Post by thread:[PIC]: RB0 Interrupt Edge Triggering

Hi, I am developing a simple hardware delay device on a 16F876 using MPLAB and Hitech C. An overkill I know. The code configures the ports, timers and interrupts as required, - RB0 as input, RB1-7 as outputs - interrupt on RB0 falling edge - timer interrupt every 40 ms. (PAL video frame rate) It looks for a falling edge on RB0 which generates an interrupt, the interrupt routine starts a software counter. The timer takes over incrementing the counter. When the counter reaches the required setting it drives an output low, RB1, and starts another software counter, which when done resets the output to high. There are two problems - 1. The INTF interrupt is generated on any edge, rising or falling, when it should only be on the falling edge. 2. Clearing the INTE doesn't disable the interrupt. I am trying to disable the interrupt after the initial falling edge and then continue to process the timer interrupt. Then later re-enable the RB0 interrupt. I searched the archive but couldn't find anything relevant. Can anyone help? Is there a know issue with this device and edge triggering? Relevant source code below. // 7 Cue Delay.c // pic 16F876 #include <pic.h> #define PIC_CLK 4000000 #include "7cue_dly.h" __CONFIG(FOSC0 | UNPROTECT ); #define LED RA0 #define STBY_IN RB0 #define CUE_OUT RB1 #define STANDBY_LED RB6 #define CUE_LED RB7 #define ON 1 #define OFF 0 static unsigned int cCueDelayCnt; static unsigned char cStdbyRx; static unsigned char cCue; main() { TRISB = 0b00000001; // Set RB1 as i/p and RB2 ->7 as o/ps OPTION = 0b00000000; // Set Option reg for weak pullup, on falling edge of RB0 etc // OPTION = 0b01000000; // Set Option reg for weak pullup, INT on rising edge of RB0 etc cCueDelayCnt = 0; cStdbyRx = OFF; cCue = OFF; CUE_OUT = ON; CUE_LED = OFF; STANDBY_LED = OFF; init_interrupts(); while(1){ // Let interrupt handler do everything } } void interrupt handler() { di(); if(INTF == 1){ cCueDelayCnt = 0; cStdbyRx = ON; STANDBY_LED = ON; cCue = OFF; INTE = 0; // Disable RB0 int INTF = 0; // clear flag // INTE = 0; // Disable RB0 int } if(TMR1IF == 1){ if(cCue == ON){ cCueDelayCnt = cCueDelayCnt + 1; if(cCueDelayCnt == 12){ // 150 frames = 6 seconds // 25 frames = 1 second CUE_OUT = ON; cCueDelayCnt = 0; cCue = OFF; CUE_LED = OFF; STANDBY_LED = OFF; INTE = 1; // Enable RB0 int INTF = 0; // clear flag } } if(cStdbyRx == ON){ cCueDelayCnt = cCueDelayCnt + 1; if(cCueDelayCnt == 37){ // 600 frames = 24 seconds // 37 frames = 1s 12f CUE_OUT = OFF; cStdbyRx = OFF; cCueDelayCnt = 0; cCue = ON; CUE_LED = ON; } } TMR1L = 0x77; // load 60535 TMR1H = 0xec; TMR1IF = 0; // clear flag } ei(); } void init_interrupts() { // timer 1 for 40mSec (1 Video frame) tick // 4Mhz/4 Clock with prescale of 1:8 with Timer 1 preload of 60,535 , gives 40mSec tick // Enable RB0 interrupt for falling edge INTCON = 0; // disable all ints PIE1 = 0; PIE2 = 0; T1CON = 0b00110001; // 1:8 Prescaler, internal clock and enable Timer 1 TMR1L = 0x77; // load 60535 TMR1H = 0xec; TMR1IF = 0; // clear flag INTF = 0; // clear RB0 interrupt flag TMR1IE = 1; // Enable Timer 1 int INTE = 1; // Enable RB0 int PEIE = 1; // Enable Peripheral int ei(); // enable ints } Thanks for your help. Regards, Graham Cheetham spam_OUTgcheethamTakeThisOuTsoutherncrossqld.com.au Regards, Graham Cheetham Software & Systems Engineer Southern Cross Ten Queensland Townsville Office 07 4726 2074 Switch 07 4726 2000 Fax 07 4726 2109 .....gcheethamKILLspam@spam@southerncrossqld.com.au -- http://www.piclist.com hint: PICList Posts must start with ONE topic: [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads

