FCDEV3B loudspeaker and microphone

[Mychaela Falconia]

Hello FreeCalypso community,

I have just performed the experiment I've been meaning to do since a
couple of weeks ago: I managed to connect the loudspeaker and
microphone that are contained in the handset part of TI's D-Sample kit
to the loudspeaker driver and microphone input circuits on our own
FCDEV3B, and tested the resulting operation.  The results are:

* When making a test call from this setup, the call downlink audio is
  heard loud and clear, just like when the speaker is driven by the
  original D-Sample board, quite unlike the little Soberton SP-1510
  speaker which was quiet as a mouse.  Thus we know that the
  loudspeaker driver circuit on our boards is fine, no worse than
  whatever TI implemented on the D-Sample, and the speaker loudness
  problem was because of a poor choice of the physical speaker part or
  some acoustical/mechanical issues around that speaker which I am
  ignorant of.

* I then made a test call to a voice mail system, made a test recording
  through the microphone, and it sounded just fine when I played it
  back.  This result proves that the microphone input circuit on our
  boards (which has never been tested before) is also fine.

* In the next test with a real person on the other end of the test
  call, she told me that my voice sounded fine through this call setup,
  but there is a strong delayed echo present from the microphone on my
  end picking up the sound from the loudspeaker.  This effect is a
  very well-known problem with all hands-free call setups, and the
  Calypso DSP includes an Acoustic Echo Cancellation feature
  specifically for dealing with this problem.  But this AEC is not
  enabled automatically, and it has a bunch of configuration settings
  which we have yet to tune.

* As a final test, I uploaded the ringtone melodies in TI's Melody E1
  format which we've extracted from the TSM30 source to the FCDEV3B
  and played them through this loudspeaker setup.  They played
  perfectly, confirming once and for all that yes, we can play nice-
  sounding melodies through a loudspeaker driven by the Calypso DSP,
  without using an external ringtone generator chip or a piezoelectic
  buzzer like Motorola and Pirelli phones do.

These results confirm that the loudspeaker driver and microphone input
circuits on our FCDEV3B are good and not in need of any further
revision.  Thus if anyone buys an FCDEV3B and wishes to have the board
drive a speaker that is loud enough to be comfortably heard while
resting on a lab bench, they can be assured that the loudspeaker
driver circuit on the board is fit to do the job, but they will need
to acquire the actual loudspeaker part themselves, as I lack the
necessary expertise to make that speaker part selection.  The same
goes for the microphone input circuit: you'll need to acquire the
actual microphone part on your own (and use your own expertise to
select that part), and wire it up to J313 on the board.  I am going to
update Appendix A in the FCDEV3B User's Manual accordingly.

At some point we'll need to find our own loudspeaker part that is no
worse than the one in TI's D-Sample handset in terms of its acoustic
output, and figure out what kind of special acoustic/mechanical
arrangement is needed around that speaker, if any - TI's D-Sample
handset only exists in singular quantity, and we need to be able to
put our development kits into the hands of whoever needs one.  I will
try disassembling the D-Sample handset to see the speaker inside, and
we may also have to hire someone who (unlike me) has a clue about
speakers and acoustics.

I am also thinking about putting together a mechanically rigid FCDEV3B
setup that would consist of the FCDEV3B itself, the FT2232D adapter
for it, the to-be-figured-out loudspeaker and a Li-ion battery that
can be easily taken out for external charging and put back in, all
rigidly mounted on a substrate such as plexiglas, so that the complete
setup with all of its interconnections can be transported as a unit,
without having to dismantle and reconnect everything.  Such portable-
as-a-unit setup will be absolutely needed in order to test truly
mobile operation with cell reselection, location updates and handovers
- I would need to be able to get into a car with this portable FCDEV3B
setup and a laptop, and exercise mobile functionality while my partner
drives the car around.  But creating this mechanical setup will be a
lot of work, so it'll be a while before we have it.

Hasta la Victoria, Siempre,
Mychaela aka The Mother