On Sat, 27 Feb 1999 10:27:26 -0700, you wrote:

>First of all, thanks for all the help in the past on other projects.
>
>I am building a unit that must be powered by batteries.
>The thought is to use 3V camera batteries.
>The unit is desired to run for months on one battery set.
>
>The unit passes a beam of light through a clear plastic tube.
>When the material inside the tube blocks the light, a 3 volt relay
>is activated momentarily.  The relay requires 1 amp.
>Because of the material being sensed, the beam must be in the visible range.
> Right now a CdS cell is being used.
>The CdS cell is slow. (response is around 10 milliseconds)
>It looks like the LED could be a major power sucker.
>I am trying out the idea of charging a capacitor and dumping it across the LED.
You'll need a big cap to supply the pulse to the relay (as you won't
want to cause a brown-out on the battery supply when it fires), watch
the leakage spec on this. Do you really need a whole  amp for the
relay? That's a pretty big relay. Have you considered a solid-state
output, or a latching relay ? 
>
>The sample rate is 3 times per second.
>For each sample, the idea is to turn on the PIC,  the CdS circuitry, pulse the LED,  read
>the resistance of the CdS cell, maybe do some calibration, determine if the relay should be pulsed.
>the relay will only be pulsed a maximum of 5 times per day.  Then turn everything off.
>
>Questions
>
>1)  What could be used to economically switch power to the relay?
Transistor or mosfet, base current & voltage drop shouldn't be too
significant compared to the relay current for a transistor. If you use
a mosfet, watch the gate voltage spec carefully - you'll have a hard
time finding anything below about 2.7v, but you could use a simple
(software clocked) charge-pump doubler to drive the gate - this would
probably be the best solution at 3v.
>2)  Is 3 volts going to work as the supply voltage (especially since the
>voltage will decline as the battery is discharged)  Would a higher voltage be better?
6V would give you more headroom & make things a bit easier (e.g.
driving a mosfet switch), but 3V should be do-able. Camera batteries
would be a pretty good choice for this app. 
>3)  Is there a clever way to reduce the on-time of the LED?
>If I had a faster sensor (in the visible range), I could shorten the
>LED on time.  Is there a low power source-sensor pair for visible
>light?  Clever ideas on reducing the power required for sensing?
If you use a phototransistor, the 'on; time should only need to be a
few uS - connect the LED and the phototransistor's pullup resistor to
the same pin to power them both at the same time. 
>4)   Any cool suggestions on power management?
Use the WDT to wake up from sleep mode - doesn't sound like the sample
time is too critical. 
back-of-an envelope calculation of power consumption : 
PIC in standby (sleep, WDT on) = 30uA
PIC @ 1mA on for 1mS, 3x per second  = 3uA
Led @ 10mA on for 10uS = 0.3uA
relay @ 1 sec, 5 times a day = 50uA
Round it up to 100uA, 
a 2CR5 6V camera battery @ 1300maH gives 13Khours, = 540 days
call it a year to be on the safe side.
Remember you may be able to use a higher voltage (= lower current)
relay with a 6V supply. Also battery contact resistance (e.g. due to
corrosion etc.) will be less of an issue at 6V.