I don't know about chip problems, but is it possible you are getting ringing or noise on your interrupt line? That could explain at least part of the problem. (but you should still be able to disable the interrupt) Try adding a rc filter if there's nothing there at the moment - especially if the line is long or subject to noise injection etc. RP Graham Cheetham <gcheetham@SOUTHERNCROSSQ To: PICLISTKILLspamMITVMA.MIT.EDU LD.COM.AU> cc: Sent by: pic Subject: [PIC]: RB0 Interrupt Edge Triggering microcontroller discussion list <.....PICLISTKILLspam.....MITVMA.MIT.EDU> 21/08/03 11:16 Please respond to pic microcontroller discussion list Hi, I am developing a simple hardware delay device on a 16F876 using MPLAB and Hitech C. An overkill I know. The code configures the ports, timers and interrupts as required, - RB0 as input, RB1-7 as outputs - interrupt on RB0 falling edge - timer interrupt every 40 ms. (PAL video frame rate) It looks for a falling edge on RB0 which generates an interrupt, the interrupt routine starts a software counter. The timer takes over incrementing the counter. When the counter reaches the required setting it drives an output low, RB1, and starts another software counter, which when done resets the output to high. There are two problems - 1. The INTF interrupt is generated on any edge, rising or falling, when it should only be on the falling edge. 2. Clearing the INTE doesn't disable the interrupt. I am trying to disable the interrupt after the initial falling edge and then continue to process the timer interrupt. Then later re-enable the RB0 interrupt. I searched the archive but couldn't find anything relevant. Can anyone help? Is there a know issue with this device and edge triggering? Relevant source code below. // 7 Cue Delay.c // pic 16F876 #include <pic.h> #define PIC_CLK 4000000 #include "7cue_dly.h" __CONFIG(FOSC0 | UNPROTECT ); #define LED RA0 #define STBY_IN RB0 #define CUE_OUT RB1 #define STANDBY_LED RB6 #define CUE_LED RB7 #define ON 1 #define OFF 0 static unsigned int cCueDelayCnt; static unsigned char cStdbyRx; static unsigned char cCue; main() { TRISB = 0b00000001; // Set RB1 as i/p and RB2 ->7 as o/ps OPTION = 0b00000000; // Set Option reg for weak pullup, on falling edge of RB0 etc // OPTION = 0b01000000; // Set Option reg for weak pullup, INT on rising edge of RB0 etc cCueDelayCnt = 0; cStdbyRx = OFF; cCue = OFF; CUE_OUT = ON; CUE_LED = OFF; STANDBY_LED = OFF; init_interrupts(); while(1){ // Let interrupt handler do everything } } void interrupt handler() { di(); if(INTF == 1){ cCueDelayCnt = 0; cStdbyRx = ON; STANDBY_LED = ON; cCue = OFF; INTE = 0; // Disable RB0 int INTF = 0; // clear flag // INTE = 0; // Disable RB0 int } if(TMR1IF == 1){ if(cCue == ON){ cCueDelayCnt = cCueDelayCnt + 1; if(cCueDelayCnt == 12){ // 150 frames = 6 seconds // 25 frames = 1 second CUE_OUT = ON; cCueDelayCnt = 0; cCue = OFF; CUE_LED = OFF; STANDBY_LED = OFF; INTE = 1; // Enable RB0 int INTF = 0; // clear flag } } if(cStdbyRx == ON){ cCueDelayCnt = cCueDelayCnt + 1; if(cCueDelayCnt == 37){ // 600 frames = 24 seconds // 37 frames = 1s 12f CUE_OUT = OFF; cStdbyRx = OFF; cCueDelayCnt = 0; cCue = ON; CUE_LED = ON; } } TMR1L = 0x77; // load 60535 TMR1H = 0xec; TMR1IF = 0; // clear flag } ei(); } void init_interrupts() { // timer 1 for 40mSec (1 Video frame) tick // 4Mhz/4 Clock with prescale of 1:8 with Timer 1 preload of 60,535 , gives 40mSec tick // Enable RB0 interrupt for falling edge INTCON = 0; // disable all ints PIE1 = 0; PIE2 = 0; T1CON = 0b00110001; // 1:8 Prescaler, internal clock and enable Timer 1 TMR1L = 0x77; // load 60535 TMR1H = 0xec; TMR1IF = 0; // clear flag INTF = 0; // clear RB0 interrupt flag TMR1IE = 1; // Enable Timer 1 int INTE = 1; // Enable RB0 int PEIE = 1; // Enable Peripheral int ei(); // enable ints } Thanks for your help. Regards, Graham Cheetham EraseMEgcheethamspam_OUTTakeThisOuTsoutherncrossqld.com.au Regards, Graham Cheetham Software & Systems Engineer Southern Cross Ten Queensland Townsville Office 07 4726 2074 Switch 07 4726 2000 Fax 07 4726 2109 gcheethamspam_OUTsoutherncrossqld.com.au -- http://www.piclist.com hint: PICList Posts must start with ONE topic: [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads -- http://www.piclist.com hint: PICList Posts must start with ONE topic: [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads

