The Forever Flashlight II - Batteries Not Required

Richard J. Nelson

I have always had an interest in unusual designs for electrical and electronic products. I am especially
interested in products that actually work as I expect them to. I don't mean that they just work right out of
the box, but rather that they work in my world, which it seems, is not part of the "normal" product design
world. Reliable, well thought out, well engineered, feature rich products will always get my attention.
How many times have you needed a flashlight and when you found one it didn't work? The reasons are
many; dead or corroded batteries, loose connections (discovered by beating on or shaking the flashlight),
or a burned out bulb. Mag Lite has addressed the latter issue with their policy of including a spare bulb in
the flashlight.
The Forever Flashlight (and others, see the last part of this report) utilizes the energy you spend shaking
or beating on a conventional flashlight to actually provide the power to operate it. See figure 1 for an
overview of the Forever Flashlight.

Figure 1 — Overview of the white LED shake powered Excalibur Forever Flashlight, batteries not required.
Seeing one a friend had bought inspired this technical product exploration. I
immediately bought one at Brookstone, store 241, supplied by Excalibur
Electronics Inc. of Miami Florida. The flashlight was actually made in China,
of course. Find Excalibur on the Internet at www.excaliburelectronics.com
Excalibur has a broad product line but the "new" Forever Flashlight II has only
a casual mention. There wasn't even a decent (sized) photograph of the product.
Since I recommend this product to friends I felt that a few words describing it
were needed.
The Excalibur website lists the flashlight at $39.95 and Brookstone sells it for
$30.00. The Brookstone website has a better photo. In addition to the shake
powered mechanism I was impressed with the lens system and its quite narrow
beam projection of light. See figure 2. Do the math using the numbers shown
in figure 2 to calculate the beam (included) angle at about 15 degrees. This
means that you will be able to see your path in total darkness as you walk out of
the mountains after a long weekend hike. Adding oil or power steering fluid to
your engine at night requires a reliable flashlight, one that has been untouched
for years. Fig. 2 — White LED Beam.
Page 1 of 14
The future of low power lighting is in the use of the Light Emitting Diode, LED. Sooner or later you will
buy one or more LED lights or lamps. The two primary advantages of an LED light source is it's efficient
light generation (exceeding that of fluorescent lamps) and long life, typically 100,000 hours. That is 50 to
100 times more life than most of the bulbs in your house.
Most traffic lights now use LED's because the cost of replacing the bulbs is labor intensive and the
electrical savings of 24-hour operation mean a very fast payback of the investment. An added bonus is
that there are many LED's for each of the colors and the failure of one of them doesn't affect the system
very much. Old style traffic lights only use two bulbs per color.

Newer vehicles also use LED's for their taillights. Hewlett-Packard is a major supplier of LED's for the
automotive industry.
I bought the clear cased model because I am technically curious and I wanted to "see" it work. I had to
ask for the clear version at the store I visited at the Valencia Mall in the Santa Clarita Valley (CA). A
clear case may not be the best choice if you don't want light spilling out from the sides of the flashlight
when being used. Tape comes to mind, or a little black paint applied to the inside of the LED Chamber.
The flashlight uses a moving magnet through a coil to charge a capacitor to power the LED. The charging
process described in the instructions suggests holding the flashlight horizontal and moving back and forth
at about three "shakes" per second for a total of 90 shakes - for five minutes of light. See more on that
topic later. Obviously shaking a heavy magnet requires shock-absorbing bumpers and shaking vertically
or too hard only endangers the flashlight mechanically and electrically.
The Forever Flashlight looks like most two "D" cell flashlights you have used. It is made of plastic and is
mostly empty space inside so it is much lighter and it actually floats. It is different from most flashlights
in that it only unscrews from the LED bulb end. Mine was quite tight and I don't believe that the design
calls for cementing the two pieces together. I can't think of any reason to open the flashlight except for
curiosity or possibility service. After you read this and study the photos you will not have to open yours.
When you first see the flashlight you will naturally want to look into the LED to see how bright it is. The
lens/reflector system, as mentioned above, is quite
effective. You will be able to look through it deep into
the LED to actually see the two gold one-mil diameter (a
mil is 0.001"- a human hair ranges from 1 to 3 mils in
diameter, a sheet of copy paper is about 3 to 4 mils thick)
bond wires used internally to connect the LED die to the
leads. You may have to look a little off-axis to see these
bond leads. The LED is not that bright. If you are
bothered by its brightness just wait two or three minutes
and it will dim. Figure 3 illustrates what I am talking
about.

Conventional LED's usually only have one bond lead.

The semiconductor die is attached to one lead post
making an electrical connection on the bottom. A bond
wire is then used to connect the top of the die to the other Fig. 3 — The lens magnifies the two 1 mil dia-
lead post. See figures four and five on the next page. meter LED bond wires to make them more visible.

Page 2 of 14
 If you are interested in LED's and how
they work you may check the Internet.
There is lots of information available
from the manufacturers, schools, and
how things work web sites.
The Forever Flashlight, FFL, is opened
by unscrewing the black rubber like
ring at the LED end. See figure 7
below. The "O" ring that seals the case
against water is shown below the
threads in figure 6. The "O" ring is
estimated to be about 37.25 mm in
inside diameter (1.465") and 2.5 mm in
in diameter (0.089"). These may be
stretched measurements. If you use
your FFL around water you may want
to apply a small amount of lubricant to
the "O" ring to insure that it stays
pliable and water resistant. The seat Fig. 4 — Conventional LED Fig. 5 — White LED wire
groove diameter is 44 mm (1.732") wire bonding, single bond wire. bonding, two bond wires

Fig. 6 — The retainer ring unscrews to show the Fig. 7 — The retainer ring and lens. The lens is glued to
black "O" ring used to insure water tightness. the retainer ring and projects a 15-degree beam.
The Forever Flashlight, FFL, uses two magnets. A large NIB magnet (Neodymium - Iron - Boron, the
most powerful magnet made) is used internally to generate the power to run the FFL. The switch,
partially seen at the lower right bottom in figure 6, operates much like your standard flashlight except it is
many times better, smoother, and electrical contact free. The switch itself is on the inside of the sealed
plastic case (it is water proof, remember) and it uses a second magnet in the button to actuate a magnetic
reed switch inside the case. You may also operate the FFL with an external magnet if you wish. Figure 8
shows the LED mounted in the LED chamber about 26.9 mm (1.06") from the end of the inner housing.
Figure 9 shows a close up of the white LED, turned off for better visibility. See figure 10 for a photo of
the inner housing.
I will describe the parts of the flashlight grouping each major section as indicated in the notations of

Page 3 of 14
figure 10, A through F. What follows is specific to model 422-M purchased in CA mid February 2004.

Fig. 8 — LED shown in its housing. Note the clear Fig. 9 — The single LED as viewed from the side
sides and flat reflector at the base of the LED through its housing. The LED reflection is on the right.

Figure 10 — Internal housing with each section identified for reference in the text.

A - LED Chamber: The LED chamber has been mostly described above when discussing the LED
light performance characteristics of the FFL. A couple of additional points need to be made. The light
itself is not pure white. There is a tinge of blue "spots" in the center of the beam. This is due to the type
of white LED being used. White LED's are not really "pure" white. This is an especially important
quality if you are taking photos using white LED's as a light source. You will definitely notice the blue.
It is not clear to me how the LED is mounted. Read more on this in the next section. It seems that the
base of the LED is glued to the flat circular reflector sheet that is, in turn, glued to the bottom of the LED
chamber. The beam focusing system consists of the dome-lens LED mounted on a flat reflector and a
double convex plastic lens to focus the beam. The lens is 40 mm (1.567") in diameter with an aperture of
31.3 mm (1.232") and focal length of about 35 mm (1.38"). The plastic lens is glued to its retainer ring

B - Circuit Board Chamber: The two flat rigid leads of the LED pass through two rectangular
holes in the wall between the LED chamber and the circuit board chamber as shown in figure 11. These

Page 4 of 14
leads must be unsoldered before the circuit board
may be removed from its housing. The two screws
(see fig. 12) holding the circuit board, upper left and
lower right corners, must also be removed with a #0
Phillips screwdriver. The coil leads may be
straightened out and left attached if desired.

One component on the etch side of the circuit board

is the reed switch mounted on a rigid standoff
straddling the two LED leads. A close up view of
the reed switch is shown in figure 13. The internal
housing has three short "tongues" at the rear end that
must be aligned with the black rear bumper retainer
when inserting the inner housing into the case. The
"tongues" insure proper alignment and positioning
of the reed switch with the magnet in the button of
the external sliding "switch."

Reed switches are composed of two magnetic blades

mounted at the opposite ends of a rigid glass tube.
When the blades are exposed to a magnetic field Fig. 11 — Flat rigid LED leads pass through the LED
they bend towards each other providing a switched chamber wall and are soldered onto the circuit board.
electrical circuit. See figure 12 for a circuit board chamber overview.

Fig. 12 — Circuit board chamber containing a single Fig. 13— Close up of the magnet actuated reed switch
sided circuit board that connects LED, reed switch, mounted on a rigid plastic stand off to position it close
storage capacitor, four full-wave-bridge-connected to the external slide switch. A magnet in the sliding
discrete rectifier diodes, and coil leads. "switch" on the outside of the case closes the switch.
An enlarged view of both sides of the circuit board is shown on
the following page as figures 15 and 16. The circuit board is
screw attached to standoffs on the inside its chamber. You must
remove it as described above to see it clearly. Refer to figure 14.
The largest component is the energy storage capacitor seen at the
right side just to the left of the upper right standoff. The four
standard rectifier diodes are seen along the bottom. Two screws
(left end) secure the reed switch stand off.
Fig. 14 — Ckt board Component side.
Page 5 of 14
Fig. 15 — Enlarged view of etch side of the single sided circuit board. Traces are accessible for measurements.

Fig. 16 — Enlarged view of the component side of the circuit board. The components are inside the housing.
Page 6 of 14
The photographs provide an idea of the physical layout of the FFL. The electrical schematic is shown in
figure 17 below. Reference designations A through E will be used to describe the operation of the circuit.

Fig. 17 — Schematic of the model 422-M Forever Flashlight. Letters A-E identify the various portions of the
circuit for reference in the text. The same designations have been added to an enlarged view of figure 15 shown
as figure 20, which will serve as a "working" guide to probe the circuit board to make measurements.
As the magnet slides through the coil, a bipolar pulse is generated proportional to the strength of the
magnet, the speed it passes through the coil, and the number of turns in the coil. When the magnet slides
back another pulse is generated of the opposite "polarity." See additional details in part E - Coil Form.
The pulses of alternating positive and negative values appear across the coil at points A & B. Diodes D1
through D4 form a full wave bridge rectifier circuit. When A is positive diodes D3 and D1 conduct and
diodes D2 and D4 do not conduct. A diode has the property of conducting current in one direction only.

When the polarity is reversed, Diodes D2 and D4

conduct and diodes D1 and D3 block current flow.
Each diode has about 0.6 volts across it when it is
conducting. The total voltage drop of a bridge rectifier
is about 1.2 volts. Another important characteristic of a
bridge rectifier is that it provides isolation from the
circuits that follow. The purpose of the bridge is to
charge capacitor C1. Once charged we certainly
wouldn't want the coil to discharge it. The capacitor,
C1, is often called a super capacitor and it may be seen
up close in figure 18. Fig. 18 — Super capacitor label.

The capacitance is 0.22 farads at a rated voltage of 5.5 volts. This voltage is important because exceeding
it could degrade the capacitor and it may no longer hold a charge and the flashlight is no longer able to
work. If you are tempted to really shake it, don't. A farad is a unit of charge, which will provide one
ampere of current for one second into one ohm, or 1 watt second. Most capacitors are measured in
microfarads or one millionth of a farad. Once charged, the reed switch, S1, is turned on and the LED is
powered by the capacitor.

LED's have a very non-linear voltage current curve and work best at about 2.5 to 3 volts. At higher
voltages their effective resistance is very low and some means, usually a series resistor, is used to limit the
Page 7 of 14
current to a safe value. There is no series resistor in this circuit. The LED, however, does not appear to
be severely over driven (very bright). Perhaps the LED itself has current limiting. Is the extra bond wire
shown in figure 5 used for this purpose? This is left for the reader to research.

If you examine the circuit board photo in figure 20 you will notice five solder pads with "X" marked on
them. These are unused holes that have filled with solder during the solder process. The corresponding A
through E points shown on the schematic are also indicated. The capacitor traces are identified with the
C1- and C1+ designations. The LED has long leads that pass through channels in the reed switch
standoff. They are soldered at their ends even though there are pads on the far left edge of the board.
This is another clue to indicate that this is a multifunction circuit board (used for several different
models). Point E indicates a trace that is not visible because of this standoff. It connects to the upper
LED lead under the standoff.
If you study figure 16, which is the other side of the circuit board, you will notice several designations for
components that are not used. One is a zener diode, ZD, which appears to be intended to clamp the
voltage applied to the capacitor to protect it. Could a properly designed coil and bumper system work
well enough to eliminate the Zener clamp diode? Reduced parts count is always a design goal.
Another designation is the circles around the capacitor. Are there models that use a larger capacitor?
Commercial sizes up to 1 Farad are readily available. This would provide five times the storage - and
increase LED operation time.
The diodes are conventional
standard garden-variety
general-purpose rectifier
diodes of the 4000 series,
1N4001. This version has a
50-volt peak repetitive
reverse voltage specification
and a 5-microampere
leakage current at room
temperature. They are
cheaper than a complete
bridge in a single package,
which would also require
less assembly time. Fig. 19 — Bridge diodes, 4 - 1N4001, 50 V PRRV, 5 µA leakage at 25°C.

The circuit is very simple to keep the parts count to a minimum. Products like this usually evolve over
time as additional engineering effort is made to further refine the circuits and component selection to
reduce costs and improve performance.
The most expensive part of this product design is the tooling required for the case and inner housing.
These are large plastic molds that require a substantial investment. Assembly labor is cheap in china so
the circuit board assembly operations are minimal. All of the electrical components except the coil (and
perhaps the reed switch standoff) are off the shelf items. The magnet is another fairly expensive part.

C - Front Bumper: The front bumper and the rear bumper are the same. They are part of the
mechanical parts that are tooled especially for the product. The purpose is to absorb the energy of the
moving magnet. You will notice the shape of the bumper shown in figure 21. The magnet is quite heavy
and it is human nature to want to shake the FFL very hard - even thought the instructions say not to. The
material is similar to a hard rubber.
Page 8 of 14
Page 9 of 14
D - Magnet: The magnet is the most powerful
type available. It is a Neodymium Iron Boron,
NdFeB or simply, NIB magnet. Neodymium Iron
Boron (and Samarium Cobalt) magnets are general-
ly known as rare earth magnets since their com-
pounds come from the rare earth or Lanthanoid
series of the periodic table of the elements. They
were developed in the 1970's and 1980's. Refer to
www.magnetsales.com for additional very general
information on magnets.
You may easily remove the magnet to "play" with
it (see the rear bumper section). It is strong enough Fig. 21 — Magnet against front bumper.
to attract metal objects with such force that you
could pinch your skin. You must be very careful in
handling it. The same force could chip objects and
flying pieces may get in your eyes. Exercise
caution. This is not a joke. Do not bring the
magnet to the surface of a TV or computer monitor.
You may move it moderately fast with the
thickness of your fingers between the magnet and
the glass surface to demonstrate the effect. DO
NOT HOLD THE MAGNET ON THE SURFACE.
It will put a permanent dark "mark" on the screen.

Even normal use of the FFL may cause small Fig. 22 — Typical objects the FFL magnet attracts.
objects to be attracted to it even though the magnet is over 8 mm (0.315") from the "outside world." See
example in figure 22. The washer on the slide "switch" shows the reed switch magnet.

The magnet measurements are: Diameter. . . . . . . 15 mm (0.587")

Length. . . . . . . . . 20 mm (0.787")
Weight. . . . . . . . . 53.0 gr. (1.87 oz.)
E - Coil Form: The heart of the FFL system is
the coil (and magnet). One advantage of sliding
the magnet back and forth is that it only requires
one hand. A cranking generator would require
either two hands or that the generator be secure in
some way. This not convenient for a personal
portable device like a flashlight. See page 13.
The coil is wound directly on the inner housing
using 30 Ga. 0.254 mm (0.01", 10 mils) magnet
wire. Refer to the coil form drawing on the next
page to follow the calculations used to estimate
the coil resistance. Fig. 23 — FFL coil form width wound with 30 Ga. wire.
Turns per layer: 0.984/0.01" = 98.4 t/l, use 99 turns per layer.
Thickness of layers: 0.335 - 0.078" = 0.257". 0.257/0.01 = 25.7, use 26 coil layers.
Total turns = 26 x 99 = use 2,574 turns.
Page 10 of 14
Fig. 24 — Coil leads taped into inner housing groove. Fig. 25 — Coil wire layers allow counting turns.

Outer turn diameter = 1,378 - 2(0.0787) = 1.2206"

Inner turn diameter = 0.748"
Average turn diameter = (1.2206 + 0.0787)/2 =
0.984" average turn diameter.
Average turn length = C =πD = 3.1415 * 0.984 =
3.092" average turn length.
Total length of 2,574 turns = 2,574 * 3.092 =
7,959" or 663 total winding feet.
Extimated coil resistance = 663 * 0.1037 =
68.8 Ω.
A measurement check using four different meter types provided values from 50 to 63 ohms. Using a DC
supply and passing 10.1 MA current with a measured voltage of 0.700 volts makes the resistance 69.3
ohms. Why were the various meter types inaccurate? They use either AC or pulse waveforms to make
the measurement. The coil has a substantial amount of inductance and capacitance (reactive components)
and this caused the error. Using an LRC meter the inductance was measured as 61 mH and the
capacitance at 398 nF using 120 HZ.
The coil was disconnected from the circuit board and an oscilloscope was connected (no load) to see the
voltage. A 68-ohm resistor was also connected across the coil. Figures 26 and 27 show the results.

Fig. 26 — Waveforms of voltage generated by magnet Fig. 27 — Coil voltage with a 68-ohm resistor load.
being shook in the coil - no load on coil. See text for waveform discussion.
Page 11 of 14
The waveforms will be dependent on a wide range of factors such as the rate the magnet passes through
the coil, the time it takes to stop and reverse, the strength of the magnet, the inductance and capacitance of
the coil, and the number of turns in the coil.
The LED used needs about 3 volts for a minimum brightness so the capacitor must be charged to a
voltage greater than three volts. This is why it takes longer for the first charge of the capacitor. The
capacitor will take all the energy the coil provides as long as the voltage does not exceed it's rating of 5.5
volts. If you study figures 26 and 27 you will note that when a load is placed on the coil its voltage drops.
A typically charged capacitor measures on the order of 4 plus volts. If, however, you monitor the voltage
cross the capacitor you will find that it "wonders around" a great deal. Once it settles down it may remain
steady for long periods of time. This is measured with the LED turned off. If you shake the magnet the
voltage will rise with each shake and it may go to 6 volts or higher quite easily. I even went as high as 8
volts before stopping. If you stop shaking when you hit six plus volts it settles down to about 4.7 volts in
15 seconds. When the voltage hits 8 volts it settled down to about 5.3 volts. After 12.5 hours the voltage
was 3.86 volts and the LED worked just fine. The LED "glowed" for over 20 minutes and the voltage
dropped to 2.52 volts. There was just a very faint light at a voltage of 2.39 volts. The voltage was still
slightly above 2 volts after 72 hours. The meaning of capacitive exponential voltage decay is clearly
illustrated - as is low leakage.
F - Rear Bumper: The Rear bumper snaps into
a plastic retainer as I have shown in figures 28, 29
and 30. Note the three notches in the retainer as
shown best in figure 29. These notches must be
aligned when the inner housing is slid into the FFL
case to align the slide "switch" magnet. Use a No.
0 Philips screwdriver for this task. Use caution
when "playing" with the magnet. Note the shake
sound after assembling. If the black lens retainer is
not very tight the inner housing will move. I didn't
want to tighten mine to the original tightness so I
placed a small piece of rubber in the bottom to take
up the slack. I can then open my FFL to
demonstrate it easily and I don't crush the "O" ring.
Fig. 28 — Rear bumper with black retainer.

Fig. 29 — Retainer is held with two screws. Fig. 30 — Remove retainer to slide magnet out.

Page 12 of 14
If you examine figure 28 (and figures 10 and 21) you will notice two holes in the inner housing at the
magnet travel extremes near each bumper. These allow air to flow freely as the magnet zips back and
fourth. If you cover these holes with tape you will notice a significant difference in the speed of the
magnet and the sound it makes when you shake it. I didn't have time to record the waveform traces to see
how they are affected by this damping.
Final Observations And Conclusions: The exploration of the Forever Flashlight was a fun and
interesting project. It provided an opportunity to explore several technical areas and learn a little about
flashlights. This is more fun than just exploring the Internet.
Opening the FFL required that it be held in a vice (wrapped with a cloth to avoid nicks) and turning the
lens retainer ring with a channel lock pliers. When the product is assembled the internal housing is held
tightly in place by the "O" ring pressure. You will notice a different sound when shaking the FFL if you
don't retighten the retainer ring fully.
I would like to see a smaller version, one that could fit into your pocket. Another consideration is a
magnet lock to retain the magnet. This could also be part of a magnetic shield that would avoid the
picking up of small ferromagnetic metal objects (as shown in figure 22). I held the inner housing of the
FFL LED up to slide the magnet to the bottom. I then put a nut (Fig. 22) on the magnet. The magnet was
easily held in place so it wouldn't rattle around while I was making tests and measurements on the FFL.
A friend pointed out the fact that the Excalibur is not the only shake powered flashlight, and not the first.
Check the following URL's for additional information.
http://www.nightstar1.com/ http://www.nightstar-flashlight.com/
http://www.nightstar2flashlight.com/comp.htm
These URL's give a great deal of technical information on a very similar product. Perhaps Excalibur is
licensing the design from Applied Innovative Technologies Inc. They list two patents and compare their
product with "others" including the Forever Flashlight. I also found - I haven't made a serious search - a
product called Mr. Dynamo at: http://www.myspace.com.tw/
This is a hand squeeze LED flashlight that gives 5 minutes
of light with 15 seconds of squeezing. If I find one - the
URL is a wholesaler (see Global Sources URL below)
and 3,000 pieces is their minimum order @ $3.30 each - I
will explore that one as well. I have a Russian made similar
model (all metal) that powers an incandescent bulb.
I also had intended to include a few ideas for modifying the
FFL (and making more electrical measurements) but that
will have to wait for an Appendix B. At least this
description will give the curious reader some technical
information about the product. Appendix A is the Excalibur Fig. 31 — Hand squeezed LED flashlight.
FFL Instruction sheet.
http://www.globalsources.com/am/article_id/9000000048123/page/showarticle?action=GetArticle
Two questions remain unanswered. (1) Is the voltage (>8V) applied to the capacitor (5.5 Volt rating) an
issue? Perhaps the 5.5-volt rating is conservative, or it self heals under these conditions. (2) Is the LED
being over stressed without a current limiting resistor? Perhaps this is built into the LED itself.

Comments and additions welcome. March '04 Richard J. Nelson, [email protected] R1

Page 13 of 14
 Appendix A - FFL Instruction sheet
Below is the FFL instruction sheet (1/2 sheet, folded). The four pages are arranged in page order

Page 14 of 14