Sunday, May 13, 2012

A mechanically-powered capacitor flashlight?

A later article about a less-impractical and potentially more useful capacitor-powered flashlight was described in the
August 14, 2012 post which may be found at this link (click here).

You probably saw them in those annoyingly repetitive low-budget commercials several years ago (similar to those for the "Faraday Flashlight" or "Faraday Torch"):   Flashlights that you "shake" to generate electricity with the claim that "you'll never need batteries".

The insides of a "shake" flashlight, showing the magnet, coil, circuit board and the two non-rechargeable CR3032 lithium coin cells.
Figure 1:
Insides of a "shake" flashlight showing the magnet, coil, etc.
Click on image for a larger version.
Why, then, do they have a battery?

Careful inspection of the flashlight - even without taking it apart - will reveal that it has a pair of non-rechargeable CR3032 lithium coin cells - the two edge-on flat metal disks as seen in figure 1Not visible in this picture is a small, green cylinder on the backside of the circuit board to the right of the coin cells which is a 0.22 Farad, 5.5 volt capacitor.  It is this component that is capable of being charged and discharged many times while the disposable coin cells aren't!

The way that these flashlights are purported to work is by the motion of the user causing a magnet to move repeatedly through a coil, this being different from the rotating fields in more conventional, rotary electric generators.  As the strong magnet moves, current flows through the coil and is rectified using a 4-diode bridge to assure that the polarity is always correct.  A simple switch with a series resistor and white LED completes the circuit, providing light.

Why the battery, then?

Several years ago - as an experiment - I took one of these flashlights on a hike with some friends, having removed the battery and discharged the capacitor completely.  As it started to get dark I took out the flashlight to see how useful it actually would be if it were necessary to generate all of the power through the motion of the magnet.

Knowing that it was going to get dark (it always does at night, around here!) I wielded the flashlight while there was still enough light to see and started shaking it, taking partial advantage of my already-swinging arms as I walked along.  It took several minutes before I got even the slightest hint of light from the light, but that was no surprise - and for a very good reason:  The capacitor had to be charged to 2.7-3.0 volts before the LED would produce any light at all and I still had another 1-2 volts to go before enough current could flow through the circuit and the flashlight had any hope of being bright enough to be useful!
Figure 2:
Circuit diagram of a typical "Shake light" - nothing special here!
Click on image for a larger version.
By the time it got dark enough to require the flashlight I'd managed to charge the capacitor to the point where it was quite bright when it was first turned on, a process that had required a mere 10-15 minutes of walking with the flashlight's magnet moving back and forth and occasionally shaking it.

Upon turning it on, the light dimmed dramatically within a few 10's of seconds, but stayed illuminated, providing barely adequate light to avoid large obstacles - but not quite enough to avoid medium-to-small rocks and tree roots in the trail.  Fortunately, being quite tall and fairly familiar with this trail - plus hiking with a group of people who had proper flashlights - I was easily able to lift my feet to avoid obstacles and avoid doing face-plants in the dirt!  I could vigorously shake the flashlight to briefly get more light, causing the light to be pointed in random, useless directions while doing so, but it faded down again soon after I stopped.

What did work was to completely shut off the flashlight and, "mooching" off the lights from the rest of the hiking party, continue on my way while shaking the light as I walked along, only occasionally turning it on.  This strategy of turning off the light while shaking permitted the capacitor in the flashlight to be charged to a higher voltage than it would be if it were on with the LED dragging the voltage back down faster than it was charged and this allowed occasional bursts of bright, useful light.

My suspicions as to why a "free energy" flashlight actually has a battery were confirmed:  It is unlikely that the consumer would actually follow the instructions for a flashlight that required some action beforehand such as:
"Shake vigorously for 3-5 minutes before turning it on.   When it gets dim, turn it off and then shake it vigorously for a several more minutes before using it again."
Thoughtfully including a few inexpensive coin cells allows for both instant gratification and light:  Without having to work just to get the light to work out-of-the-box probably went a long way toward preventing customer returns on flashlights that didn't seem to work because people didn't read and follow the directions!

If you look at the diagram in figure 2 you'll notice that the non-rechargeable battery is placed in parallel with the energy from the coil and the capacitor.  What this means is that after extensive use with the coin cells exhausted, the battery is still in place, adding a bit of loss to the mechanical generation of power as one attempts to "charge" them.  Ignoring this issue, in the case of a light with a dead battery, the user may not be aware that it might be necessary to shake the such a light for a while before it becomes useful - and that assumes that the battery isn't going to badly drag down the charge accumulating on the capacitor!

