PiBoy-Zero

just using wiring-pi or the python GPIO classes to drive a software pin did work, but tended to flicker ever so often - since there is so much software overhead involved and the kernel rescheduling the python process :P

what seems to work much better is a software-pwm that uses the hardware PWM/PCM subsystem of the broadcom chip - like

https://github.com/sarfata/pi-blaster/

with some minor modifications to the source and the '-pcm' switch set (can't use the PWM subsystem, since it will be busy generating audio) the LED (used as placeholder for the backlight circuit) stopped flickering :-D

NB: there is another trade-off: by using the PCM subsystem for software-pwm we lose the serial-console. so if one needs the (hardware)serial to debug, the backlight service has to be disabled

and 3d-components!

this revision fixes these errors:

- missing trace connecting powerboost:bat-low to the pi-zero
- wrong transistor pinout Q3 - turned out to be not needed and removed anyway

- clearance for protuding hdmi-connector tabs

- mounting holes undersized
- rerouted usb-traces (equal length, less vias and traces crossing)

- corrected usb-solder pad position

- wrong pinout on BJT transistor for backlight

diff of the copper layers:

putting the components for the audio filter in, only for one channel though - since there will be a follow-up revision to address the other bugs found so far. still this tells us if audio this way works or not

and it does! :-D

somewhat audible sound came out of this mini speaker (just salvaged it out of a broken tablet to see if it works, don't intend to keep it hooked up) - the target in-ear headphones worked even better :-)

the software side followed this adafruit tutorial:

https://learn.adafruit.com/adding-basic-audio-ouput-to-raspberry-pi-zero/pi-zero-pwm-audio

the next steps on the audio topic would be to read-up on the device-tree configuration to properly setup the two pwm channels instead of the on-the-fly configuration

while the front solder pads line up nicely

the through-hole solder-points for the usb connection do not :-(

, the hdmi socket has some mounting tabs which protrude a tiny bit

will have to include some sort of cutout to accommodate those :-\

this is what happens if you start designing a PCB without the major components at hand to recheck the dimensions :-P

oh well, i'll keep on adding and testing the other functional groups - maybe there are more bugs waiting to be uncovered :-)

just came with the mail :-D - point of origin: Pasadena CA

thanks Adafruit and Hackaday!!

putting the components on the pcb, grouped by function - power is first

and temporarily mount the power-boost pcb (to ease removal in the worst case) - looks ugly, but gets the job done...

the two wires connect to the LiPo Battery - in the final assembly the battery will be directly soldered to those pins, since the battery jst-socket is (intentionally - was a trade-off decision...) blocked by the USB-A socket

And the next bug... ordered the wrong transistor that pulls on the enable pin of the PowerBoost... have to fix the BOM and order another one :-\

replacement transistor arrived, turned out it didn't work as intended either :P

after some playing around with it, noticed that the pulldown (to the powerboosts enable pin) without the inversion/buffer transistor in-between just works...

since the voltage levels in the soft-power part (tow latching mosfets) are already on V_Batt level, and the connection to the raspberry is "protected" by another resistor, everything should be fine. the whole soft-power circuit was designed to turn 5V on/off, so the 3-4.2V of the lipo shouldn't hurt the pi (should, should, ...).

updated the schematics and boardlayout accordingly (see github)

word to the wise: don't put smd parts for multiple projects into one order, or see to it that the part-reference is reflected in the packaging listing of the distributer - spent the better part of the evening picking them apart due to a slight mistake in the BOM (which has been fixed and pushed to git) :-P

beautiful purple :-)

still waiting on the shipment with all the other components though - got delayed because of some parts where not in stock :-\

time to sanity-check the pcbs...

just noticed that i didn't add a license text to the github repository, which one to pick...

- the CERN Open Hardware License looks interesting ( http://www.ohwr.org/projects/cernohl/wiki )

- so does the (L)GPL...

but the former only seems to cover hardware design files, whereas the later only covers software/libraries (the scripts and code that will be added later) :-/

... let's fall back to an MIT license for the time being :-)

still waiting for the first pcbs to be made and arrive, and already found the first bug - forgot to wire the low-battery pin to the raspberry :-P

already fixed it in the schematic, but will have to jump-wire that one on the pcbs...