I am programming an FPGA that uses push buttons as input signals. I am programming a finite state machine with 11 states that transition from state to state using specific button presses. For example, in my design, state 0 goes to state 1 using button1. This is the same transition case from state 1 to 2 and from state 2 to 3. This state transition system is implemented in my vhdl code using case statements much like the "switch" statement in c++. I am lighting up certain LEDs in each state to keep track of which state the board is currently in. My issue is that when I press button1 while in sate 0, the board shows that it has moved to state 3. What i think is happening is that it is indeed going to state 1 and 2 but the same button press in state 0 is also being read in state 1 and 2. This happens so fast that the boards doesnt have enough time to show the LED indications for state 1 and 2. It stops at state 3 because state 3 moves to state 4 using a different button. So my question is, how do I make the button press signal have a rising edge and a falling edge such that a single button press is read only in one state and not the ones that follow it? The press button signals are debounced but this only makes the signal a uniform wave. Any help is appreciated, thanks. I attached my vhdl code. The "btnC, btnL, btnR,..." are the push buttons.
You could insert additional states which are waiting for button release. For instance, instead of transitioning directly from state 0 to 1, insert another state which is waiting for button release and then switches to state 1.
Just switch on the rising edge? Something like : http://insights.sigasi.com/tech/clock-edge-detection.html
FSM: process(clk, my_btnC)
if(my_btnC ='1') then
my_state <= s0;
NEVER EVER use a async input directly as a reset source. You may encounter spurious problems when leaving the reset. This is because your FSM are some flipflops in reality. And the reset goes to their reset input. When the reset is deactivated at the same moment (+-a few ps) when a rising_edge on the lock occurs (and that is about 10..100 million times per second) then some of the FSM flipflops see a deactivated reset and some not. And therefore the FSM starts with a undetermined wrong state... Calibroflower wrote: > Any help is appreciated, thanks. Run a simulation. And see how your code reacts to "short" and "long" key presses. Then you will find out that you edge detection (and before that debouncing) on each button input.
: Edited by Moderator