Just to elaborate on Garth's post above and his second paragraph
here ...
Because making PCB traces traces wider does not significantly reduce their inductance, even thin traces are acceptable for creating a
grid of ground traces (as an approximation of an actual ground plane). And thin traces mean such a grid needn't add significant area to the board.
Neither would the grid complicate the routing much. Regarding
signal wiring, it's generally true that north-south traces tend to go on one side of the board and east-west traces on the other. And the ground grid can "go with the flow" of whatever else is there, interleaving its north-south ground traces with north-south signal traces, and likewise for east-west.
At every point where a N-S ground trace crosses an E-W ground trace you'll want to connect the two (otherwise the grid doesn't approximate a plane -- current must be able to flow N-S, E-W or in a zigzag for diagonal travel). And if the N-S and E-W ground traces are on opposite sides of the board then you'll need to use a via to connect them. Vias do add a bit of inductance, but hardly enough to negate the benefit of the grid.
I'm not advocating grids as the answer to every situation. But they certainly bear consideration as one useful option on the continuum between outright rats-nests and rigorous, controlled-impedance designs. As a moderately interesting data point, it's worth noting that portions of my 1988
KK Computer use 74AC series logic running at 80 MHz. KK is implemented in wire-wrap, and there's no ground plane. But there
is a ground
grid, which in this case means
almost every IC has four connections (N, S, E and W) emanating from its Gnd pin to those of its neighbors. (Also, almost every IC's Gnd pin has a bypass cap running directy to the IC's Vcc pin, just as Garth suggested in the linked post).
-- Jeff
_________________
In 1988 my 65C02 got six new registers and 44 new full-speed instructions!
https://laughtonelectronics.com/Arcana/ ... mmary.html