Post by thread:[PIC]: Intrpt Bounce

Hi all, I have a reed relay on the RB0 Interrupt. It's very slow, say a second on, and a few seconds off - to ground. RB0 is made to detect the interrupt on the falling edge, that's OK. A falling edge - jump to ISR - wait there if necessary till reed stops bouncing, then exit ISR. Problem occurs when reed opens again, a second or so later. The contact bounce creates a couple of rising and falling edges again, the latter of which trigger interrupts. Can't put delays in, because reed on and off times vary considerably - seconds to many, many, seconds, even minutes. I have used a 555 to create a one shot which works OK, just wondering if there was another way. Schmitt trigers capture the bounces too. Regards Howard Simpson. -- http://www.piclist.com hint: To leave the PICList spam_OUTpiclist-unsubscribe-requestTakeThisOuTmitvma.mit.edu>

> Problem occurs when reed opens again, a second or so later. > The contact bounce creates a couple of rising and falling edges > again, the latter of which trigger interrupts Put a trap at the start of the ISR to test the state of the b0 pin. Interrupts are turned off on entry to the ISR, so further transitions on b0 will have no effect. If you don't want to run the ISR code then simply wait until the bouncing has stopped and then jump to the retfie, which will re-enable INTE for the next legitimate closure. Make sure to clear INTF before leaving the ISR as the bouncing will set this flag, causing immediate re-entry into the ISR as soon as retfie is executed and interrupts are re-enabled Alternatively a simple RC filter should get rid of most of the noise so that b0 sees only one transition -- http://www.piclist.com hint: To leave the PICList .....piclist-unsubscribe-requestKILLspam@spam@mitvma.mit.edu>

{Quote hidden}> -----Original Message----- > From: Howard Simpson [SMTP:hjsbraKILLspamBIGPOND.COM] > Sent: Monday, February 10, 2003 11:53 AM > To: .....PICLISTKILLspam.....MITVMA.MIT.EDU > Subject: [PIC]: Intrpt Bounce > > Hi all, > I have a reed relay on the RB0 Interrupt. It's very slow, say a second > on, and a few seconds off - to ground. > RB0 is made to detect the interrupt on the falling edge, that's OK. A > falling edge - jump to ISR - wait there if necessary till reed stops > bouncing, then exit ISR. > Problem occurs when reed opens again, a second or so later. The contact > bounce creates a couple of rising and falling edges again, the latter of > which trigger interrupts. > Can't put delays in, because reed on and off times vary considerably - > seconds to many, many, seconds, even minutes. > I have used a 555 to create a one shot which works OK, just wondering if > there was another way. Schmitt trigers capture the bounces too. > Regards > Howard Simpson. > Well, if the maximum switching speed of the relay is of the order of a second or greater I see no reason why you could not use a delay effectively. One possible method is to change the edge on which you interrupt, so the interrupt service always catches the first edge and can then wait until all bouncing has finished: e.g. if RBIE = 1 if INTEDG = 0 // falling edge // do your stuff here // debounce routine here INTEDG = 1 // check for rising edge next time else // debounce routine here INTEDG = 0 // check for falling edge next time end if RBIE = 0 // clear interrupt end if Cheers Mike -- http://www.piclist.com hint: To leave the PICList EraseMEpiclist-unsubscribe-requestspam_OUTTakeThisOuTmitvma.mit.edu>

Howard Simpson wrote: > > Hi all, > I have a reed relay on the RB0 Interrupt. It's very slow, say a second > on, and a few seconds off - to ground. > RB0 is made to detect the interrupt on the falling edge, that's OK. > Problem occurs when reed opens again, a second or so later. The contact > bounce creates a couple of rising and falling edges again, the latter of > which trigger interrupts. Force a timed delay period of 0.2 seconds etc where the contacts MUST be open. Only after this "safe" period is the system enabled ready for the FIRST falling edge. -Roman -- http://www.piclist.com hint: To leave the PICList piclist-unsubscribe-requestspam_OUTmitvma.mit.edu>