> > There are two problems - > > 1. The INTF interrupt is generated on any edge, rising or > falling, when it should only be on the falling edge. > > 2. Clearing the INTE doesn't disable the interrupt. There has been a suggestion of glitching on the RB0 input. But I thought that after the first interrupt, (glitching possible during manual testing), the RB0 interrupt should be disabled until after the two timeouts periods, where it is re-enabled. > > __CONFIG(FOSC0 | UNPROTECT ); > > The default state of the watchdog timer is "on." You're not clearing > it in your code, so you're probably resetting frequently. The assembly code listing from the compiler shows the WDTE bit set to zero, which is off. > > void interrupt handler() > > { > > > > di(); > > ei(); > > Don't manipulate the global interrupt flag in the interrupt handler. I'll give that a try removing these. > > > if(INTF == 1){ > > If you're disabling INTE, that doesn't prevent INTF from being set. > Your test should be > > if (INTE && INTF) I still don't understand that if INTE is off (set to zero) how an interrupt can be generated. Thanks. Graham Cheetham -- http://www.piclist.com hint: PICList Posts must start with ONE topic: [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads

> On Thu, 21 Aug 2003, Graham Cheetham wrote: > > > > If you're disabling INTE, that doesn't prevent INTF from > being set. > > > Your test should be > > > > > > if (INTE && INTF) > > > > I still don't understand that if INTE is off (set to zero) how an > > interrupt can be generated. > > As the datasheets says, "Individual interrupt flag bits are set, > regardless of the status of their corresponding mask bit, or the GIE > bit." > > An interrupt is not being generated, but your software is testing the > INTF flag, regardless of the reason that it is executing the ISR. > I.e., when your timer interrupt pops, your test for INTF will be true > if there has been an RB0 edge, regardless of the state of INTE. > Thank you - it is making sense now. I will go and redo the code. Graham Cheetham -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics

> > If you're disabling INTE, that doesn't prevent INTF from being set. > > Your test should be > > > > if (INTE && INTF) > >I still don't understand that if INTE is off (set to zero) how an interrupt >can be generated. It isn't; your periodic interrupt occuring, and during the course of testing for that interrupt, the INTF flag is spotted, causing the machine to inadvertantly go of course and service the INT interrupt. Or at least, that's what would be happening if this were written in assembler. I don't know if the C compiler is smart enough to catch that little snafu. Mike H. _________________________________________________________________ <b>Help protect your PC:</b> Get a free online virus scan at McAfee.com. http://clinic.mcafee.com/clinic/ibuy/campaign.asp?cid=3963 -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics