U6002 mystery and solution

[Werner Almesberger]

There's an interesting inconsistency eagle-eyes Alvaro discovered when
we discussed the symbol for the low-pass filters for gta02-core: in
the non-public GTA02v6 schematics, a low-pass filter symbol is used:
http://people.openmoko.org/werner/filter.png

Note that, if we assume that pins 1 through 4 are numbered in the same
way in both places (starting with pin 5, they diverge, but that's an
artefact - the LCM wouldn't work at all if that divergence existed in
reality. This was probably done to reuse some footprint with different
numbering.), they are connected to the inputs of the filter. Here's
the data sheet of it:
http://www.acxc.com.tw/product/fa2012/FA2012-KR40PAA_070827.pdf

In the public schematics, the low-pass filter symbol has been replaced
by a resistor block. What's odd there is that the inputs have changed
place with the outputs.

Now we wondered if this is something that happened for real or if it's
just a glitch that was introduced when cleaning up the schematics. The
good news is that, according to the published netlist (which is
straight from the full GTA02v7 schematics), pins 1 through 4 connect
to the Glamo and pins 5 through 8 (should probably be 6 through 9) to
the LCM connector.

So hopefully nobody will get confused by this :-)

I don't know what this is like in the GTA2v8 schematics, but given
that no GTA02v8-based devices have been made (yet ?), that's not so
critical.

By the way, while we're dragging the bugs out of the old schematics,
that must make them look pretty awful. However, there are also nice
things to find there. E.g., the low-pass filter symbol is much more
elegant than what I would have suggested.

- Werner