FCDEV3B V2 boards are here

[Mychaela Falconia]

Hello again everyone,

Our FCDEV3B V2 boards were reworked at Technotronix yesterday, and the
loudspeaker amplifier output now has the expected Vpp swing of about
3.6 V on each leg at maximum volume with sufficient power supply
voltage.  As I suspected, someone swapped the resistor values between
R331&R332 (feedback resistors) and R351&R352 (input resistors), putting
the amplifier gain below unity instead of the expected 1.8 value.  But
it is fixed now, as I made them rework all 8 boards.  They did the
rework at their expense because the incorrect resistor population was
their mistake (my BOM document is unambiguous), and hopefully they
will populate these resistors correctly to begin with on our subsequent
batches.

In terms of the count, we have 5 fully working boards (out of which I
need to keep one for my own development) and 3 partially defective
ones, leaving us with 4 defect-free boards available for either
commercial sale or subsidized give-away to strategically chosen
parties.  Of the 3 partially defective boards, the status is as
follows:

* Prior to yesterday's rework we had one board with completely dead RF
(both Rx and Tx in all 3 bands); I suspected the antenna switch
component and asked the folks at Technotronix to take a closer look at
it.  They saw nothing wrong on X-ray, but they resoldered it as an
experiment - same part, no replacement, but added a little more solder
to the joints.  After this rework the board perfectly passed the full
calibration suite (Rx and Tx in all 3 bands) and appears to be
perfectly good.  This is the board which I am keeping for my own
development, so we don't take chances with boards sent out to other
people.

* One board has a bad EGSM (900 MHz band) Rx path, while all other RF
paths and all other functionality work fine.  It was subjected to the
same antenna switch resoldering as the other board, but nothing
changed.  I don't know when or if I am going to attempt further
experiments on this board, as I have far too many other work items.

* Two boards have perfectly good RF and loudspeaker circuits, but have
trouble with the SIM socket: poor or no contact is made with the card
in the socket, and the firmware throws up a "SIM not inserted" error.
On one of these two boards the SIM socket appears to be a solid fail,
but on the other board it exhibits intermittent behaviour: sometimes
works, sometimes doesn't.

The SIM socket has now certainly earned the reputation of being a
repeating trouble spot (we had two boards with SIM socket problems in
our very first batch as well), and I will definitely use a different
socket part on our next design.  But for the current FCDEV3B product
we may just have to deal with poor yield in this regard: there is no
universal standard PCB footprint for SIM sockets, different
manufacturers' socket parts have different footprints, and there is
absolutely no way I can pay for yet another PCB respin to change the
SIM socket footprint.

One possible solution to the SIM socket problem may be to sell the
boards with defective SIM sockets (at a steeply discounted price, of
course) to those specialized users who don't need a local SIM directly
connected to the modem - such a directly connected SIM is required for
Standard Modem operation, but I know of some users who are using or
thinking about using our modems for some highly non-standard
applications, and to the best of my understanding, those applications
don't need a working SIM socket.  If you are in that category, I now
have two boards with perfectly good RF in all 3 bands but flaky or
dead SIM sockets, and you can have them for very cheap.

Hasta la Victoria, Siempre,
Mychaela aka The Mother