Post by thread:Timer1 interrupts interfere with LCD

More playing about with timers. I was using timer1 with a 32768 xtal, interrupting every second. Xtal was running slow. So I decided to try using timer1 driven from the internal osc (20mHz xtal on a '877). Figured out how to load the timer in the interrupt routine without screwing up the time interval. Now I get one interrupt every 10 milliseconds, and can keep track of time of day. My problem: I'm displaying the time on an LCD - the code uses bit-banging to do the output. (I'm using PicBasic Pro, so I didn't code the LCD stuff myself). Every once in a while the display hiccups - presumably because the timer interrupt routine went off in the middle of sending data to the LCD, and disrupted the timing. Same thing happens if I send the output via bit-banging serial output (again in PicBasic Pro) I presume that I could disable interrupts before doing the I/O, and turn them on after. In this case, I think that the timer1 counter would contain a number equal to the number of instruction cycles taken by the LCD routine. Since I have to preload the counter in the interrupt routine anyhow, in order to get an interrupt every 10 msecs, I think I can compute a new preload based on the existing counter, so the next timer cycle will be shorter, to compensate for the time lost in the LCD routine. This is my first time experimenting with timers and interrupts. Is what I'm proposing a good way to do this? Anyone got any better ideas? Thanks Larry Bradley Orleans (Ottawa), Ontario, CANADA -- http://www.piclist.com hint: To leave the PICList spam_OUTpiclist-unsubscribe-requestTakeThisOuTmitvma.mit.edu>

I had a similar problem - turned out to be occasionally interrupting when it was in page bank 1 - took me ages to work it out and debug it. Just save the current page at the start of the interrupt and put the pic back into whatever page it was in when the interrupt occurred at the end of the int. routine. Dont forget to pop and push the W register (and carry etc) before returning to the point where the interrupt occurred. May not be the right answer but worth looking into. Dom {Original Message removed}

In my interrupt routine, I'm saving W, STATUS and PCLATH, and restoring them at the end. I thought that was all I needed to save. But I'm no expert yet. Larry At 09:19 PM 12/17/2002 +0000, you wrote: {Quote hidden}>I had a similar problem - turned out to be occasionally interrupting when it >was in page bank 1 - took me ages to work it out and debug it. Just save the >current page at the start of the interrupt and put the pic back into >whatever page it was in when the interrupt occurred at the end of the int. >routine. > >Dont forget to pop and push the W register (and carry etc) before returning >to the point where the interrupt occurred. > >May not be the right answer but worth looking into. > >Dom > Larry Bradley Orleans (Ottawa), Ontario, CANADA -- http://www.piclist.com hint: To leave the PICList .....piclist-unsubscribe-requestKILLspam@spam@mitvma.mit.edu>

"Larry Bradley" <lhbradleyKILLspamIGS.NET> wrote: > I presume that I could disable interrupts before doing the I/O, > and turn them on after. In this case, I think that the timer1 > counter would contain a number equal to the number of instruction > cycles taken by the LCD routine. Since I have to preload the > counter in the interrupt routine anyhow, in order to get an > interrupt every 10 msecs, I think I can compute a new preload based > on the existing counter, so the next timer cycle will be shorter, > to compensate for the time lost in the LCD routine. > > This is my first time experimenting with timers and interrupts. Is > what I'm proposing a good way to do this? Yes, that's a good way. In practice, you "compute a new preload based on the existing counter" by replacing your current MOVWF TIMER1 with ADDWF TIMER1. -Andy === Andrew Warren -- .....aiwKILLspam.....cypress.com === Principal Design Engineer === Cypress Semiconductor Corporation === === Opinions expressed above do not === necessarily represent those of === Cypress Semiconductor Corporation -- http://www.piclist.com hint: To leave the PICList EraseMEpiclist-unsubscribe-requestspam_OUTTakeThisOuTmitvma.mit.edu>