RE: [MEG_builders] MOSFET Coil-Driver for "MEG-2" Experiments
- Hey David,
I read your E-Mail and I thought it might be a good idea to bring this
details to your attention.
I bought a "magnetic viewer film" a few months ago and I played a lot with
it. It gives you an idea on how the magnetic field is behaving by changing
color; from dark green to light green.
I think putting magnet stacks on a MEG is not a good Idea. With the Viewer
Film, I see a great difference between a whole magnet and a magnet stack.
The field is not the same. There is some small flux intersections in the
"joints" of those magnets. Although stacking magnets give you a stronger
general field, I think it should be avoided in the MEG building.
I just started experimenting and building my MEG, but knowing that fact will
influence the way I'll build it. I'd rather buy a magnet that fits
"perfectly" the space in the MEG, or even a little larger.
I'd like to point out a fact about adding "shims" to make the magnet fit. If
we are using a special material as a transformer core, we should avoid
putting a normal iron shims to fit the magnet in place; This would waste the
overall effect, since the system would have to go over those shims.
I also think it would be a better alternative to buy a magnet that it a
little larger than the place you want to fit it and cut small holes into
your core until there is a perfect fit.
I'd like to know if someone had to opportunity to try those configurations
and builds, it would be interesting to know the outcome, "practically".
I'll surely try all those alternatives when I'm done winding all the set.
Keeping you informed guys!
>From: "davidj95650" <djenkins@...>_________________________________________________________________
>Subject: [MEG_builders] MOSFET Coil-Driver for "MEG-2" Experiments
>Date: Wed, 18 Aug 2004 19:50:20 -0000
> This message is to provide some thoughts and details of experiments
>I am conducting and some observations.
> In message #1230 Dave Narby posted pictures of another
>implementation of the MEG by Dr. Bearden and his colleagues. There
>are two aspects of this construction which are particularly
>interesting: The drive coils are very thin and both drive and output
>coils use thick plexiglas as forms. This type of construction results
>in more "leakage" flux, that is flux from the coils "leaks" away from
>the core. In usual circuits, this is a problem because it contributes
>to losses and through Lenz-effects creates large voltage spikes that
>can damage associated circuitry. Thus the use of this construction in
>the MEG must mean that it has some bearing on the performance.
>Dr. Bearden has stated on more than one occasion that the extra energy
>comes from "outside the core". Leakage flux from the drive coils
>which does not enter the core would still be available to wash over
>the turns of the output coil. Because this flux is not in the core,
>and thus its changes are not slowed by the core response, it can be
>made to change very rapidly, inducing a consequent large E-field by
>the relation E = -dB / dt where dB is the change in this leakage-flux
> I have made a bitmap drawing to illustrate this. Go to "Files"
>then go to the folder "MESSAGE ATTACHMENTS", go to the folder "MEG2
>Experiments", and open "LeakFlux.bmp".
> The electrons traveling on the surface of the windings are most
>affected by this dB / dt. Their velocity increases dramatically due
>to the induced E-field. As a result of this velocity, there is an
>additional E-field as electrons travel through the gradient of the
>A-potential which is the result of the magnetic field produced by the
>permanent magnet being retained in the nanocrystalline core and the
>associated A-potential freely existing outside the core. See message
>#1204 for further discussion of this.
> It seems reasonable that a large perturbation of the A-potential,
>if it is to have the greatest effect, should occur when there is an
>abundance of electrons on the surface of the output-coil windings. I
>designed a MOSFET driver to do this. Go to "Files" then go to the
>folder "MESSAGE ATTACHMENTS", go to the folder "MEG2 Experiments", and
>open "Driver1.bmp". All the rectifiers and MOSFETs are attached to a
>heat-sink with suitable insulating materials although there isn't much
>heat developed. Just want to be safe heat-wise.
> This driver circuit stores the high-voltage, generated by the Lenz
>effect when the drive coil discharges at the end of each half-cycle,
>in capacitor CP. Near the end of the next active half-cycle, this
>high voltage is applied to the drive coil, increasing its current
>during a very short interval. This is the time when the output coil
>voltage is at its peak and there is an abundance of surface electrons
>on the windings. To simplify the description, the Lenz-effect voltage
>is typically 400 volts, the length of the driver coil winding is about
>4 meters, which gives an E-field of 400/4, or 100 volts/meter,
>roughly. For an output coil length of 80 meters, if this E-field were
>to be present across the entire winding, the induced voltage would be
>80,000 volts. However, only a portion of this perturbed field reaches
>the output coil, so the effect will be less, but still a considerable
>amount. Another way to look at this is that the drive-voltage is
>normally 24 volts, but when the end-of-cycle pulse occurs, it is for a
>short period of time 400 volts.
