LS,
Is there a thumbs rule connecting two layers.
On the bottom and on the top layer I have a GND polygon.
To reduce stray current I would like to connect these two using via's.
Regards
Harry
LS,
Is there a thumbs rule connecting two layers.
On the bottom and on the top layer I have a GND polygon.
To reduce stray current I would like to connect these two using via's.
Regards
Harry
The plane should carry a current of approx. 3A at 18V DC
"Gary Gofstein" <nospam@use.forum.net> schreef in bericht
news:hl1u5s$8mk$1@cheetah.cadsoft.de...
No. You would need to know the distribution of the current for DC. For AC
currents in the RF range you need to know the shape of the pulse and solve
the wave equation to know where the current will flow and what skin depth
it will flow in. Nobody does that.
For "normal circuits", it's seldom an issue. How much current are we
talking about here anyway? DC, audio or RF current?
To get a comparison for DC currents you can look at trace temp rise vs
size charts for ordinary traces. You will see that unless you use super
thin "digital" traces, it's usually okay. I always use the biggest traces
I can fit anyway. There's a ULP that calculates temperature rise on traces
for DC currents; you'll see that your plane is likely to be much bigger
than necessary. You don't need to know how many vias to add, because you
should just add as many as you can to get the best performance. You can
estimate the temperature rise of a via by thinking of it as a wire,
figure out the equivalent wire guage, consult a current vs temp rise site
or chart to get the temp rise. You should specify your vias to be solder
filled for best DC and thermal conductance. If you will never solder to
them, you don't need thermals on vias (EAGLE default anyway). Actually,
it's possible to tack solder a wire even on a via with no thermal, so it's
rare to need a thermal on a via. All this applies only to DC currents,
RF/pulse currents are a whole nother story...
The plane should carry a current of approx. 3A at 18V DC
"Gary Gofstein" <nospam@use.forum.net> schreef in bericht
news:hl1u5s$8mk$1@cheetah.cadsoft.de...
No. You would need to know the distribution of the current for DC. For AC
currents in the RF range you need to know the shape of the pulse and solve
the wave equation to know where the current will flow and what skin depth
it will flow in. Nobody does that.
For "normal circuits", it's seldom an issue. How much current are we
talking about here anyway? DC, audio or RF current?
To get a comparison for DC currents you can look at trace temp rise vs
size charts for ordinary traces. You will see that unless you use super
thin "digital" traces, it's usually okay. I always use the biggest traces
I can fit anyway. There's a ULP that calculates temperature rise on traces
for DC currents; you'll see that your plane is likely to be much bigger
than necessary. You don't need to know how many vias to add, because you
should just add as many as you can to get the best performance. You can
estimate the temperature rise of a via by thinking of it as a wire,
figure out the equivalent wire guage, consult a current vs temp rise site
or chart to get the temp rise. You should specify your vias to be solder
filled for best DC and thermal conductance. If you will never solder to
them, you don't need thermals on vias (EAGLE default anyway). Actually,
it's possible to tack solder a wire even on a via with no thermal, so it's
rare to need a thermal on a via. All this applies only to DC currents,
RF/pulse currents are a whole nother story...