Interface push button rotary encoder on 1 analog input

[David Bender 28]

  I had a lack of digital input pins for a pushbutton rotary encoder switch so I used an analog input.

  I wrote up my results here: https://analog10.com/posts/rotary_encoder_analog_input.html

  It works pretty well except for an occasional reverse tick but that's probably a flaw in my code.

 

[roadrunner84 466]

Alternative: use a one-wire I/O expander and make a software library for that

[timotet 44]

Great idea!

Good thing your 1 left over pin was an ADC pin!

[dubnet 238]

Made the Hackaday blog....

[greeeg 460]

  It works pretty well except for an occasional reverse tick but that's probably a flaw in my code.

 

The reverse ticks might be due to a "blocking" part of your code when it calls "begin_lcd_write". (Blocking probably isn't the right word.)

But while your main loop is inside this function it could miss an update from the ISR. I'm probably wrong. Just a thought.

 

I've found whenever I used rotary encoders I had to proccess the transitions quickly because there was a fair amount of phantom transitions, I was probably using very cheap rotary encoders too.

[David Bender 28]

begin_lcd_write() just buffers data and initializes the TA0.1 IE.

The LCD data are clocked in the timer interrupt handler so there really isn't any blocking.

 

I think an improvement would be to supersample 8 times, average and then use the change in value as a determination of how far the encoder turned, rather than only comparing discrete levels.