Fix issue #56 by increasing debounce time to 20 ms

[sruggier]

On my own ANSI keyboard module (a rigorously chosen sample size of 1) this is necessary for the keyboard to work correctly. Otherwise, I frequently see duplicate key down events.

I also tried 10 ms, and it was much better with that configuration, but I was still seeing the occasional bounce, so I doubled it to 20.

I wasn't sure whether to apply this to the four keyboard modules only, or to include the numpad and macropad as well, but I chose to start with all of them, because the consequences of the debounce time being too low are much worse than if it's set more conservatively than it needs to be. It would be nice if the debounce time were dynamically configurable without having to flash new firmware, but unless/until that happens, I think this is a safer default value.

[biglizards]

had a similar issue on an iso keyboard (limited, as far as i could tell, to the enter key)

this certainly fixes it, but occasionally (especially for weak keypresses towards the edge of the key) it doesn't register them at all, wheras before it would have registered them once or twice. and i'm not sure if this is my imagination, but typing feels ever-so-slightly less responsive

i'll try this version out for a few days, then 10ms, and report back

[sruggier]

Yeah, it's probably worth trying values in between, to try to find the minimum value that still fully fixes the problem.