Have you considered using interrupts to handle the GPIO pins connected to the keying inputs? ?Interrupts might reduce or eliminate the possibility of missing the first dot before a dash. ?
?
A priority interrupt handler would need to be written with good real-time coding practices (bounded time algorithm, no locks, no API calls that could involve memory management, etc.). ? But this should be possible. ?Then post the key inputs with time stamps to be handled by the priority keying sidetone process, which perhaps will run only a few milliseconds later. ?The time stamps on the interrupt data would prevent the dash from being handled before any dot, if that was the key closure sequence, and solve at least one issue with fast iambic keying that polling with a slow CPU on a loaded system might never be able to completely solve.
?
73, Ron, N6YWU
