kony wrote:
The low end boards reduce costs in every possible way.
Yes, but that is certainly NOT always done by re-engineering
or parts substitutions when they also had a more elaborate
or more stringent certification necessary for other global
regions. This is quite evident with many boards.
They may be using
non-UL approved PC board material, and are probably omitting every
non-essential component.
No, not every... you're guessing. Even junky, lowest of the
major low-end PCChips boards have a few SMD caps, resistors
and such going to ports whose physical sockets weren't
installed.
They are probably stuffing the minimum number
of bypass capacitors, using the "remove capacitors one at a time until
the board fails, then add one back" method.
Yes they do reduce capacitor counts or quality in "some"
designs. This is not new, but cannot be assumed to apply in
many cases, either... the boards themselves are the proof,
we need not overgeneralize when there are so many concrete
proofs of these things.
This would be the type of
board where they would change the power layer to supply +5V power
directly to the ports, eliminating the expense of a fuse or jumper, or
of adding a trace.
No, absolutely not. If the upper pads on the top layer are
those for the port/features supplied (working) on a
particular board, they do MOST DEFINITELY NOT just go and
rework the inner layer. How hard is this to grasp? Look at
some boards. I don't mean to be rude about this, but you're
overgeneralizing to the point where the resulting logic is
in error and it's useful to see why that's the case.
It is obvious that with the traces already in place, if
there were to be any cost-reduction in layer rework, the
very most they would do is bridge the two solder pads, for
several reasons. A couple - Power traces (rather than large
planes) should not be in inner layers when fine signal lines
are above them (as the port data lines are). They heat up
more and could delaminate, plus cheap boards tend to have
only 4 layers, no ground layer between these so you would
have the potential for noise pickup.
There are probably traces going under the fuse,
You are again jumping to conclusions. We cannot assume
"probably", and with all the hundreds (if not more)
resistors and caps surface mounted, it is rather trivial to
put a 0 ohm resistor across the pads. To even think about
doing it any other way costs more than just doing it the
same way as (ironically enough) everybody has been doing it.
which
would make it impossible to add a connecting trace without a board
relayout, which is very costly.
Still you are drifting along with ideas based around
erroneous conclusions. What they do, put simply:
The entire board is designed, period. They do not rework
the layers. Surface mounted pads are used, populated for
differing compliance and features. This is the way it IS
done. Look at some boards. No, look at ALL boards. This
is the way it is done. "IF" the board the OP observed, the
board that started this whole thread, deviates from this, it
is unique in this regard.
However, we still have no evidence that these missing
components/pads are actually supply for the implemented
ports on that board, rather than there being another
trace(s), on the surface layer, for power.
I worked for the 2nd largest motherboard company (at the time) and we'd
do anything reasonable to save a couple of cents per board, including
changes to PCB layers, if the volume was high enough.
Sure, if there was justification, a way to do so it could be
done for some things. That's not evidence that this is
what's happening to modern motherboards for port power
traces. The motherboards themselves are evidence.