A Pendulum should really work

[conradelektro]

An other idea:

A unbalanced flywheel is fixed to a table and the table rests (swims) on twelve piezo elements harvesting the electricity from the vibrating table.

The flywheel could turn very fast and does not have to be very unbalanced to induce vibrations into the table (20 Hertz = 1200 rpm). May be there are higher frequency harmonics building up in the table if the table is 2 mm steel.

I imagine a 250 mm diameter flywheel made from 10 mm acrylic with one heavy magnet, which is for the pulse drive and is at the same time the mass for the imbalance.

Greetings, Conrad

[elgersmad]

It'll work.  But, you'll need to bolt down your generator or it'll be walking all over the house.  How many vibrating pets do you have?  How about sticking that eccentric wheel on a plywood dog cutout?  One big angry chuaua.  The imbalanced wheel is a good idea.  If you use really good bearings and solid mounts it might even last awhile.  Given a push, wouldn't take much energy to keep going.

[conradelektro]

Attached a possible pulse drive circuit with piezo elements.

The reed switch is not a very good solution, but will work with very low power. For proof of principle it will suffice.

One can add more piezo elements, each one needs four diodes. Stacked piezo elements might be connected in series?

The pulse drive could also be used for a pendulum.

In general (imbalance wheel or pendulum) the power from the piezo elements should not be enough to drive any setup as a perpetuum mobile (according to accepted theory of everything).

I will go hunting for parts, just for fun, it is weird enough to be built. Should not make a racket as long as the flywheel is driven rather slowly.

May be someone has an better idea how to drive an imbalanced flywheel with piezo elements. Better circuit to harvest the energy from the piezo elements? Other drive method?

Greetings, Conrad

[elgersmad]

I would use a higher impedance circuit.  The reed switch uses power, like anything else.  But, MOSFET OpAmp would use less power triggering Power MOSFET, MESFET, or IGBT.  Battery saver style ICs can do a better job and use less power, even when the MOSFET kicks in to dump the capacitor.  A Piezoelectric Element works as a source of AC.  When you compress electrons move in one direction, as it decompresses electrons go back to where they came from.  So, every element requires a bridge rectifier.  They have no real polarity even though they often connect red and black leads when you're generating power.  It's like a sponge filled with electrons, they leave one side of the disk and go to other when you compress it.  The electrons return to their initial starting point when the disk decompresses.

You have a bridge rectifier and connect the cheap piezoelectric disk too it.  Then you'll get a pulse of light from a single LED when you press on it, and again when you let go.  Just a simple bridge rectifier experiment will reveal that.  Just press and keep the pressure on it, then suddenly let go.

Good start though.

The importance of polarity really applies to producing sound and which direction it can move first.  Most piezoelectric materials must be compressed first.  So, which direction the electrons move is important when you are driving it to produce sound.
This MOSFET

This OpAmp

Then use 2 Zener Diodes in series to control power on after the bridge rectifier.  Then it goes even farther to detect the right voltage based upon a threshold knob.  Open Loop Gain.  When it's on, it's on.  The output of the opamp directly drives the MOSFET's gate.  The capacitor should be a large value and feed a capacitor that holds the ideal charge for a transformer/converter.  So, if we want 10 volts out of the transformer, if we want 30, we can change the turn ratio and have 30.  So, first it fills a fat capacitor to keep the spikes away from our circuit.  Then it fills a small one that will produce the pulse width we want.  Capacitor in parallel, resistor in series with the load, pulse capacitor, zener diode pair in series and parallel with the load.  Then off of the low zener diode the opamp is biased and above the high is the full voltage we want for the MOSFET about 30 volts.  You don't want to dump anything off but if it's just pendulum that should work.

You'd need a higher voltage MOSFET for the eccentric wheel and motor.  That would do allot more faster and may require some effort in splitting up the outputs of the piezoelectric disks to produce a split supply and operate a Bridge H MOSFET Circuit.  FM style relatively high frequency.  On that one, I would really need to sit down with a meter, measure the voltage and current to think about which style of power up I'd want to use.  Basically, when there's not enough, the zener voltage divider prevents anything from getting power or starting up.  The opamp doesn't take enough to hurt that.


If you wind up hearing the circuit, each pulse should the same width when the circuit fires a pulse for the transformer.  The resistor between the large capacitor and the small one is chosen to produce a square wave with a 50% duty cycle maximum.  Each on pulse is always the same width.  The off pulses vary in time with the amount of power produced by the piezoelectric disks.  As the pendulum starts from it's highest point and falls to it's lowest, the frequency of pulses will rise, then when it starts riding back up, it will drop in frequency to almost clicking.

[conradelektro]

@elgersmad: sorry, I have difficulties imagining a circuit from words. May be you could draw your circuit and post it.

I am afraid, the piezo elements are not able to power an OpAmp. The one you mentioned has a
Quiescent Current per Amplifier of 25µA at 6 Volt resulting in 150 µW.

150 µW is a top performing multilayer special production piezo element for "energy harvesting". What we will see from inexpensive run of the mill piezo elements is less (hopefully 10 µW).

This is the reason why I came up with the reed switch. The energy for actuating the reed switch comes from the permanent magnet of the pulse drive and does therefore not have to be provided by the piezo elements. All energy (besides the loss in the capacitor) goes into the coil of the pulse drive when the reed switch closes (when the magnet has a certain position near the coil).

I would switch the pulse drive in a "push away" set up: when the magnet is exactly at the mid point (when its natural attraction to the core of the coil is a maximum), the reed switch should feed the electricity to the coil in order to push the magnet away (in the direction of rotation). The reed should stay ON till the magnet is well away from the coil. In case of little power fed to the coil, this means, the magnet is less attracted to the core of the coil than before and can therefore leave the sticky point with a little net gain. The capacitor will be exhausted completely. There will be no electricity left for any active component.

Keep in mind, piezo elements show a "static electricity phenomena" which means rather high voltage (up to 50 Volt), but very very very low amperage.

The US-army wanted to produce shoes which generate electricity when walking by help of piezo elements. Well, they gave up. One had to walk for hours to produce very few Milli-Ampere-Hours.

Attached see a photo of one of my pulse motor experiments. I used toroids instead of coils. The magnet is attracted to the ferrite material of the toroids, but when current is driven through the windings on the toroids, the magnet "does not feel" the ferrite material any more and can leave the sticky point. I like to call that "attract the magnet and then hide at the sticky point". This allows to drive such a pulse motor with very little power (but it will have very little torque). I achieved 800 rpm with 0.16 Watt (no load, just free run). But with the piezo elements it should work with 0.000016 Watt (16 µW), which is to be doubted.

Greetings, Conrad