-=> Jasen Betts said to Roy J. Tellason
 -=> about ""FREEZER TOO HOT" ALARM" on 09-06-04  06:55.....

 JB> so, if you want it run for three months non-stop (amd leaving no
 JB> margin to sound the alarm) that's approx 2000 hours, so from 500mAh
 JB> it's 250uA, from 9V that's 36 megaohms resistance.  which is near the
 JB> edge of CMOS performance, but with a resistance like that it's not
 JB> something you can build on veroboard. 

If the veroboard is properly cleaned there should be no problems
with leakage. This may require washing it with alcohol to remove
all traces of flux. If necessary the whole circuit can be sealed
in silicone sealant -- be sure to use the non-acidic type.

Using a microcontroller as the core and powering all the
external stuff from one of the ports would be a better way.

According to the data sheet a 16F84 in SLEEP mode draws about
28uA max (typical 7uA). A 500mAh battery would run it for well
over 2 years at maximum current, or up to 8 years with the
typical current value.

OTOH 9v is too high for the 16F84, so a bank of four AAA cells
would be more suitable. Unfortunately I don't have any figures on
their capacity. My guess is that they would be higher than the 9v
battery.

 JB> Possibly an old smoke detector could be modified, and the "sign of
 JB> life" pulse (which flashes the LED once per minute) could be used to
 JB> trigger the temperature sensor. you get a low battery alarm this way
 JB> too. 

That's an interesting possibility. I have some old smoke
detectors, but none of them has the "sign of life" LED.

   ,-./\
  /     \ From Greg Mayman, in beautiful Adelaide, South Australia
  \_,-*_/    "Queen City of The South"    34:55 S  138:36 E
       v

... An apple a day keeps the doctor away..if you aim it right
___ Blue Wave/386 v2.30

--- FLAME v2.0/b