Hey guys,

I'm having some problems with the 14000 chip, and am about
to pull my hair out.  I'm using Hi-Tech C.

[1] I swiped the synchronous serial code from the serial.c
    sample provided with the compiler.  Seems to work fine
    on the 14000, as long as you change the loop delay
    variable.  After a bunch of fiddling, I got it to
    work fine (for Tx only - it's my diagnostic tool
    for debugging, which is hooked to a serial LCD).

    Then I added a/d code, which I hacked together from
    AN626 and updated to reflect the capbilities of the
    Hi-Tech compiler.  Surprisingly, this compiled
    and worked perfectly the first time I tried it.
    (NOTE on A/D code and JW parts: cover the window
      with something opaque, as light throws off the calibration).

    The only problem is that because of some necessary
    delay loop code, I had to compile with full optimization (-O -Gz9).

    This completely screwed up the serial port code.  I thought it
    was just that the loop delay variable needs to be reset,
    but alas, no, nothing I try works (and here comes the weird part)
    on 8 out of 10 of the 14000's I've got.  Works just peachy
    on the other two.  I figured it was probably just minor
    timing differences of the on-board oscillators, but I
    can't get any value to consistently work on the other 8.

    Anybody have any ideas, or know where I can find a simple
    synchronous Tx serial driver that doesn't depend on
    counting instruction overhead?

[2] I'm having real trouble with the i2c sample code provided
    with the compiler - I can apparently write to an I2C
    slave, but read cycles are a total mess.  I'm sure this
    has to do with some of their timing assumptions, but
    I really don't want to mess with those - there's just too
    many permutations to deal with.

    Anybody know where I can find some 14000 sample code
    for using the on-board I2C module?

Thanks,

-Will
"Life is a meritocracy"