Common batteries are free energy sources

[mscoffman]

@All


Erratum;

1N914  (Do-7) = 1N3148 (DO-41) is silicon diode not Germanium  !!

1N54 !n60 1N270 are usual Germanium diodes
PESE-SEMICONDUCTOR (since 1964)
Gustav Pese

True...I just checked - sorry, I was thinking of the 1N54 RF signal diode. The 1N914 is std silicon 1Vdc forward voltage signal diode MAX If = 200ma.

Thank you for the correction Gustav

:SMarkSCoffman

[mondrasek]

I guess I'll be slowing down on this for a bit.  I can't see any way around redesigning and making further extensive changes to the relay.  Currently the relay still has it's original steel armature plate on the back of the pivoting relay contact arm.  It is to that plate that I have my little neos attached to allow the air coil solenoid to force a push-pull oscillation as it switches between +9 and -9 V.  Unfortunately there is probably attraction of the coil to the steel plate as well, and that never switches to repulsion.  So that side of the switch will always see an additional attraction and I will not be able to achieve 50-50 switching.  The plate has to go.

I'm torn between just taking the baby step of replacing the steel armature plate with a similar one of plastic or wood or going all out with a redesign that incorporates more improvements.  Some things I was considering:

1)  Make the attenuator plate rise up past the pivot for a considerable distance and attach a weight that can adjust up and down to set the frequency.  Right now the "counter weight" I've been using that works best is still the alligator clip attached to the old spring tab.  I had tried gluing a thin steel strip to replace the alligator clip and give me something to attach weights (neos) at different heights to set different frequencies, but it was too flexible and actually lagged the switch arm, causing bouncing.
2)  Raise the pivot so I am not working on such a small scale.
3)  Spread out the contacts so there is a longer gap in the switching so the traces can be better analyzed.  I don't think the dead time in the switching hurts anything so why not slow the whole system down?
4)  Maybe move the solenoid to the other side of the pivot to spread out the components further.  There might be some benefit in changing the distance from the pivot and magnet/solenoid arrangement that is difficult to do in the tight confines of the current relay set up.
5)  Maybe flip the relay so the contacts are up.  Right now gravity pulls the switch arm away from the contacts, into the dead zone of the switching.  Inverting will cause gravity to pull it to one side or the other.  I believe one of these types of arrangement I can help it self regulate the 50-50 split by using gravity to assist if one set of batteries is creating a weaker magnetic field than the other.  But I need to think that through some more.  Maybe it is better the way it is now?

I'll also need to order the Schottky diodes.  Don't have access to any old Computer PSs that I know of.

One other test I thought to do yesterday was to place a 200V 6.8mF cap across the high voltage side of the transformer/rectifier to see how high the spikes would charge it.  It leveled out about 59.8 V.  I also ran a small DC motor from the same set up.  It ran slooooow.

M.

[mondrasek]

The Schottkys arrived and I've built a new "Franken-relay".  I replaced an original relay's steel attenuator plate with model plywood (1/16" x 2 laminates) and left it long enough so I wouldn't need an alligator clip on top to stabilize the oscillation.  I also removed the steel back and top plate for the solenoid to eliminate other Lenz losses.  I checked everything with a magnet and am 99% confident the relay is now non-magnetic except where I introduce the new permanent magnet to throw the switch plate from side to side as the solenoid oscillates from +9 to -9 V.

I was very supprised at how hard it was to tune all the mechanical variables in this new solenoid to even get it to run.  I guess I got lucky with the first one.  Now that all the unnecessary magnetic materials are removed I can see how much the conductors between the common input and the switch plate contacts act as a spring.  I had to tweak the bends in those connections for about 15 min. before it would run reliably.  I'm guessing this will be a big obstacle in getting a 50-50 switch cycle.

The Schottskys are a good improvement over the bridge rectifier from RadioShack.  I tested both by using the high voltage output of the 1:1 transformer through them in bridge rectifier arrangements to charge a cap.  The Schottskys took the cap to a higher voltage over all and ramped up much quicker than the RS unit.  Thanks MS.

"Race conditions".  That appears to be the latest hurdle.  Can't feed back directly to the batteries because the spikes occur at the same time as the switch is opening and creates a short?  Damn the speed of light!  This may be a stupid question, but I'm guessing there is no lossless delay circuits for voltage?  If not, then I'm stuck with charging an output source and not the input directly.  May still be able to feed that output source back to the input, but with more complexity and therefore more losses.

Assistance requested.

M.

[mscoffman]

"Race conditions".  That appears to be the latest hurdle.  Can't feed back directly to the batteries because the spikes occur at the same time as the switch is opening and creates a short?  Damn the speed of light!  This may be a stupid question, but I'm guessing there is no lossless delay circuits for voltage?  If not, then I'm stuck with charging an output source and not the input directly.  May still be able to feed that output source back to the input, but with more complexity and therefore more losses.

Assistance requested.

M.

@modrasek;

An engineer would generally separate a circuit into two parts. The logic part
where he creates the circuit's behavior and the power part where he uses
enough components to accomplish what he needs in terms of power processing
then he links the logic part of the circuit to the power part with buffer amplifiers
called drivers.

You are trying to combine the two and have the whole circuit implemented
from power components - this would be called "functional overloading". It
makes things more complex and tricky to analyse and difficult to modify
because of sneak paths, etc.

