Some issues about key detection in STM32

1. Compare the method of detecting keystrokes by delay and debouncing with the method of detecting keystrokes by interruption: For the method of detecting keystrokes by interruption, if you do not do some processing in the program to detect whether the keystrokes are pressed again, its accuracy is not as reliable as the method of detecting keystrokes by delay and debouncing.

2. The reason is: when we use interrupts to detect buttons, there is no delay in the interrupt program, then there will be a problem. When the button is pressed, there may be multiple jitters. Although we set the falling edge to trigger the interrupt, there may also be multiple falling edges in these multiple jitters, so the interrupt may be triggered multiple times.

3. The only solution is to use delay. The corrected code is shown in Figure 3 below. However, the code in P90 of the "STM32 Library Development Practical Guide" will cause false triggering.

Figure 1: Traditional delay debounce detection key program

Figure 2: Rough delay function

Figure 3: Corrected code for detecting keystrokes using interrupts

4. So the question is, if delay is also used in the interrupt program, what is the difference between it and the traditional delay debouncing?

The difference is: the traditional delay debouncing program should be written in the while loop of the main function, constantly detecting whether the button is pressed, occupying CPU resources in vain; while the interrupt method will not affect the execution of other programs. Only when the button is pressed will the interrupt be triggered and the interrupt service program will be entered.