| View previous topic :: View next topic |
| Author |
Message |
scottc
Joined: 16 Aug 2010
Posts: 95
|
|
 Posted: Tue Oct 26, 2010 4:51 pm |
 |
|
PCM, Thats an excellent explanation,
I tested the routine in a TIMER0 ISR and it appears to work fine also.
I'm using a 18f part so timer0 is 16bits
Here is how I implementated it.
| Code: |
#int_timer0
void Push_Button(void) //Function increments Count by 1 regardless
{ //of how long Enc_SW is pressed in
set_timer0(0);
static int1 pushed = false;
if (!pushed && !Enc_Sw) // Switch is Pressed / down
{
pushed = true;
while ( get_timer0() <53036) ; //5MS Interrupt
if (++count == 4) count = 1; // Increment the count
} else if (Enc_Sw) // Switch is Released / up
pushed = false;
}
void main(void)
{
setup_timer_0 (RTCC_DIV_1|RTCC_INTERNAL);
CLEAR_INTERRUPT (INT_TIMER0);
ENABLE_INTERRUPTS (GLOBAL);
ENABLE_INTERRUPTS (INT_TIMER0);
SET_TIMER0(0);
}
|
I believe its correct, but comments welcome if it is not.
Thanks Scott |
|
 |
pmuldoon
Joined: 26 Sep 2003
Posts: 218
Location: Northern Indiana
|
|
| Posted: Thu Oct 28, 2010 7:57 am |
|
|
| Are you really intending to spend 5 out of 6.4mS in the interrupt routine? |
|
|
scottc
Joined: 16 Aug 2010
Posts: 95
|
|
| Posted: Thu Oct 28, 2010 10:41 am |
|
|
Intent was to spend about 5ms in the isr as a pseudo debounce for the
pushbutton switch.
Thanks Scott |
|
|
pmuldoon
Joined: 26 Sep 2003
Posts: 218
Location: Northern Indiana
|
|
| Posted: Thu Oct 28, 2010 11:21 am |
|
|
It works, but just a bit unusual since your PIC will only have 14/64ths of its potential processing time available to do anything else.
Here is yet another debounce technique. I had an existing piece of hardware that had 2 pushbuttons (SELECT, ENTER) and a 2-line LCD for an interface.
I needed to make the buttons responsive & debounced, yet I wanted a way to keep a careless user from accidentally getting into the setup/config
stuff. I detect a button press quickly, and can also detect a button held for about 2-secs (to access setup screens). The interrupt routine sets the flags that tell the rest of the program that a button was pushed. It is the responsibility of the routine responding to the button press to reset the flag.
I'm just showing the SELECT button to keep it simple.
It's running with a 4MHz crystal and no divisor on the Timer_1 setup, so each count is 1 uS.
| Code: |
#define PB_SELECT PIN_D6 // SELECT pushbutton
#define DEBOUNCE_TIME 3 // # of (10mS) interupt cycles Pushbutton must be held for before validating
#define LONG_PRESS_TIME 200 // number of 10mS interrupts before keypress is considered to be a long press (up to 254)
// Global Flags set by interrupt when button press detected, cleared by responding routing
int1 PBselectHit=0; // Flag indicating PB has been detected and debounced.
int1 PBselectLongPress=0; // indicates a LONG_PRESS_TIME has expired
/********************* Timer1 Interrupt Handler *************************************/
#int_timer1
void Timer1Handler()
{
static unsigned int SelectActive=0,
set_timer1(get_timer1()-10000); // set the timer to fire again in 10mS
// pushbutton debounce SELECT
if(input(PB_SELECT))
{
if(SelectActive++ == DEBOUNCE_TIME) // button must be held continuously for this many interrupt cycles
PBselectHit = TRUE; // flag must be reset by function waiting for keypress
if(SelectActive == LONG_PRESS_TIME)
PBselectLongPress = TRUE; // flag must be reset by function waiting for keypress
if(SelectActive > LONG_PRESS_TIME) // long press detected
SelectActive = LONG_PRESS_TIME+1; // clamp counter
}
else
{
SelectActive = 0; // otherwise reset debounce counter and start over
}
}
|
|
|
|
scottc
Joined: 16 Aug 2010
Posts: 95
|
|
| Posted: Thu Oct 28, 2010 6:52 pm |
|
|
Thanks for sharing your example pmuldoon, interesting code.
I am using my routine with a pic18f2331 running at 10Mhz, I did not
notice any problems with the pic slowing down. The code may not be optimal but does appear to work good.
I get in and out of the button routine fast. Each press of the button will increment the count variable. I test the variable and depending on its value jump to a rotary encoder routine and display another value on
a LCD running in 4bit interface mode. Similar to say using a car radio
and tuning a station.
The actual lcd routine running in 4bit mode, 20chr x 2 line is running at
1.25 Ms for each write to the display. Tests so far seem fairly fluid.
Hit the button, jump to a menu, Spin the encoder, write to the display.
The only ISR in the code is the push button routine, so all in all not too
bad.
Got to just love the pic processor man, awesome little chip 
and lets not forget the ccs compiler thats awesome too.
Thanks Scott |
|
|
MassaM
Joined: 08 Jun 2019
Posts: 31
|
|
| Posted: Sun Jul 14, 2019 6:01 pm |
|
|
| pmuldoon wrote: |
Here is yet another debounce technique. I had an existing piece of hardware that had 2 pushbuttons (SELECT, ENTER) and a 2-line LCD for an interface.
I needed to make the buttons responsive & debounced, yet I wanted a way to keep a careless user from accidentally getting into the setup/config
stuff. I detect a button press quickly, and can also detect a button held for about 2-secs (to access setup screens). The interrupt routine sets the flags that tell the rest of the program that a button was pushed. It is the responsibility of the routine responding to the button press to reset the flag.
|
Hello pmuldoon,
Apologies as I don't intend to hijack this discussion thread, but your code is helping me in my project and may I ask your help on this thread please?
https://www.ccsinfo.com/forum/viewtopic.php?p=224938#224938
Appreciated!
_________________
while(!dead)
{
keepLearning();
} |
|
|
|
FAQ 
Search
Register
Profile 
Log in to check your private messages 
Log in
