The way Ashhar is using a single ?C I/O line to key the ?BITX is pretty clever, but some people are having issues with different keys/keyers/paddles. I think the main reason he used a voltage divider for key sensing is because he was running out of I/O pins
on the Nano. Looking at the code, the LCD display uses the standard LiquidCrystal library, which eats up 6 I/O pins. I know Allard has Rel. 2.0 of the software going, but I'm wondering if A4 and A5 could be moved and the I2C interface used for the LCD display.
That would free up enough lines for the keyer to dedicate a dit and dah pin for each. Quite honestly, I have not looked at Allard's code yet (too many alligators), so I'm just tossing this out there for discussion.