Why a New Design?¶
Published on 2022-05-09 in PewPew LCD.
So far the most successful among the PewPew family of devices has been the PewPew Standalone – I ran many workshops with it, I sold it on Tindie, I made several badges for conferences out of it, and now @Makerfabs is selling it for general population to use. Despite being a bit clunky, it’s simple and cheap and gets the job done very well. I have later experimented with other designs, adding various display screens, stronger microcontrollers, a variety of buttons and touh pads, different LEDs and recently even WiFi, but the resulting designs were inevitably more complex and more expensive. Now I feel like I can put together what I have learned through all this experimentation, and design a device that is even better for running workshops, less complex, and cheaper than the Standalone.
The problematic component is the LED matrix. Yes, it’s cheap, but that’s pretty much its only advantage. The original FeatherWing version used a bi-color matrix, with four possible colors. In the Standalone I replaced it with a much easier to source and to drive, and also cheaper, monochrome matrix, that is driven in the background by a special software library compiled into CircuitPython. There are no longer any colors, just shades of light, and to save pins, the same library also scans for the button presses. This leads to some visual artifacts and of course displaying text is not easy.
I have tried replacing the LED matrix with individual LEDs – either bi-color or monochrome, but while this makes the device considerably slimmer and lighter, it increases the assembly cost – with all those discrete LEDs having to be soldered.
Adding a color TFT screen required more memory and more processing power, and while the resulting device is very nice on its own, and lets you make much more advanced and nicer games, it is also more expensive and complex – not suitable for quick workshops anymore.
Finally, I was pretty happy with a monochrome OLED screen. The additional code required forced me to shrink the filesystem on the device a little bit, from 64kB to 48kB, but you can still fit all the games that have been written for the device so far on it all at once, so I don’t consider it a big drawback. The screen is the wrong shape, landscape rather than square, so a lot of it is wasted, which is not great, but acceptable. The additional components required by the display, and the large number of connections of the display itself contributed to both the complexity and price, though. I tried to offset it by shedding components such as buttons, battery holder, USB socket and expansion socket with pads and cutouts in the PCB itself – and I am happy with the results – but I feel like I can do better still, with a different display.