b) While I think that max(Voltage batA + Voltage batB + Voltage batC)
is a probably valid way to measure the system state of charge, it is
not a technical proof of the the total energy state of charge
in the system. Just don't get too disappointed if the above is true.

While playing around with a 12V Bedini/Imhotep fan I tried charging NiCd batteries.  I immediately noticed that charging with the BEMF spikes (~300V) would cause the batteries to charge to a higher voltage.  A hicap 9V NiCd (nominal 8.4V) would charge to 9.8V.  An AA 1.5V NiCd (nominal 1.2V) would charge to 1.4V.  I did not charge these for a long time, just until the voltage stabilized, and in all cases they did not run a load as they had almost no current.  But then I could charge them all with a normal NiCd charger and they would charge to the same higher voltage and also have current to run a load.  I never learned exactly why this happened and did not pursue testing to see if the overall energy state of charge in the higher voltage batteries was greater than that in their original lower voltage forms.  My *guess* is that the high voltage charging spikes of the Bedini/Imhotep fan "conditioned" these batteries so that they charged to a higher voltage but probably held the same or less less overall charge.  They have never appeared to run longer than their unconditioned counterparts (actually less) but that is purely a subjective observation.

Reading about high voltage (Bedini type) charging of SLA batteries will lead you to information on "conditioning" of that battery type.  IIRC the lead oxide crystals form finer patterns on the plates leading to a visibly lighter gray coating on the plates than with conventional charging.  I do not remember reading anything that says the peak voltage increases like I witnessed in my small NiCd batteries.  But if they do, you might be seeing a voltage rise with an equal or greater drop in current capacity leading to no overall gain.  I am not sure how you would check to see if your batteries are achieving a higher overall energy state of charge and not just a higher peak voltage.

Good luck with your research.  I look forward to reading your results.

M.

why don't you guys use ammeters to measure how much power is actually being used by the chargin battery and how much power is absorbed by the others.  Or use a device that measures the instantaneous voltage and current every second or so and have it graph V*I of all three batteries.  The energy is the area under the graph.  Judging energy using only the battery voltage doesn't seem very scientific.

why don't you guys use ammeters to measure how much power is actually being used by the chargin battery and how much power is absorbed by the others.  Or use a device that measures the instantaneous voltage and current every second or so and have it graph V*I of all three batteries.  The energy is the area under the graph.  Judging energy using only the battery voltage doesn't seem very scientific.

Don't forget that current is not energy just like voltage is not energy.

You can do this, but it is very tough to do without slowing the time
dynamics down so that the signal doesn't pulse. Removing the pulses can
be done with a buffer - a capacitor. The motor's nature is to be dynamic -
so what you have to do is drain some energy off as DC and measure that
excess. Once you have it, you adjust to drawing off as much as possible
and have the motor still run. This is your estimate which you divide by the
amount of time, known as the integration period, and start integrating
again for the next value.

---

To: Plengo

You really want your goal to be "Science" put before "Finishing
the Project". To do this you want the instrumentation and control
power supplies to be independent of the "core circuit activity"
until you have proven or have an estimate of how much extra energy
has come from the core circuit (/time). Reed relays and opto's can do the
isolation. This leaves the overunity energy intact in the core circuit.
Raw transistors don't isolate (especially bipolar transistor) so for this
part of the project you simply have the transistors drive reed relays,
that do the actual signal switching.

Now once you have proven,the overunity energy and approximately how
much, you can design the control and instrumentation circuitry with merged
power supply with the core circuit that use just the sufficient fraction of
what is being created...that has to take place before the unit is claimed
to be full overunity – but it need not be the first goal. It’s the overunity
of the core circuit that is important, but technically the control and
instrumentation need to be “paid for” from the core overunity eventually.
– leaving user energy.

---

By the way, the source and charge battery that are back-to-back you
might try placing power diodes backward across these batteries. Current in
one direction discharges one battery or the other and the extra diodes
would provide a  “non-discharge” pathway. (If it doesn’t break circuit
operation), that is.

---

It's very easy to convert an old spare slow PC computer, into a
process control computer, and it can be somewhat
easier to program in your favorite HLL,-then it is to program
a microcontroller. The input and output bits can then be read
and written into something called a Port Expander. I've done several
projects in this way and they are a lot of fun. If you can control a process
manually, then it's easy to control it by computer. Let me know if
you would like to do this and I can send you a port expander circuit
expandable to any number of Input and Output signal bits.


---

In the long term - gross battery voltage increases are state of charge
increases, but only in the long term - assuming the batteries are functioning
normally.

Also, if the circuit works as advertised then the control microcomputer
can be used to draw off the excess energy and dissipate it in a load
(while measuring the total amount). This will need to happen
by some method if it is functioning.

:S:MarkSCoffman

Hey mscoffman, great advices, thank you.

Can you post a quick drawing of what you mean with the diode backwards?  And yes I would accept your help concerning that PC communication "expander" and how to do it. It would really accelerate my project since I can only switch automatically today using a function generator and a relay and it is under a fixed timed pulse which is not at all optimal.

Fausto.

See if I can do an ascii(or whatever diagram)

    Charge                     Source II
             |                        |
<----*---| |------*------|  |-----*---->
       |     |           |            |       |
       | (-)   (+)      |    (+)       (-) |       
       |                  |                    |
       *------>|----*------|<------*
              1n4001           1n4001

The diodes blocks each of their own batteries but lets
current flow around it when the other one is driving.
This is especially important so that the charge battery
does not get hit with a drive pulse. The circuit will definitely
function differently though...In Bedini circuit the charge
battery is opposite polarity then the Souce because BEMF
(back from the wheel) is reversed. If the both don't
work one should then the other.

