> > =A0I intend to output the data on the serial at about 9600/19200,
> > would you think an RC clock or internal clcok would be accurate
> > enough for this simple task?

> Some newer PICs have internal clock accurate enough to get away with it i=
n
> some cases when you know the other end is right on. =A0Otherwise, this is=
 not
> a good idea. =A0The early PIC internal RC oscillators are too inaccurate =
for
> this.
>
> Look up the accuracy of the internal oscillator in the datasheet for the
> operating range you will be using.

Guide.

As Olin says, older PICs were generally not stable enough and as he
and Harold say, newer ones often are. Appreciating what the
requirement is and how to tell from the spec sheet is useful.

Worst case your timing is allowed to be out by half a bit time if it
is synchronised in the middle of the start bit at the start of
transmission.
Close enough, to avoid missampling  you are allowed a drift of 1/2 bit
in (1/2 start + N data bits + stop bit) =3D 1/2 /(9.5) for 8 bits + 1
start + 1 stop ~=3D 5%
You can argue over whether you can save 1/2 a bit if the drift is
towards missampling the stop bit late etc but that's close enough.
Given that worst case the transmit clock can drift one way and the
receive clock the other, the worst case where you can guarantee
accuracy is 5%/2 =3D +/- 2.5%

You can decide from the spec sheet how much error you can expect.
Drift across full temperature range will usually be greater than
across limited temperature range and if both your devices are in the
same thermal environment then the relative drift will probably be
less.

With older parts and restricted environments eg office temperature
range and both devices in same thermal environment., asynchronous coms
would probably work with RC clocks.

Grabbing a random data sheet - 16F877, dated 2008 they specify

+/-1% AT 25C, 3.5 V Vdd
+/- 2% 0-85C and 2.5 - 5.5 Volt Vdd (almost all applications)
+/- 5% -25 to +125C and 2 - 5.5 V Vdd. (overnight in snow to auto
firewall mounting)

ie it will work OK for 1st and 2nd ranges and will very probably fail
at the extremes of the 3rd range unless devices are tortured
similarly.



               Russell

--=20
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
.