I designed a PCB. What''s wrong with it?
Archiver
Posts: 46,084
--- In basicstamps@yahoogroups.com, "jmccorm" <jmccorm@y...> wrote:
> Is this realistic? Is this something I could give to an
> experienced person to produce? Are the holes for the LEDs and the
> wires too close to work with? Too many resistors too close together?
> Any obvious mistakes?
"Is this realistic?"
Absolutely.
"Is this something I could give to an experienced person to produce?"
Most definitely. Anything single or double-sided you can easily etch
and drill yourself in small runs.
"Are the holes for the LEDs and the wires too close to work with?"
Certainly not. Those willing to pay the premium are fabricating with
.002" lines and .002" spaces!
"Too many resistors too close together?"
Not at the currents you're working with. Re-calc the wattage. You may
be able to get away with much smaller resistors.
"Any obvious mistakes?"
Obvious? No. But since I happen to be a board designer by trade,
I have a few critiques (I tend to be anal about some things.:-P):
1) I'd like to see more meat around the lowest most holes on the LEDs
(Ground?) to help prevent breakout during the drilling operation.
You've got enough room to fatten all the traces up to the size of the
smallest pad.
2) You might think about rotating the LEDs 90 degrees (either way) to
get a straight run in and out of them.
3) Check the datasheet for the LEDs. Are you certain the drills are
the correct size? The look very small relative to everything else.
4) The four pins at the bottom of the key are going to be a challenge
to solder. That fat trace will make a nice heatsink.
5) Unless you're sure you need chassis ground, I'd remove the plating
around the mounting holes. And even if I was sure, I'd look for an
alternative method. Are you planing to have the holes plated through?
6) It doesn't matter with traces this heavy, but try to clean up the
heavy trace at the bottom of R1. The goal is to avoid creating traps
that hold the etchant before the board gets into rinse.
All in all, nice design. And it's fun, isn't it? UNTIL you're under a
deadline to get a 10K+ part, 20-layer, 128 differential-pairs, matched
length, controlled impedance, zero-crosstalk monster done in 30 days!
UGH! :-\
Stephen Weller
American. Veteran. Patriot.
> Is this realistic? Is this something I could give to an
> experienced person to produce? Are the holes for the LEDs and the
> wires too close to work with? Too many resistors too close together?
> Any obvious mistakes?
"Is this realistic?"
Absolutely.
"Is this something I could give to an experienced person to produce?"
Most definitely. Anything single or double-sided you can easily etch
and drill yourself in small runs.
"Are the holes for the LEDs and the wires too close to work with?"
Certainly not. Those willing to pay the premium are fabricating with
.002" lines and .002" spaces!
"Too many resistors too close together?"
Not at the currents you're working with. Re-calc the wattage. You may
be able to get away with much smaller resistors.
"Any obvious mistakes?"
Obvious? No. But since I happen to be a board designer by trade,
I have a few critiques (I tend to be anal about some things.:-P):
1) I'd like to see more meat around the lowest most holes on the LEDs
(Ground?) to help prevent breakout during the drilling operation.
You've got enough room to fatten all the traces up to the size of the
smallest pad.
2) You might think about rotating the LEDs 90 degrees (either way) to
get a straight run in and out of them.
3) Check the datasheet for the LEDs. Are you certain the drills are
the correct size? The look very small relative to everything else.
4) The four pins at the bottom of the key are going to be a challenge
to solder. That fat trace will make a nice heatsink.
5) Unless you're sure you need chassis ground, I'd remove the plating
around the mounting holes. And even if I was sure, I'd look for an
alternative method. Are you planing to have the holes plated through?
6) It doesn't matter with traces this heavy, but try to clean up the
heavy trace at the bottom of R1. The goal is to avoid creating traps
that hold the etchant before the board gets into rinse.
All in all, nice design. And it's fun, isn't it? UNTIL you're under a
deadline to get a 10K+ part, 20-layer, 128 differential-pairs, matched
length, controlled impedance, zero-crosstalk monster done in 30 days!
UGH! :-\
Stephen Weller
American. Veteran. Patriot.