latest c139 build

Mychaela Falconia mychaela.falconia at gmail.com
Wed Oct 26 03:30:53 UTC 2016


Hello everyone,

While it is still as unusable in practice as before, I just made another
C139 UI demo fw build corresponding to the latest commit in fc-magnetite:

https://www.freecalypso.org/members/falcon/experimental-builds/c139-uidemo-build20161026.tar.bz2

Ever since I first got it running on the piezo-buzzer-equipped C139
about this time last year, I've always wondered why the ringtone sound
produced by TI's demo UI code is so horrible.  Well, I finally figured
out why: yet another miscommunication between different parts of TI.
The German knuckleheads at TI-Berlin (former Condat GmbH) assumed that
the BZ_Tone() function takes the desired tone frequency in Hz as its
argument, whereas the actual implementation in the chipset driver code
from TI-France takes the timer value to be loaded into the hardware -
the two are inversely related.

As a result, all BZ_Tone() arguments sent by Condat's code were well
out of range, and clipped to the maximum timer load value of 255,
which corresponds to about 99 Hz, the lowest frequency which the buzzer
driving circuit can produce.  Thus no matter what tone frequency was
intended, the broken fw was only generating 99 Hz tones.

Naturally I could not allow such an egregious bug to remain, so I fixed
it, and made a new binary build too.  All of the other show-stopping
bugs are still there (the only other change since the last build is
the addition of my audio debugging AT commands which end users can
safely ignore), so there is no improvement in usability unfortunately,
but once again, I could not allow such an egregious bug to remain.

If you do flash this new build and make a call to the phone with this
fw, you will now hear a much nicer ringtone - super-basic for sure,
but not horrible any more.

M~

P.S. Yes, as programmers we all make bugs in our code all the time -
but what blows my mind is that TI *shipped* their code like this, and
had this bogosity in their shipping code for many years, it seems: the
ancient 20020917 fw image that came with the D-Sample kit from that
era produces the very same horrible 99 Hz "ring" sound.


More information about the Community mailing list