Monochron - Programming thru FTDI

[mcw]

kit build was smooth - very satisfying. initial programming thru a usbtiny.

but when I arrived at reprogramming time, I got the dreaded "avrdude: stk500_getsync(): not in sync: resp=0x00"
(actually sometimes I get other resp= values, but always errors out.)

so far I have established:
1. FTDI Friend works (can wiggle the TX line from hyperterm & see on scope)
2. The clock is resetting (screen flash + beep each time avrdude makes an attempt)
3. TX light flashes a few times before giving up
4. All connections are good from AVR pins 2,3 to FTDI chip pins
5. The bootloader seems to be there, when I leave bootloader only on the chip it flashes the backlight of the LCD

and one odd fact
6. it does occasionally get into a mode where the RX light is flashing once a second as the clock operates (?)

next step is to write some test firmware to checkout both the chip's usart and the friend,
but as that'll take me awhile to get round to-- anyone have other troubleshooting suggestions ?

[adafruit]

we think its because there isnt a pullup on the reset line. one thing that does work - but is annoyingish - is to physically plug in the power cord and program within 3 seconds?

[mcw]

thanks for the heads up, if I run into any intermittent function I'll look there.
but looks like it wasn't part of the trouble here.

i'll start at the result and work backwards: something was up with the bootloader. works now.
something I can't recreate was wrong, maybe the .hex was corrupted when it was extracted from the archive, I'll never know...

on the way, I compiled the bootloader myself so I could read the code & tweak to confirm it was all there, and in the process i ran into a standing arduino bug: http://code.google.com/p/arduino/issues/detail?id=152
which could just be fixed in the monochron firmware... EEWE should be EEPE for the '328

as soon as i loaded my fresh build, everything worked.
when i went back to the stock august 10th .hex file, it still magically works. huh.

for the sake of others troubleshooting/playing/learning it is worth noting that the DEBUG statements in the code exercise the FTDI/serial path... just open a terminal on your FTDI port and watch the debug info stream out of the monochron. I tweaked the code to echo typed characters, thus getting a full loopback test out of it.
(of course it showed that everything was functioning.)

and for the smallest mystery: the dalichron branch is sitting in github with DEBUGGING enabled,
that is where my once-a-second RX light was coming from.

again, thanks. fun clock. now back to modifying code...