> I have a reed relay on the RB0 Interrupt. It's very slow, say a second > on, and a few seconds off - to ground. > RB0 is made to detect the interrupt on the falling edge, that's OK. A > falling edge - jump to ISR - wait there if necessary till reed stops > bouncing, then exit ISR. > Problem occurs when reed opens again, a second or so later. The contact > bounce creates a couple of rising and falling edges again, the latter of > which trigger interrupts. > Can't put delays in, because reed on and off times vary considerably - > seconds to many, many, seconds, even minutes. > I have used a 555 to create a one shot which works OK, just wondering if > there was another way. Schmitt trigers capture the bounces too. Configure OPTION_REG to interrupt on the next expected edge. On a real edge interrupt, disable RB0 interrupts completely. Clear timer 0 and enable timer 0 interrupts. With the prescaler you can get a delay up to 13mS even with a 20MHz oscillator. Now process the interrupt and return normally. Any new bounces will be ignored because the RB0 interrupt is disabled. On the timer 0 interrupt, disable the timer 0 interrupt, flip the RB0 edge appropriately, clear any pending RB0 interrupt condition, then enable RB0 interrupts again. In other words, this scheme ignores any new reed switch edges for a debounce blockout period after the first edge. Another scheme I usually use for mechanical switches is to poll every millisecond or so from an interrupt. The interrupt routine sets the global flag bit to indicate the switch state only when it's been the same for 50mS in a row. The foreground event loop then handles the switch state. ***************************************************************** Embed Inc, embedded system specialists in Littleton Massachusetts (978) 742-9014, http://www.embedinc.com -- http://www.piclist.com hint: To leave the PICList @spam@piclist-unsubscribe-requestKILLspammitvma.mit.edu>

Howard Simpson wrote: > > Hi all, > I have a reed relay on the RB0 Interrupt. It's very slow, say a second > on, and a few seconds off - to ground. > RB0 is made to detect the interrupt on the falling edge, that's OK. A > falling edge - jump to ISR - wait there if necessary till reed stops > bouncing, then exit ISR. > Problem occurs when reed opens again, a second or so later. The contact > bounce creates a couple of rising and falling edges again, the latter of > which trigger interrupts. > Can't put delays in, because reed on and off times vary considerably - > seconds to many, many, seconds, even minutes. > I have used a 555 to create a one shot which works OK, just wondering if > there was another way. Schmitt trigers capture the bounces too. > Regards > Howard Simpson. Thanks to repliers. Hmmmm. I see the trick is to determinie how long the unwanted trigger lasts, and allow for it. Unwanted trigger from bounces will be short, and wanted ones, even the shortest ones, will be much, much longer. One thing I didn't mention that, at some point, the reed may be closed for a long time, even days, but that should not matter as RB0 triggers on edges only. (falling in this case) So basically, if there is a falling edge, the input must stay low for a period longer than a bounce, for the edge to be assumed valid. Howard Simpson -- http://www.piclist.com hint: PICList Posts must start with ONE topic: [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads

> Hmmmm. I see the trick is to determinie how long the unwanted > trigger lasts, and allow for it One trick that may work is to start a counter on receipt of the first transition. You set the counter to increment until it reaches a certain value, but only if there are no pin readings opposite to the one you're looking for, otherwise it resets to 0 and starts again. If the counter reaches the certain value you've set then it's safe to assume that the bounce noise is finished. For instance, detecting a 0 to 1 switch closure. You have a delay loop that takes 1ms. Read the pin, if 1 then increment the counter. If 0, then clear the counter, and so on. When the counter reaches 50 you know that 50ms has elapsed with no 0's on the pin, even though it may have been longer than 50ms since the first 1 that started it all -- http://www.piclist.com hint: PICList Posts must start with ONE topic: [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads