Working duplex matrix!

2022-11-09

A working POC of my duplex matrix! In rust!

Wooooo! I just got this working. But it wooorks! It’s a really simple test really. I make use of the neopixel that is on the xiao to make it easier for me to test. Lots of colours in the same led so simple to get lots of data into it.

Full code can be found here https://github.com/qwelyt/couscous/blob/e6329efceba72f70873f93c6cf267ccd8ebec83e/firmware/couscous/src/main.rs

The thing is actually really simple.

loop {
        pin5.into_pull_down_input();
        pin6.into_push_pull_output();
        pin6.set_high().unwrap();

        let blu = pin5.is_high().unwrap();
        pin6.set_low().unwrap();

        pin6.into_pull_down_input();
        pin5.into_push_pull_output();
        pin5.set_high().unwrap();
        let gre = pin6.is_high().unwrap();
        pin5.set_low().unwrap();


        if blu && gre {
            ws.write(brightness(once(wheel(192)), 32)).unwrap();
        } else if blu {
            ws.write(brightness(once(wheel(128)), 32)).unwrap();
        } else if gre {
            ws.write(brightness(once(wheel(64)), 32)).unwrap();
        } else {
            ws.write(brightness(once(wheel(255)), 32)).unwrap();
        }
        delay.delay_ms(200);
    }

As you can see, pin5 and pin6 scan in one direction first, and in the other direction right after. And based on that we change colour of the neopixel (ws).

Super exited about this actually. It means that what I want to do is possible. 60 keys on 11 pins. That kind of blows me away. The next step is either to code the full matrix scanning, or moving this example to rtic. Using real time interrupts can be nicer going forward, but I’ve always beaten my own path before. We’ll see. Very nice to see it actually working.