PiBoy-Zero

been using the Pi-Boy on and off for the last couple of months - works beautifully :-)

some things need to be redone though;

- messed up front panel (heat-setting M3-thread *after* painting was a dumb idea...)

- the power button needs a redesign (button-cap is missing)

- have to add a bit material to the back/front to close the slight gap between the two

on a side note: what turned out really well where the ABXY buttons - the combination of the soft-tactile omron switches plus a coat of epoxy feels just right while playing :-D

the way i build them was:

- print the button-piece (one part, all four are connected by a bridge - to ease assembly) with the letters indented

- fill the letters with acrylic paint

- put a drop of XTC-3D (some other epoxy might work too) on top to form a slight dome

for the thirtieth log something special - it lives! :-D

- battery: check

- display: check

- controls: check

- sound: check

all systems go:

there are still many rough spots:

- on the software side there are many daemons still missing (pwm back-light, power management, ...)

- the system needs further configuration: the buttons currently only work in emulation-station menu, not the emulators (probably sth minor)

- the case needs further refinement, lots of sanding, priming and a few layers of paint :-)

- the setup/configuration steps need more documentation or better: scripted as far as possible (rough script already on github)

- print and assembly instructions?

creating a 3d printable case from the rough sketches done earlier

the overall shape was put toghether relatively quickly - but getting the d-pad to feel right took some iterations, the first few where still cross-shaped

turned out, that shrinking and making it round works better for the alps switch. connecting the ABXY buttons eases installation

also playing with the back part, for example splitting it in two with overlapping part along the seam:

which would be printed standing up:

this would have the upside of a sturdier frame - but on the downside prints waaay longer, with increased risk of print failure and the split/seam doesn't look as nice as a single-pice backside does, paintability would probably be another issue - so i'll stick with the single-piece for now

btw: the front piece in its current form uses the pause-print-insert-M3-trapped-nuts-resume-print approach:

STLs to print your own are on github :-)

another wrong footprint...the push buttons used for start/select group their four pins in pairs - the other way round

... really need to learn to double check the datasheet for all components that have more than two leads - even trivial ones like buttons :-P

oh well - the fix was to desolder the button, cut some traces and amputate a leg >:-)

updated the pcb layout and pushed the changes to git and osh-park

for anyone who happen to have this buggy version, the fix looks like this:

cut the start/select traces at the red lines, and do not solder the button pads marked with the circle...

mainly buttons and the usb socket...

the usb-socket has mounting tabs which i decided to trim down at the side close to the alps-nav switch, just in case the room is needed later on by the 3d printed that'll come ontop of it

the omron switches chosen for ABXY have some position keys on the back, which have to be trimmed also -- which was a design decision, including holes for those would have interfered with the smd-pads on the other side

in retrospect it would have been a bit easier to mount the tft last, since it sits close to some solder-pads... but it was still doable

the adafruit guide for pwm-audio on the pi-zero ( https://learn.adafruit.com/adding-basic-audio-ouput-to-raspberry-pi-zero/pi-zero-pwm-audio ) configures the gpio-alt functions on the fly to remap the audio-pwm pins from hidden ones to pins on the gpio-pinheader
after reading up on

device tree configurations and tracking down all necessary pieces, i got this this code snippet/overlay to do the same at boot-time:

/dts-v1/;
/plugin/;

/ {
  compatible = "brcm,bcm2708";

  fragment@0 {
    target = <&gpio>;
    __overlay__ {
      pinctrl-names = "default";
      pinctrl-0 = <&pwm_audio_pins>;

    pwm_audio_pins: pwm_audio_pins {
	brcm,pins = <13 18>;   /* gpio no ('BCM' number) */
	brcm,function = <4 2>; /* 0:in, 1:out, 2: alt5, 3: alt4, 4: alt0, 5: alt1, 6: alt2, 7: alt3 */
	brcm,pull = <0 0>;     /* 2:up 1:down 0:none */
      };
    };
  };
 };

this maps the left channel to pin 33 (BCM=13) and the right channel to pin 12 (BCM=18)

compiled with:

dtc -@ -I dts -O dtb -o pwm-audio-pi-zero-overlay.dtb pwm-audio-pi-zero-overlay.dts

dtoverlay=pwm-audio-pi-zero

helpful links:

http://elinux.org/RPi_BCM2835_GPIOs - pins and alt functions for the pi-zeros cpu

https://learn.adafruit.com/introducing-the-raspberry-pi-zero/audio-outputs - the other adafruit guide, touching opi-zero pwm audio

https://www.raspberrypi.org/documentation/configuration/pin-configuration.md

https://www.kernel.org/doc/Documentation/devicetree/bindings/pinctrl/brcm,bcm2835-gpio.txt - documentation on the gpio pinconfig via dtoverlay

https://github.com/fivdi/onoff/wiki/Enabling-Pullup-and-Pulldown-Resistors-on-The-Raspberry-Pi - good example of gpio config via dt-overlay

ok, again repopulating the board with the components grouped by function (the raspberry-zero and the powerboost 500 are already mounted)

first: soft-power

connecting the lipo battery and hitting the power button the pi boots, and pulling the gpio down powers it of :-D

second: audio

NOTE: the components in the top right corner are left/right channel = two times the same set, but their placement got a bit mixed up = is not symmetrical... oh well, just have to follow the silkscreen :-P

Testing the audio with a pair of headphones: works! :-)

(when nothing is playing - eg the channels are silent - there seems to be a tiny bit of high frequency noise coming from somewhere - and sd-io is also slightly audible... i wonder if that is just due to the anaolog/pwm nature of the system or bad board design/soldering :-S )

third: backlight and tft display:

the 2.2" display is held by a metal bracket, which in turn is glued with strips of double sided tape to the donor-pcb.

a sharp knife makes short work of these:

reapplied double-sided tape to the bracket and glued it to the pi-boy pcb - aligned to the right edge of the silkscreen square

transplanting the tft screen:

* the ribbon cable of the screen was desoldered from the donor-pcb with the previously mentioned tea-candle method

* pre-tinning the contacts on the pi-boy pcb

* and finally re-soldering the 14 contact

* double/tripple check the solder-joints/solder bridges since the contact pitch is ultra small :-P

powering the hole thing up and we have picture! :-D

(upside down, but thats easily fixed in the config)

similar projects that mounted the raspberry onto another pcb seem to reflow it in some way or another, with a hot-air rework station or similar tools. the same is doable with a good fine tipped soldering iron.

these steps seemed to produce good results:

1) widen the mounting hole near the hdmi to 3mm, which will be used as position key (and later on by a screw holding the case together) - done by hand, very carefully :-)

2) pre-tin all the surface mount pads on the base-board with a slight (!) amount of solder

3) use a M3 screw to get an accurate position, clamp the boards together and solder by first heating up the rim of the pin-hole, then feeding solder into it until the hole is roughly half filled. at this point it should connect with the pad below and visibly suck the solder a bit down

the tip i used fit halfway through the pin-holes, which i could put into the soldered holes that seemed to have a hard time connecting to the pad below (was the case with some of the holes connected to the ground plane, since they sucked the heat away pretty fast)

not wanting to discard the perfectly good smd-components that are already on the Mk1 pcb a candle-light desoldering session was in order: the Mk1 pcb below - roughly 1cm above the flame of a small tea light candle - and the Mk2 above ready to receive the transplant

moving the candle under the specific components to desolder; small tweezers in one, the solder iron in the other hand - worked like a charm :-)

things of beauty :-D

time to desolder the Mk1 and populate the Mk2...