How would I poll 14 inputs (p0 to p13) for a...

deno

single button press and debounce that button press?· The "debounce" part is what I am really interested in.·

Right now I am using the following code, which works, but every once in a while I get a "bounce" and the routine will run twice instead of just one time.

'.........................................VARIABLES...............................................
button_command········ VAR·············· INS
'.........................................CONSTANTS.............................................
'.........................................MAIN PROGRAM.........................................
reset:···················· 'program will return here when finished with the selected routine
PAUSE 500·············· 'pause time to get finger off the button
DO
LOOP UNTIL button_command > 0
· SELECT button_command
··· CASE· 512····························· 'typed_char-for serial communication on PC
······ ....
··· CASE· 2048
····· ....
··· CASE· 4096
····· ....
··· CASE· 1024
····· ....
··· CASE· 8192
····· ....
··· CASE· 256
····· ...
··· CASE· 16
····· ....
··· CASE· 8
····· ....
··· CASE 1
····· ....
··· CASE 32
····· ....
··· CASE 128
····· ....
··· CASE 2
····· ....
··· CASE 4
····· ....
··· CASE 64
····· ....
· ENDSELECT
Is there a way I could use the BUTTON command?
Thank you for your input
Deno

Jon Williams

No, you can't use the BUTTON command for that.· A better approach, in my opinion, is to use a loop to debounce the inputs and then alias the bits your return variable.· Something like this:

myButtons··· VAR··· Word
startProc····VAR··· myButtons.BIT0
endProc····· VAR··· myButtons.BIT1
idx··········VAR··· Nib·

Main:
· GOSUB Get_Buttons
· IF (startProc = 1) THEN
··· ' start the process
· ENDIF
· IF (endProc = 1) THEN
··· ' end the process
· ENDIF
· GOTO Main

Get_Buttons:
· myButtons = %111111111111········ ' assume pressed
· FOR idx = 1 TO 10·················' test 10 times
··· myButtons = myButtons & INS···· ' validate inputs
··· PAUSE 5
· NEXT
· RETURN

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Jon Williams
Applications Engineer, Parallax

KenM

Jon,
Please help me to understand your custom debounce routine.
For the moment, ignoring the overhead portion of your routing, if the input pin (myButtons.BIT0) is still low after 50 mS then indeed the button is low.
What is the difference between waiting for the button to go low, pause 50 mS test again for low?
It seems that your code basically will return low only if the button is low during the last iteration of the FOR NEXT loop.
In my tiny programming mind, I don't see the difference between simply waiting 50 mS and updating every 5 mS as in your code?
Also, since the GOSUB is constantly run, what is to prevent the button being pressed 40 mS into the FOR NEXT loop, then a 50/50 (?) chance myButtons.BIT0 is low during the last iteration?
I'm sure there is a good reason, what am I missing?

myButtons    VAR    Word
startProc    VAR    myButtons.BIT0
endProc      VAR    myButtons.BIT1
idx          VAR    Nib  

Main:
  GOSUB Get_Buttons
  IF (startProc = 1) THEN
    ' start the process
  ENDIF
  IF (endProc = 1) THEN
    ' end the process
  ENDIF
  GOTO Main

Get_Buttons:
  myButtons = %111111111111         ' assume pressed
  FOR idx = 1 TO 10                 ' test 10 times 
    myButtons = myButtons & INS     ' validate inputs
    PAUSE 5
  NEXT
  RETURN

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Ken

Jon Williams

You get better deboucing by doing several short checks than one long one -- my opinion and how I do it.· Yes, one could simplify by doing this:

Get_Buttons:
· myButtons =·%111111111111 & INS
· PAUSE 50
· myButtons = myButtons & INS
· RETURN

The *problem* (for me) is that PAUSE 50 is a long time and you can't see any contact bouncing during that period.· So, if you happen catch an input that has bounced closed during the two points you scanned, it looks like a valid input -- when it may not be.· By scanning more times you have more opportunites to detect an input that hasn't actually settled.

In my original routine the loop will cause a 0 to be returned for any contact that bounces open at any time during the cycle.· Why?· Because anything ANDed with zero is 0 -- and that 0 will stick through the end of the loop.· If you look at it carefully, the subroutine will only return a 1 for a given input if the pin is high on entry to the subroutine and remains in that state all the way through it.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Jon Williams
Applications Engineer, Parallax

Post Edited (Jon Williams (Parallax)) : 7/14/2005 2:54:48 PM GMT

KenM

Thank you for the explanation.



▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Ken