Getting last joules out of batteries

[klugesmith]

Cordless mouse on this PC failed a few minutes ago, and I suspected its battery. 
Swapped the battery, a single AA cell, with one in a quartz-regulated, analog-faced wall clock.

Apparently the mouse needs less juice than the clock.
Mouse is working fine with ex-clock battery.
Clock second hand is barely twitching with ex-mouse battery, even with clock in horizontal plane.
Normal failure in a wall-mounted clock leaves the seconds hand pointing to 9, and clock runs longer if laid flat.

Has anyone here tried a "joule thief" LED blinker circuit,
with a battery at the point of failing in mouse or clock service?

[davekni]

Apparently the mouse needs less juice than the clock.

Much lower peak current required by mouse, but higher average current.  Mechanical clocks fail mostly due to battery internal resistance.

[Weston]

There have been one (or more?) attempted products to use a joule thief like circuit to extend battery life in existing products. I don't think any of them were commercially successful https://en.wikipedia.org/wiki/Batteroo_Boost

There is also not that much energy remaining in the end of the battery discharge curve. If the issues is ESR I don't think a boost converter would help much but I wonder if a large capacitor would help.

Regardless, the joule thef is a fun electronics project and circuit to analyze as a circuit design problem. I remember one of my first electronics projects in middle school was a joule theif based flash light using a a small common mode choke I pulled out of a power supply. I was quite proud of it until I think someone stole it on a school trip.

More recently I made a jfet based joule theif that could start up and light a LED down to a ~40mV input, which was enough to be powered by a flame and a few type K thermcouples!

The open source SKY130 PDK has some zero threshold mosfets so I was considering making a similar energy harvesting circuit and submitting it to one of the open source IC tapeouts, but I am a bit burnt out on IC design stuff at the moment.

[klugesmith]

Nice work, Weston.
In case this drifts into "useful engineering with very low voltage"...

Thermocouples are ubiquitous in gas valves for appliances with standing pilot lights.

A single junction (IIRC it's type T or J, with constantan on one side) generates about 25 mV.  That powers a solenoid to hold gas valve open, opposed by a spring that closes valve if pilot light is not burning.  The holding current requirement (with magnetic circuit closed) is much less than the pull-in current.  That's why, when lighting the pilot, one needs to keep mechanical button pressed until thermocouple is hot enough.

The solenoid stands as an example about wire windings in general.  Designer selects wire diameter, to trade between voltage and current, with no change in power requirement or coil dimensions or ampere-turns.  For the gas valve it's a few turns of thick wire, to run on 25 mV.

[rikkitikkitavi]

I thought that pilot flames where some kind of bi-metal that you need to mechanically push into the flame with the push button, and when heated by the flame it of course bends, and then maintains valve open by some mechanical contraption?

[davekni]

I thought that pilot flames where some kind of bi-metal that you need to mechanically push into the flame with the push button, and when heated by the flame it of course bends, and then maintains valve open by some mechanical contraption?

That seems like a logical alternative too.  The couple gas heaters I've personally used were thermocouple/solenoid.

[klugesmith]

In my experience, the bimetal bending parts are only for thermostatic control of main burner valve. (in a semiconductor-free water heater control).  There may be systems that work as Rikki described.

The conventional safety interlock has a poppet valve, normally held closed by a spring, which blocks all gas flow.

To light the pilot, step one is to push and holds a mechanical button which
1) opens the valve all the way
2) closes the magnetic circuit of a very-low-low voltage solenoid.  Holding voltage requirement may be 0.015 or 0.020 V.

The job of thermocouple, heated by the pilot flame, is to give the solenoid enough ampere-turns to hold the valve open when mechanical button is released.  Jump to 2:30 in video if URL doesn't do that.

/>
An easy physics demonstration employs a hand-held solenoid unit mated to a thermocouple. Push the plunger in, watch it pop right out.   Heat thermocouple with a lighter flame until the plunger stays in when pushed all the way.   Voltage and current metering may add value for EE nerds, but are not trivial.

[rikkitikkitavi]

Ok, cool. I bow to your knowledge. Smarter every day (thanks Destin for your programs) due to this forum

 Definiately some cunning engineering applying physics.

[klugesmith]

Storage space cleanup revealed 50-year-old heat controls using both methods.

The furnace part on right has a probe that extends into heat exchanger space. 
Bimetal strip, formed into a helix, rotates metal disk with adjustable set-point tabs.  There's enough torque for the tabs to operate snap-action electrical switch that controls main air blower.

The water heater control on left has pilot light safety valve, as we have discussed, and a non-electrical (?) thermostat mechanism for main gas valve.

In this specimen, the gas inlet opening gives a view of safety valve.   Opened by manual button in line above it.  Held open by solenoid with thermocouple fitting directly below it.