> The timing diagram at the bottom of the picture illustrates the
>timing of the MOSFET drive. Just prior to the end of the drive
>applied to the "DRIVE MOSFET", the "PULSED MOSFET" is turned on
>which applies the voltage stored in CP to the drive coil. This
>results in a rapid change in the current in the coil, which means
>a rapid change in its magnetic field.
> I have been experimenting with an AMCC-630 core, which is probably
>smaller than the core in the MEG2 pictures. I'm using a CP with a
>value of 22 nF, which discharges in about 1 uSec when applied to the
>drive coil. This results in an increase of current in the drive coil
>of about 100% (a doubling), as measured with a current probe on one of
>the leads of the drive coil.
> My core has 15 turns of #16 magnet wire on each drive coil, one
>turn per layer, and the first turn is spaced 0.18-inch from the
>surface of the core. The output coils are 240 turns of #18 magnet
>wire, also spaced 0.18-inch from the surface of the core. Three
>columns of 0.5-inch x 1.0-inch NIB magnets are placed at the center of
>the core, and wedged firmly in place with 1.0-inch transformer
>laminations to minimize leakage flux from the magnets. This results
>in a magnet stack that is 1.0-inch wide and 1.5-inch deep.
> The output capacitors are 6,000 uF at 350 volts. The load is a
>variable resistor nominally set at 1,750 ohms for the present
>experiments. The setup is almost identical to the pictures posted in
>message #1230. This latest build-up has several interesting problems:
>it radiates a lot of high-frequency energy which interferes with radio
>reception as well as scrambling a TV which is about 3 feet away,
>operating with a coaxial antenna cable. Also, the control circuit,
>based on a high-frequency power-supply integrated circuit, suddenly
>draws very high current from the drive-supply voltage during the time
>that the high-voltage pulse is being applied to the drive coil. So
>far, heavy filtering, running on batteries, shielding all external
>wires, enclosing the control circuitry in a solid metal box, and
>widely separating the components has not solved this problem. It's
>intriguing to me, as I haven't seen an interference problem this
>severe in my years as an EE. It's also a pain because I can't run the
>setup at full power.
> I haven't been able to duplicate the drive-voltage waveforms seen
>in the MEG2 pictures. My input current looks like a standard ramp of
>increasing current as the drive coil charges. The MEG2 pictures show
>an almost constant current.
> Efficiency is about 70%, nothing to get excited about.
> As yet I haven't tried changing the values of the CP capacitors to
>determine if a larger value will make a difference. In this case, the
>pulse would have a smaller dA / dt because of lower voltage in CP, but
>a longer time-constant which would allow the nanocrystalline core to
>be more involved in the process. This might also solve the
>interference problem in the control circuitry.
> Many things to try, have to find the time to do it.
Designer Mail isn't just fun to send, it's fun to receive. Use special
stationery, fonts and colors.
Start enjoying all the benefits of MSN� Premium right now and get the
first two months FREE*.
- Hi Jon,
I haven't paid close attention (exacting measurements) of the
efficiency when it is so low. IMO when OU strikes, it will make
a large difference because it will build upon itself (the "ping-pong"
effect). However, I could be very wrong, and your suggestion is
a good one. I will finish making my control circuit interference-
proof (since the A-potential can not be shielded, this could be
a significant challenge if it is the cause of the problem !!),
then compare the results using this driver with results
using a full-bridge driver which limits charge and discharge voltages
to the supply voltage. Finding the best efficiency is a time-
consuming process as the frequency, load-resistor, and supply
voltage must be all be varied independently and the results
An e-mailer raised the question of using stacks of magnets where
using one magnet of the right size, or grinding the core to make
a single magnet fit might be a better choice. These nanocrystalline
cores do not respond well to any machining: the layers flake and
fall apart. Also, using a gauss-meter, the stray flux from a stack
of magnets is only a few (less than 20) gauss where the magnets
join. The largest leakage of flux is around the transformer
laminations used to wedge the magnet stacks in place where I measure
up to 400 gauss adjacent to the laminations. As the field in the
magnets is about 10,000 gauss, this leakage represents about 2.5%
of the total field. The leakage measured near the output coils,
where it really counts, is much less, typically less than 100
gauss adjacent to the laminations, and even less farther away.
