Help: How not to fry a pocket watch B
Archiver
Posts: 46,084
Having fried my second DS1302 clock/calendar chip, I'm trying out a pocket
watch B. As these cost lots more than a DS1302, I'm leery of doing any
chancy experiments with it, so I'm looking for help.
The goal is to have a backup battery for the pocket watch (hereinafter, the
PW), but NOT for the BS2 or other components. (I want days of backup power
for the clock, and I won't get that if the BS2, etc. is running on the same
capacitor.)
One possibility is to put a diode between +5 and the PW power pin, and put
the super cap behind that diode, too. Then the +5v will charge the cap when
available; but if +5v fails, the cap will keep the PW going (but nothing
else, because of the diode.) Of course, if there's 0.7v of voltage drop
across the diode, the PW will be running at 4.3v, not 5v. The doc says
that's ok...
But is it? The BS2 would be running at 5v and talking (via SERIN/SEROUT) to
the PW. The PW would be at 4.3v. Doesn't that amount to 0.7v of overvoltage
on the PW pins, and a fried PW? Can I fix that by putting diodes on the
BS2->PW signal path, to give it the same voltage drop? Or maybe just put the
BS2, etc. behind its own diode, and run it at 4.3v too...
And then there's what happens if the power goes off, and then on. At that
point, the cap might be drained pretty low, and it won't come up to 5v
immediately. But the BS2 (at 5v) will come right up and start talking to the
PW, which might still be at around 2v or so. Won't that kill the PW? (I
think this is the scenario that killed the DS1302's).
I'm certain people have solved this problem. Battery backup of a clock is a
pretty common scenario. How is it done?
watch B. As these cost lots more than a DS1302, I'm leery of doing any
chancy experiments with it, so I'm looking for help.
The goal is to have a backup battery for the pocket watch (hereinafter, the
PW), but NOT for the BS2 or other components. (I want days of backup power
for the clock, and I won't get that if the BS2, etc. is running on the same
capacitor.)
One possibility is to put a diode between +5 and the PW power pin, and put
the super cap behind that diode, too. Then the +5v will charge the cap when
available; but if +5v fails, the cap will keep the PW going (but nothing
else, because of the diode.) Of course, if there's 0.7v of voltage drop
across the diode, the PW will be running at 4.3v, not 5v. The doc says
that's ok...
But is it? The BS2 would be running at 5v and talking (via SERIN/SEROUT) to
the PW. The PW would be at 4.3v. Doesn't that amount to 0.7v of overvoltage
on the PW pins, and a fried PW? Can I fix that by putting diodes on the
BS2->PW signal path, to give it the same voltage drop? Or maybe just put the
BS2, etc. behind its own diode, and run it at 4.3v too...
And then there's what happens if the power goes off, and then on. At that
point, the cap might be drained pretty low, and it won't come up to 5v
immediately. But the BS2 (at 5v) will come right up and start talking to the
PW, which might still be at around 2v or so. Won't that kill the PW? (I
think this is the scenario that killed the DS1302's).
I'm certain people have solved this problem. Battery backup of a clock is a
pretty common scenario. How is it done?