for making a top to bottom via would = also make plated through holes.

I just know that setting up something I just HAVE to = do on the very last run before submitting is a surefire recipe for = disaster for me. I would rather put up with seeing the short every time = I rand DRC than risk sending the files off to hastily.

Mike

On Fri, Mar 11, 2016 at 12:20 PM, Peter Clifton = (petercjclifton AT googlemail DOT com) [via geda-help AT delorie DOT com] <geda-help AT delorie DOT com> wrote:
(And be CAREFUL)... I've missed nets because = of this. Be sure to make the short at the last possible moment.

Peter

On 10 = March 2016 at 22:16, DJ Delorie <dj AT delorie DOT com> wrote:

There's no change since forever for this, it's been the same all = along
- you have to manually short the two grounds as the last step in
layout, and ignore the DRC errors for those grounds after that.

--
"Creativity is intelligence having fun." =97 Albert = Einstein

= --Apple-Mail=_AD149F33-C66F-4870-879F-6D74A6FF6C60--