According to my research, the only purpose of the permanent
magnet is to provide a large static field in the core of the
MEG. Because the core acts like a magnetic sponge, very little
magnetic field is measured outside the core: typically a few gauss
(less than 5). Thus according to quantum mechanics there is a
large A-potential outside the core which is due to the magnetic
field inside the core (to be precise, the B-field in the core is
due to the "curl" of the A-potential in the core). According to
the "Convective Derivative", an electron moving through this A-
potential will experience an acceleration or deceleration,
depending on whether the A-potential is decreasing or increasing
in the area where the electron is moving. Note two important items:
the electron must be in motion, and the A-potential must have a
gradient (a change in value for a change in location). The greater
the motion and/or the greater the gradient of A, the greater the
acceleration seen by the electron. I have shown in previous posts
that this can be observed indirectly by measuring changes in
magnetic field within the output coil (see message #1216).
More work, hopefully more to report.
--- In MEG_builders@yahoogroups.com, "jonfli" <jonfli@c...> wrote:
> Hi David,
> You wrote:
> [snip a lot]
> > Efficiency is about 70%, nothing to get excited about.
> This is an interesting line of MEG experimentation but I have a
> this 70% efficiency with your switched 22nf caps? If so, what is the
> efficiency without them in the circuit?
> Jon F
- --- In MEG_builders@yahoogroups.com, "davidj95650" <djenkins@r...>
> This message is to provide some thoughts and details ofexperiments
> I am conducting and some observations. .....David,
> <<SNIPPED FOR BREVITY>>
> David J.
Hey, thanks for sharing the results of your recent MEG experiments
with us at this site! Great!
I find your driver circuit with a capacitor being connected via a
Vpulse MOSFET to the input coil to be very novel and interesting. As
I didn't quite understand the circuit, I decided to check it out by
an electronic circuit simulator program, doing a simulation for the
basic parts of the input circuit, treating them as inductances,
capacitances, etc. It does indeed appear to add a real kick to the
input coil drive. I was also curious about why you were seeing large
current spikes in the supply voltage rail and in my simulations, I
observed currents for your circuit with and without the pulse circuit
and I found a very definite but very brief current spike, flowing
first to (supportive to)the supply and then away from the supply at
the start of the pulse drive. This spike does not occur in my
simulations of your circuit that do not incorporate the pulse
circuitry. At present I don't understand the reason for this spike
but with time perhaps I will.
While simulations are just simulations and are always questionable,
perhaps this info could be of use to. I will share this info with
you by personal email and perhaps if it checks out, I'll post it to
the site. Will try to get in touch with you shortly.
ANYWAY, thanks for your new MEG driver idea and sharing your MEG
experiment results with us. It's good to see at least one
experimenter sharing his results with this group :-)
- Hi David,
I have found a concept called "hidden momentum", which I am not sure
totally agree with. In a B field free region with an A field, there
a momentum change = qA, for a charged particle (electron) moving in
the region. However there is no force or momentum change on the
electron. This is what I seem to have verified by passing an
electron beam in an CRT through a toroidal coil with DC current.
There is a QM "phase shift", whatever that means.
This was written about by 2 authors Shockley and James, but I have
found a good on-line link. I can email you a paper, if you like.
--- In MEG_builders@yahoogroups.com, "davidj95650" <djenkins@r...>
> I have made a bitmap drawing to illustrate this. Go to "Files"due
> then go to the folder "MESSAGE ATTACHMENTS", go to the folder "MEG2
> Experiments", and open "LeakFlux.bmp".
> The electrons traveling on the surface of the windings are most
> affected by this dB / dt. Their velocity increases dramatically
> to the induced E-field. As a result of this velocity, there is anthe
> additional E-field as electrons travel through the gradient of the
> A-potential which is the result of the magnetic field produced by
> permanent magnet being retained in the nanocrystalline core and themessage
> associated A-potential freely existing outside the core. See
> #1204 for further discussion of this.