---

Ok, good! On the expander, I usually use PAL programmed array logic drive
circuit but I have done a pure 3 TTL IC integrated circuit version. An old PC
usually has RS232 signal connectors, which I use. Newer PC's would require
an available USB to RSR232 converter dongle adapter.

:MarkSCoffman

Fausto:

I read through your thread and just want to mention a few things.

In this circuit B1 never gets charged through D1.  When I see a schematic diagram that is not too complicated I try to look at it and work out how it works in my head.  I think that is a good thing to do for everyone that wants to experiment with your circuit or simply follow the thread in detail.  I will leave it as an exercise for those that are curious enough to see why B1 never gets any charging energy from the spike.

When you do get a spike going through D1, the energy goes to two places, but it's hard to say precisely what the proportion is between the two.  The first place is the B2-B3 battery combination, where the spike will result in a small discharge of B2 ("helping" the spike), and a small charging of B3.  The second place is through D2.  I assume that you would consider it to be undesirable to discharge through D2 because that represents lost spike energy that doesn't charge any of the batteries.  That is another investigation - trying to determine if D2 is "stealing" your charging energy from the pulses.  Why do you have D2 in the circuit anyways?

For your microcontroller or PC-based control system, why don't you just keep it simple and just use it to program switching intervals in software instead?  You can do any timing you want and if you assume the system can demonstrate over unity, it has to arrive at some sort of steady state.  This implies that you don't really need to measure the battery voltages.  Of course you are already doing that with your pulse generator setup, but changing the timing for the switching over to a microcontroller or a computer would be much more "elegant."

Someone mentioned monitoring the fan speed with the sense line and that is a fantastic idea.  I am just not sure if you will still have a working sense line any more because you destroyed the motor controller board.  You could easily make an alternative sense line by picking up a signal from the spike itself, and connecting that to the microcontroller input.  Certainly the microcontroller in conjunction with it's on-board timers and the development software (or PC equivalent) could make an extremely accurate measurement of the fan speed or the pulse frequency of the spike, and this is indirectly giving you the net voltage across the three batteries in series.  However, I have a feeling that you want to measure the individual battery voltages.  If the microcontroller has an on-board A/D converter you should be able to use that to measure the battery voltages quite accurately through signal conditioning.

MileHigh

Fausto;

I think you are wasting your time trying to control your Bedini unit by
dead-reckoning with fixed time periods for on-and off. I can pretty much
guarantee that the overunity in these Bedini machines is a *process* that
needs to be controlled via monitoring. Fausto, if you complete this project
I promise you, based on the probabilities of what I have seen about things,
the world will not soon forget you.

You see there are two different approaches to this problem, one is giantism
– Build a giant machine, because the overunity process then becomes a
dominant process. The problem with that is it becomes potentially dangerous
and it no longer fits on a desk and expensive to do, requires space to build
etc. The better approach is to use one of monitoring and control of a
desktop size system, by inducing sufficient efficiency into the device’s
operation. That requires instrumentation expertise, which is what you are
developing. Believe me, once one understands what one is doing
systemically, safely scaling it up in size will be no problem. Any size
system will require rotation of batteries to prevent HV from burning-in to
a particular position, and therefore levelizing wear among batteries.

---
   
Also, those back-to-back polarity canceling acid/lead batteries are a
complex system and not easily analyzable. Their net voltage will go above
and below zero. Any static charge would move around between them in an
unpredictable fashion. These could emulate an ice-scraper by boosting one
voltage above the other, in a kind of a self regulating mode, they are not as
simple as most other components connected with back-to-back polarity
appear to be.

---

I had to rework my RSR232 Port Expander interface because I had forgotten
some drawbacks that I previously experimentally removed. But I am slowly
completing the documentation of it. It is a very good demonstration for why
one uses PALs Programmable Array Logic and why for an IC integrated circuit
designs above two IC’s…Microcontrollers often make more sense.
 
---

Below is a link to the datasheet  for PIC microcontroller you mentioned. The
PIC16F84;

http://ww1.microchip.com/downloads/en/DeviceDoc/35007b.pdf

---

Below are weblinks to a number of advertisements of inexpensive
commercial DVM’s that output their readings to an RS232 serial port as
well as their internal LCD display. They are generally less than $100. They
have a “gate” time of about one reading per second. (Having two of these,
one for voltage the other for current would be what is required for a formal
proof of overunity energy production for any device.) Several have built in
frequency meter setting. (ie. for RPM measuring)…Nice.

http://www.electronic-supply.com/DMM-page.htm

http://www.radioshack.com/product/index.jsp?productId=2103962

http://www.cs-sales.net/veindimuwida.html

http://www.apogeekits.com/multimeter_dvm345di.htm

http://www.jameco.com/Jameco/Products/ProdDS/1581214.pdf

http://www.jameco.com/Jameco/Products/ProdDS/137462.pdf

http://www.jameco.com/webapp/wcs/stores/servlet/ProductDisplay?langId=-1&storeId=10001&catalogId=10001&productId=137462&

---


:S:MarkSCoffman

Thanks MileHigh,

I already been through that road with Bedini and I do know if it works or not already too. But again, thanks for the good advice.

Fausto.