Fifty years... That's a lo n g time. Kind of an unreasonably long time. Are you telling me no one can be bothered or afford to come by and change batteries once every 25 years? There are batteries qualified for probably 25 years but I haven't seen any qualified for 50. Also since your app will require something to be powered up doing useful work, (the strain gauge and its 'measurer'), the power consumption will be more than inconsequential.

I believe you will need to think out-of-your-current-box. Can you remote the batteries to somewhere accessible?

Does this all happen in the pitch black? Can you not cram a tiny solar cell somewhere?

Is this somewhere energy harvesting can be applied? If it's a bridge or a tower there may be sufficient vibrational energy to run the system.

This company has THE longest lasting batteries available. My 25 years "guess" seems to be spot-on if you note the headline on the home page.
http://www.tadiranbat.com/

Keith Cress
kcress - http://www.flaminsystems.com

Seems to me that you'd have problems with the hardware and software way before you hit year 20 even; it's very likely that the mating transceiver will simply be long lost or broken well before the 50 years is up.

TTFN
FAQ731-376: Eng-Tips.com Forum Policies

In the AMR (automatic meter reading) market they do make some specialized water and gas metering radio devices that can survive up to 20 years on batteries. These are used on some buried meters meant to measure and report at intervals. But this is at the extreme of this market. They use LiMgO2 (photo lithium) non-rechargable batteries, but even these batteries are used far beyond their normal shelf-life. Electronics are designed to sleep at no more than a couple microamps.

What you want is beyond the self-discharge of typical battery technologies. You will have to go to a parasitic power approach - be it thermal, vibration, EM field, micro-solar, whatever. Alternately you can go for a type that would be periodically recharged using a RF or magnetic field - somewhat like a RFID reader that would query the device and at the same time put more energy into it. But even the typical power storage source for parasitic power or RFID devices (supercapacitors) have leakage currents 5X what the AMR industry would consider adequate.

The military does use some special battery technologies that can be 'activated' after years of storage - usually by a small pyrotechnic charge that opens a small chemical vial. Maybe a setup with several of these where when one battery reaches the end it activates a new battery every 15 years or so.

I know - everything I can think of is too complex, too expensive, or too large for your application. Good Luck!

Thinking aloud... how long do batteries for medical implants for pacemakers last? And what cell chemistry do they employ?

As a reference, pacemakers are generally intended to last about 5 to 10 yrs.

http://www.mayoclinic.com/health/pacemaker/MY00276...

TTFN
FAQ731-376: Eng-Tips.com Forum Policies

Thanks for all of the suggestions guys, I need to do some research into all these possibilities. Im intrigued by the parasitic power approach. Since this unit will be inside the walls of a house, solar is definitely out of the question. We cannot tap into the existing electrical lines in the house because of a huge increase in installed cost to the end user and it would be hard to work around the building code for internal electronics. My question is, how could thermal, vibrational, EM-Field etc. be used to give a battery a slow charge over many years. What if we used a 20 year battery and used one of these technologies to give it a slow trickle charge? Could this effectively increase the longevity of the device? How can a stationary unit attached to the framing of a house use one of these techniques to charge a battery? Remember tha the device does not require much power to operate...

?? a wall wart should only add about $10 to the base cost

TTFN
FAQ731-376: Eng-Tips.com Forum Policies

Battery electrolytes are either caustic or acidic. In either case, their containers are never intended to survive that long. You could probably commission a custom battery, but that would be way more expensive than a wall wart.

TTFN
FAQ731-376: Eng-Tips.com Forum Policies

You've already mentioned the device is to measure a strain gauge. How much energy is in this phenomenon you're trying to measure? Could a piezo-stack generate enough charge to power the device which then records the data? Basically use the piezo for both the sensor and power source? Parasitic power is presently a popular subject of electronics right now, but few possible applications will prove both cost effective and reilable.

In the realm of impractical exotic expensive power sources, your question made me think of a device I read about a few years ago. A radio-active alpha emitter packaged so it bombarded some substance creating free electrons, and would work as a power source providing a few uA of power to a device for years. Was about the size of a tube of lipstick. I believe it had a price tag that would even make the military wince!

50 years is a LONGGGGGGGG time for a power source - even the Voyager 1 and 2 spacecraft are only 35 years old, and I don't think they're projected to last 50 years on the RTG that powers them (radioisotope thermoelectric generator). BTW It's expected any day now that NASA will announce that Voyager 1 has officially left the solar system into interstellar space.

By continuously monitoring, do you mean measuring the strain every 1/60 of a second? That doesn't give the electronics much time to sleep. The AMR meters I have seen only leave sleep mode to record a measurement every several minutes to several days. Can you use something like a mercury-in-rubber strain gauge to turn on the electronics, or even to open a chemical via?

Jumping in a bit late, but...

One of my lighting products spends most of its time in sleep mode, waking about 10 times a second (if memory serves) to check on the state of a pin (relatively high-impedance, say 100k)... average power draw is under 50uA. That includes time (and power) to start and stabilize the clock with each wakeup. I was duly impressed with such low power requirements, but you need to power up and read an A/D converter, as well, and that typically requires time to stabilize, which means even more power. I wouldn't expect to eat less than several hundred uA, at best, though to be fair that is merely a guesstimate. Interested in hearing how it turns out.

Dan - Owner
http://www.Hi-TecDesigns.com

I don't know your wiring methods but in North America a wire run parallel to the cable to the hot water tank or electric heating element could pik up enough induced voltage and current to power micro electronics and charge the battery.

Bill
--------------------
"Why not the best?"
Jimmy Carter

Ah, now that's pretty clever Bill.. I like it.

Keith Cress
kcress - http://www.flaminsystems.com