> Hello :-)
>
> I've been reading this list via the archive on www.infosite.com for a few
weeks
> now.  It's been VERY USEFUL !
>
> Does anyone have experience of getting wrong values from the lower byte of
the
> timer 1 counter on a 16F877 ?
>
> I'm clocking tmr1 at 3.5 - 3.6Mhz   (Those with a Ham Radio background
will
> quickly realise what I'm trying to do!)
> The processor clock is 3.2768Mhz and I'm genterating 100 interrupts a
second.
>
> So Timer1 should advance by 350000 to 360000 between each Tmr0 interrupt.
> To get around the problems of reading the high and low bytes I used this
method
>
> Step1 Read High byte
> Step2 Read Low byte
> Step3 Reread High byte
>
> If the two high bytes are the same then low byte did not overflow while
> reading the values.
>
> If the two high bytes are different then look at the low byte to decide if
> the overflow happened between Steps 1 and 2  or  Steps 2 and 3.
>
> If lowbyte > 0x80 then the overflow happened after reading the low byte so
>  use the second high byte value
>
> If lowbyte <= 0x80 then the overflow happened before reading the low byte
so
>  use first high byte value
>
> This seemed to work as expected except once in a while it produced the
wrong
> result.
>
> Note I can't afford to do the normal method of rereading the low value
again if
> the high bytes differ as I'm trying to measure the frequency as accuratly
as I
> can.  Timer 1 is being clocked at about 4 times the instruction rate so
going
> back to reread the values will introduce an unacceptable error.
>
>
> It seems that sometimes the value read from the timer low byte was wrong!
> How did I come to this conclusion ?  Well I have a DAC on PORTB and I just
> replace my main loop with on the read the low byte and writes it to the
DAC.
>
> I expect to see a nice ramp on the scope increasing by about 4 DAC steps
between
> each sample.  But every so often the value skips by a much larger value
(And
> yes I had stopped the 100Hz interrups when I tested this).
>
> In the end I found another way to get an accurate count (*) but this does
seem
> like a hardware bug to me.
>

Hi

It wasn't clear to me: Are you using timer 1 as an assynchronous counter
????

I have used many times the timer1 module of the F872 and F876 and never saw
this (in fact, only synchronous mode !)

Fabio Pereira

--
http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads