Formula1, I did get it to some what work but it did not work as I hoped it would. I need to build different coils to accommodate the action. The coils I used were those that were part of the flashlight and they are spread apart to far.
 I got off in to something else and I haven't had time to get back to it. This project I am working on now is the real deal and I cant wait to get in finished. It's a pulse motor unlike any other I can find. I was going to post my findings but I am not ready yet.
 

... yeah, looks like half the companies selling "shakable flashlights"
are not actually selling the version with a magnet in it,
but rather a version where smart use of an inductor coil and shaking action
can extend the battery life significantly.
I recently bought a real magnetic inductor shakeable torch at www.magictorch.nl
which is just a typical "Faraday flashlight" like www.everlifeflashlight.com also
sells them.
A friend of mine claimed he had a much better one that was cheaper too;
but something didn't feel right when I shook that one, and when I pulled out
my neodymium magnet it didn't "feel" any magnetic field in my friends torch...
... so we opened it up, and indeed it did not contain any magnet at all, it
contained something that seems to be an iron rod inside a coil, and two
small lithium batteries. Haha, the joke's on him.

Here's a piece of text ripped from www.skinz.org/shake-flashlights/shake-flash-lights.html
which describes the two different types of shakeable torches quite adequately:

(Type 1) Non-Rechargeable Shake Lights w/Lithium Batteries and LED Bulbs
The Type 1 Shake Lights are actually not rechargeable and run on two lithium batteries.  If you know about lithium batteries, they last for years and are used on such items as lawn sprinklers, smoke alarms and cameras to ensure years of backup power.  But the truth is the shake and rechargeable part of it is false and does not work.  But these are still a great value.  Why? 

Think about a typical flashlight that takes batteries.  Most good flashlights take 4 D-size batteries which have an average price of $8.  If you put these new batteries in a flashlight and turned it on how long would the power last?  2 Hours?  Maybe 3 Hours?  And you are out $8 plus the cost of the flashlight.

Why does this matter?  We conducted a series of tests to determine the average 'light-life" of these lights by leaving them on until they went out.  Each light tested lasted a minimum of 30 hours before they got dim and some even over 40 hours of continuous use!

(Type 2) "As Seen On TV" Rechargeable Shake Lights

Almost impossible to find anymore.

Type 2 shake lights look very similar to Type 1.  These are from the famous "Seen on TV" ads and are not really being sold much today.  These do have magnets and recharging ability, But as you use them the working ability diminishes into nothing and you are left with a inferior product in our opinion.  Every rechargeable cell wears out after repeated use.  So at some point they will not work.  Also, magnets lose their magnetic ability when they are moved around, and rechargeable powered devices lose power when not used.  How many times did you charge something, and if not used in a long time the battery is dead?  Same story here.

Do you want to grab your flashlight and have to shake it for 30 seconds before use?  Imagine you hear a noise in the house.... do you want to make noise shaking a flashlight?  NO!  In a tight spot fixing something?  How can you shake this light?  Seems impractical.

I underlined those two parts to indicate what I believe is a nonsense argument against these true Faraday torches;
That true permanent magnets demagnetise as they are moved around is not per definition true, it is only true for
magnets with internal conductivity and typically untrue for ferrite or other 'ceramic' magnets. Yes, obviously time
decay will effect a magnet like it does any phyical material and over the course of hundreds of years even permanent
magnets will lose magnetism, but certainly not fast enough for your shakeable torch to become useless during your lifetime.
And that rechargeable batteries lose charge over time when they aren't used is true... but the same goes for those non-
rechargeable Lithium batteries they use in the other type... So that is a non-argument.
The last paragraph is complete bull. Do I want to grab a flashlight and shake it for 30 sec before use?
If I never have to buy new batteries for the thing? And as an alternative to rummaging around the house looking for full
batteries or a battery charger which would need an additional 30 minutes to charge my empty batteries? Hell Yeah!
Imagine I hear a noise in the house, do I want to make noise shaking a flashlight? No of course I don't, I'll grab
the baseball bat and switch on the full illumination of my indoor lighting, of course! Or do they believe you can scare
the noisemaker away by sneaking up on him with a flashlight? What crock. In a tight spot fixing something?
Also not the typical situation for using a flashlight, is it? Bad arguments in order to try convince you to buy a product
that is not by far as cool as the original.

I bought a couple of these torches cheaply at a car boot sale. They were fitted
with 2032 batteries which do not recharge at all. What sort of capacitor
circuit should I use instead?
Paul.

Hi Paul,

If your torches include the cylinder magnet and coils to be able to induce voltage at all, then I think you have a good chance to build in a schottky diode bridge plus one or two so called supercapacitors.
See this link for explanation and schematics etc: http://www.discovercircuits.com/H-Corner/shake-chgr.htm

To get an insight how much supercapacitors cost I think you may wish to look at Farnell for instance http://www.farnellinone.com/ in your country, then write supercapacitor in the Product Search window at top of that page and study the results. Unfortunately the cheap type supercaps of 5V or 5.5V have got a rather high series inner resistance/impedance (around 30,  75 or 100 Ohms),  though you can consider putting some in parallel to reduce their resultant impedance and increase their capacity for a desired lighting time length. Supercaps of 5V with less than 1 Ohm inner impedance do exist but of course they cost rather more.

Of course you may wish to search for other vendors to find the cheapest supercap source for yourself.

rgds,  Gyula

To get back to that first post in this thread;

That was clearly an attempt to design a similar setup, in which
the magnet is cuased to move up and down inside the "collector" coil
by alternately pulsing an electromagnet on either side of it.
Seems to me you'd be expending at least an amount of energy equal
to that produced in the "collector coil" every time you make the magnet
move by pulsing the electromagnet. And most likely quite a bit more.
Also, seems to me the magnet will get stuck at one electromagnet
as it contains an iron core and the permanent magnet will still be attracted
to that when the electromagnet is no longer active, unless you actually
push it away with a pulse of opposite polarity in the electromagnet...
... which obviously will take additional energy.
All in all, seems like an inefficient way to make the magnet move,
and even if that was an efficient way, you're using a coil to
move a magnet through a coil, using current to induce a magnetic
field which moves a magnet which induces a current... You'd be
better off using the feed current directly.
Or at least, that's what it looks like to me.

Nevertheless, if we do away with that suggested mains input,
and if we were to use a slightly different electromagnet arrangement,
we might be able to turn it into something cool.
So let's say we use a sort of "crystal radio" receiver to "leech"
energy from radio waves, but instead of simply using its coil as
a filter and using the amplitude modulation to feed the voice signal
to our speaker or headphones, we simply collect those fluctuations
in charge in a capacitor (with rectifier if needed).
We could have these charges accumulate untill they reach a certain
minimum value, then allow the cap to discharge (using a Zener diode
for example), and use these periodic discharges to for example make
the magnet move. (By feeding this pulse directly into the coil around
the magnet, thus making the magnet jump up)
We don't need two electromagnets on either side of the magnet to make
the magnet move. We only need the coil around the magnet.
We also don't need to pull the magnet back down, we've got gravity doing
that for us. So we only need to give the magnet a good "kick" up,
it will drop down by itself.
And yes, although we might be able to use electrical charge collected
by using a "crystal radio" circuit, the efficiency of the magnet+coil
arrangement will still be far below unity, obviously.

But now, what happens when we use a second magnet at the bottom of
our cylindrical setup (I am assuming here that we have wound the coil(s)
around a cylinder with the magnet inside it, and the cylinder standing on one end),
which is in repulsion to the moving magnet?
It is obvious that the moving magnet will hang suspended in repulsion above
this bottom "stator" magnet. It is also clear that the magnetic field of this "stator"
magnet will have the same field orientation as the coil does when it "kicks" the
moving magnet upward. So during this "kick", the "stator" magnet and the coil
align their fields.
As the current pulse fed to the coil stops, the coils field collapses, and the magnet
no longer experiences a "push" up, and starts to drop down to the bottom of the coil.
During the entire drop, it will induce current in the windings of the coil. Proper use of
rectifiers or diodes should allow for collection of all of that.

We might be able to increase the output by using a stack of magnets in repulsion,
but that would require a slightly different setup. For one, we'd need a number of
"collector coils" to make good use of the multiple field line orientations, and this would
result in a "collector coil" array that cannot be used as the "kicker" coil anymore.
So we'd need to create the desired magnetic field pulse in some other way.
Perhaps putting a coil around, under, or above the "stator" coil  (or a combination of these)
would allow us to produce a similar "kick" if we pulse that coil...
A second objection would be the increased weight of the moving magnet stack,
which would obviously need more input to actually move up the cylinder.
So the higher output would also mean a higher input, and the stacked opposing
magnets would mean the collector coils and the "kicker" coil(s) cannot be the same.
On the other hand, we may be able to use a second "stator" magnet atop the cylinder,
which should be in attraction to the moving magnet, but with a spacer in between so that
even when the moving magnet is closest to this top "stator" magnet and does "feel"
its attraction, the attraction is not strong enough to actually hold the magnet there against
the pull of gravity, and the magnet will drop down. This might reduce the "kick" power
somewhat.

Ok, perhaps I should have posted this in the "Half-baked Ideas" section??

Well, any reactions, suggestions, or other types of reply are most welcome!!!

Kind regards,
Koen

Probably too much friction all in all, he also could not show it in his videos, that he reall yachives more output than input...
These are different oscillation systems, as it is dependant from one water column going back and force through All the coupled outputs,
than for example the Rick Friedrich coupled LC coils...where there is really coming more power out as has been put in...as each coil can oascillate on its own frequency...

Regards, Stefan.