> Paul, sorry to ask you this, but how can you get your /MCLR input
> down to a secure level of 0.15 Vdd (as stated as the max level for a
> clear "0" on the /MCLR pin) by just having an RC network? As you
> describe it, Vdd is reduced to 3V and that is what you get at the
> /MCLR pin as well. It doesn't help if the 3V level is delayed an
> hour, it won't go below 0.15 Vdd or ...?
> With the DS1231 you will get a /MCLR tied down to GND as soon as the
> input voltage drops to 4.5V or 4.0V (settable) and that level will be
> held until the voltage is stable above the 4.0 or 4.5V limit. It also
> has a small extra delay to allow any eventual crystals to start up
> sufficiently.
> Wouldn't this solve your problem? At any level, rise time and so on
> you will be granted a secure power-on-reset.
> As mentioned in other replies Maxim has lots of uP watchdogs. The
> reason I use the DS1231 from Dallas is that I have quite the opposite
> experinces with Dallas as the person who suggested Maxim. Maxim makes
> a lot of good stuff but mainly they just make a LOT of stuff. In the
> last three years they have produced 10.7 cm of just "New Releases
> Data Book"s (measured, not estimated nor exaggerated). I have the
> continous fear that they just will unsupport a device without prior
> notice and then you have to change your construction to use another
> circuit.
>
> /Tomas

You know, you're right - when I have a "brown-out", my reset line does
not actually reset since my Vdd line doesn't go to 0V.  I, however,
found the line in the databook that said Vdd must start from 0V for a
proper bring-up and stopped looking at the problem.  I guess that, I
MAY be able to ignore the line in the databook and use the reset pin
and the DS1231.  I'll give it a try - testing - testing - testing.....
Thanks...!!!

--

            -- Paul Greenwood --  (pablo@austin.ibm.com)

All true wisdom is found on T-shirts.