What's worse, there's the impression by the user that it "never needs batteries" and that it can always be counted on:  While this is (technically) true, it's unlikely that someone finding the battery dead in this thing from having accidentally left it on would be accustomed to shaking it for a few minutes just to get any light at all.  These factors probably explain why it is that one hardly sees these flashlights advertised anymore!

A practical "shake-light" is possible, but the folks that made the unit described above didn't go about it in the right way for a number of reasons, no doubt related to keeping costs down:
  • The coil/magnet's mechanical setup isn't as efficient as it could be.  The use of springs (or other magnets) would improve the mechanical efficiency and "spring" the moving magnet away from the end-stops rather than the mechanical energy being absorbed in the plastic/rubber bumpers at each end.
  • More efficient coil arrangement.  A series of independent coils (e.g. a number of "linear poles") stronger magnet(s), etc. might allow for more energy recovery rather than a single, large coil.
  • Resistive losses in the coil.  The current of the of the coil is dumped into the capacitor fairly inefficiently.  As noted below, even if the coil was loss-less, the capacitor's internal resistance would rob the effort of efficiency. 
  • Improved efficiency of energy recovery from the coil.  Ideally, one would use a small, switching converter to optimally take the energy from the coil and use it to charge the capacitor.  As it is, a 4-diode bridge (using 1N4001-type diodes) has a minimum "2-diode" voltage drop (of around 1.2 volts) is used, wasting a significant portion of the effort - much of it due to effective impedance mismatch between the energy source (the coil) and the load (the capacitor).
  • Better capacitor.  A standard "supercap" is used in the above flashlight.  These capacitors have fairly high internal resistance (10's of ohms, typically) which means that a fair amount of energy is actually lost when the capacitor is charged or discharged - particularly if either is attempted quickly!  In comparison, "Ultracaps" and similar have far lower internal resistance (fractions of ohms) and would work much more efficiently.  It's worth noting that two "fresh" lithium coin cells in series will output about 6.4 volts - this, into a capacitor that is rated for just 5.5 volts!
  • A switching converter to run the LED.  This could be as simple as something along the lines of the "Joule Thief."  As noted above, the LED doesn't even begin to light until 2.7-3.0 volts or so appears across the capacitor and it isn't usefully bright until there is 3.6-4.2 volts available which means that we have over "3 volts of effort" before we get any light at all!  A switching converter could not only allow the LED to be illuminated with as little as 0.8-0.9 volts across the capacitor and utilize more of its charge, but it could more efficiently and consistently power the LED over the voltage range instead of burning power in the resistor shown in the diagram above!  Practically speaking, in order for this to work one would need a better capacitor as noted above.
  • Regulating the LED's current.  As odd as it sounds, reducing the LED's current would probably help make the light a bit more useful.  As it is, an 18 ohm resistor is present - most likely to prevent the two coin cells from being exhausted as quickly, but this simple arrangement causes there to be too much LED current when the battery is fresh and, perhaps, too little when the battery is weak - or possibly when running from the capacitor alone.  A simple current regulator circuit to keep the LED current lower all of the time would help to maintain a more consistent light level under most conditions even if it were lower overall and allow the charge from either the capacitor or battery to last longer.  More ideally, a simple regulator could be incorporated into a simple circuit like the "Joule Thief" mentioned above and increase usefulness and efficiency and charge longevity all at once!
  •  Some sort of power management.  All of the above tend to conspire to reduce efficiency, but if there were some "smarts" involved better use of the available energy from the moving coil could be made.  One growing field of interest has to do with "energy harvesting" and some of the techniques and available chips to do this might be adapted to improve overall efficacy.
While doing all of the above could actually make such a device practical, it would also (significantly) add to the cost.  It is unlikely, however, that a "shake light" would ever be as "human-efficient" as a simple crank-type generator owing to the mechanical nature of this action:  Turning a crank takes a bit less effort than moving one's entire arm back and forth!

What about any sort of practical, capacitor-based, "battery-less" flashlight that would provide a useful amount of light, run time and storage longevity - even if one that's not charged by the user's muscle power directly:  Is this even possible?

The answer is "Yes" - sort of, - see the August 14, 2012 entry - click here.


This page stolen from

1 comment:

  1. I got something similar specimen with dead supercapacitor. It looks like the CR2032 cells in this can be charged as the voltage does go up and stays that level when shaking it for 30 minutes or so. There's a good method for charging it by holding it in 45 degree angle and rotationg it 90 degrees so the magnet slides from one end to another by the gravity alone.



While I DO appreciate comments, those comments that are just vehicles to other web sites without substantial content in their own right WILL NOT be posted!

If you include a link in your comment that simply points to advertisements or a commercial web page, it WILL be rejected as SPAM!