With the *above* I wanted to show how this would be done in the real world.

(a)With this in mind one way to do what you want is to use a
capacitor to bypass the b-emf pulse through it to the battery, the
capacitor will pass the pulses but block the dc. You can use a series
diode to select pulse polarity and parallel resistor to 'slowly' drain or reset
the capacitor to a base dc state. This is tricky to analyse and some
calculations would be necessary to select components - then scope probing
to make sure it operates the correct way.

(b)In reality you would like to treat the batteries symmetrically. You
could do this by having two coils, one half the circuit would fire into
the static part, then pass control to the other half where it would
perform the same function on the now static first part. With what we
referred to above, an engineer would use a logic flip-flop to guarantee
correct circuit operation, rather than having one coil fire the other
somehow. This way the batteries would be treated exactly
symmetrically with respect to time.

---

web link to latching relay circuit;

http://www.discovercircuits.com/circuit-solutions/latching.html


Finally; I found this circuit design on the Web using FET transistor logic
and a latching relay I wanted to show you. Latching relays are not all
that common of a component that use two coils to pull the contact "off"
or "on" then stays in that state forever until the next pulse comes along.
It shows that a 25ms pulse has enough energy to relatch the relay to the
correct state. Some relays use only 75mw so you can multiply the 25ms
duty cycle times 75mw and you get a decimal point followed a number of
zero's after it representing continuous, per second, power dissipation.

The illustrated circuit uses 3volt batteries and a capacitor as a one
stage voltage multiplier to get 6Vdc= 3volts from the battery plus 3volts
from the capacitor switching a 5volt latching relay. The 25ms input logic
signal will cause one 6volt pulse just long enough to latch the latching
relay into the new state.

The way this would be used in our case would be to issue a setup pulse
to alternate sides of the relay(s) maybe once per minute or so. then issue a
number of recharge pulses while keeping the latches in that same state.
Then hitting the latching relays to swap battery state.

What I would probably start with is 5 double voltage pulses into parallel
batteries then one with b-emf pulse then 5 more double voltage pulses.
Then a little idle time then repeat this for one minute and then hit the
latching relay to swap the battery states. If that doesn't create overunity
in the batteries, I don't know what will.

Mondrasek; Maybe you can combine some of these thoughts with your
own and have it satisfy your needs somewhat easier or with more certainty
then with the first asymmetric relay approach.


:S:MarkSCoffman

[mondrasek]

@MSC

Functional overloading?  That sounds awfully negative.  In my line of business being able to accomplish multiple tasks with a single component or system is "design optimization" and considered a good thing!  Maybe it is that type of thinking that got me playing with the relay as both the switch and load.  But I agree that has complicated the design, construction and testing most likely.

I realized late on Saturday that I have two unused contacts in that relay that might be usefull for feeding voltage back to the batteries.  One NO connects to the + of one battery directly and to two others, but with the solenoid in the way.  And one NC that connects to the - of three batteries.  I was beginning to play with that when I lost stable operation of the relay.  As I mentioned earlier, the wires in the relay drag on it like a spring so it will only run if they are bent just right.  That is the biggest lesson I think I learned from this second relay build, so I went after building the third generation to eliminate the wire drag.

Relay #3 has the supple conductors from the relay switch replaced with #24 magnet wire that can be bent to hold a shape.  The four leads are exiting the top of the relay cube (top is removed) and are then bent to the sides, then down, and then back out to the sides, co-axial with the relay hinge pin.  I'll leave the co-axial sections about an inch long before soldering them to the wires connecting to base connections.  That way the only spring force the relay arm will see is the torsion on the four magnet wire leads in that area as the relay arm oscillates.  I estimate that oscillation is only through one degree or less so the torsional spring effect should be near nothing.  Relay #3 was just near complete when other duties called, and then Ike dumped our power for the rest of the evening, so I haven't been able to complete and test.

Thanks for the ideas on feeding back through caps.  Could be the simplest thing to try.

Also love that new circuit!  I had not considered latching the relays (or using latch relays), but I had considered building a double coil version of my existing set up.  One asymmetrical concern I have now is that the push and pull forces induced by the solenoid on the neo magnet are probably not equal, since both poles of the neo are in the electromagnet field and are at different distances from it.  My guess is that will result in one direction of force being higher than the other for equal but opposite solenoid charging currents, right?  I had considered next to place a longer cylindrical magnet through a hole in the switch plate arm so that opposite poles were on opposite sides.  These would both be repelled by separate solenoids that would fire alternately as the contacts on each side switched the voltages from +9V to -9V and back.  Diode, solenoid, diode in series, paralleled with an identical diode, solenoid, diode with opposite polarities.

On a side note, I was curious about why Imhotep used a signal diode vs. power diode in his Bedini fan circuit for the collector "output".  Maybe it was just because this was a component easily sourced through RadioShack.  I thought I'd test my fan's ability to charge up a cap through the signal diode vs. my Schottkys.  So I attached a Schottky to the collector as well, and then set the fan to charging a 250V 6.8mF cap from the signal diode first.  It took it to ~220 V (no wonder it bites when you grab it).  I then switch the output lead clip from the signal diode to the Schottky and blew my transistor.  I hadn't thought about the fact that my little power diodes are only rated at 200V.  So that was the end of my test and my second transistor.  Luckily I had purchased two after I blew the first one and had one left to get the little battery charger functional again.

M.