The Tom Bearden

Help support the research


Subject: RE: MEG in process.
Date: Tue, 20 May 2003 09:57:15 -0500




You're getting the idea; there is real power to be developed by these units in small packages.  But the phenomenology also is rather fearsome, and much of it is not in the textbook.


The units are indeed highly nonlinear.  High nonlinearity brings in a host of effects one does not have with a normal transformer.  Such as reducing the input to increase the output, in certain situations.  Such as sensitivity to initial conditions (be sure to turn it off and turn it on in the same way each time).  You can have unanticipated interactions between your switching units and the rest of the unit, including both input and output at the various sections.  The nanocrystalline core material seems to be quite sensitive to moisture and corrosion, with exposure to moisture having essentially destroyed two of our own units.  Once the nanocrystalline core corrodes, it's of no use except for a paper weight.  So we will have to find a good way to seal these cores.  Otherwise, in a fairly humid climate like Alabama, they tend to gradually degrade due to corrosion.   In their commercial use in normal transformers, of course, they are usually immersed in oil, so hydration corrosion is not a problem.  But in experimenting with open cores on the bench, on really rainy days etc., it is indeed a problem.


There are some other effects we're being very close mouthed about at this time, since we are still in the patent-filing process on some aspects of the units and on our solutions.  All I can say is to please check the relative and simultaneous timing aspects of all the coils and switching parts of the system, very rigorously.  Be sure you are okay on that.  You will find you can have multiple inputs not predictable by normal theory, and they can easily be subtractive rather than additive.  You must insure the phasing so that you maximize the multiple inputs as additions. Else your unit can literally pit one part of itself against another, etc.  So if you can beg, borrow, or steal an 8-channel data sampling and storage oscilloscope, 300 to 500 MHz, with the software on it to already integrate under the curve etc., get it.  It is really, really, really needed.  And as you know, the quality of the probes and the use of the proper probes are just as important as the oscilloscope itself.  Sometimes even more so.


Our own best guestimate at this time is that the little units we ran at 25 watts are actually capable (with modification) of producing a kilowatt at least, and more probably 2.5 kilowatts.  Once we have obtained development funding, our plan is to develop first a fundamental 2.5 KW unit with synchronizer, so that up to 6 units can be added together and sync'd.  That covers from 2.5 KW to 15 KW and it will require one year of very hard work from where we are right now.  A year after that, we would hope to have a basic 10 KW unit with synchronizer, for covering 10KW to 60KW.  And so on.  The extreme nonlinearity of the phenomenology means that one must allow sufficient time to do extensive exploratory phenomenology at every new scale-up stage. It is certainly doable, but it is also certainly not simple or easy.  It's somewhat comparable to doing some of the nonlinear work that goes with a new re-entry vehicle with heat shielding, etc.  And it's complicated.  The ideal development team necessary must have specialists in geometric phase, nonlinear oscillation theory, nonlinear oscillation control theory, math modeling in a higher group symmetry electrodynamics (can't just use electrical engineering; it's totally inadequate and it doesn't describe the phenomenology), etc.


But be prepared to wrestle mightily with nonlinear phenomena, the theory of nonlinear oscillation, and the theory of nonlinear oscillation control theory. These areas are quite different from their normal linear oscillation and linear oscillation control theory counterparts.


Best wishes on your project,

Tom Bearden

Date: Mon, 19 May 2003 21:11:37 -0700

Greetings, and I hope you are well.   
I've succumbed to the MEG fever and have ordered materials, based on JLN's BOM, and expect to be winding coils in a few days.
Item I.  I  expect delivery of a 1" x 2" x 1.375" (pole to pole) Neo magnet which purportedly has a rating of 5350 gauss and a pull of 170 lbs.  Assuming this to be gauss per sq. cm., this calculates to a total field of 69,030 gauss.  My references indicate that 1 ampere-turn will yield a 1 gauss field and this implies that with one turn my magnet will have the equivalent of 69,030 amperes flowing!  (And, with no heat!  Certainly a case of super-conductivity at room temperature?)  So, in that 69,000 amperes flowing for one second is that many Joules, or watt-seconds, then in one hour we have 69 kw-hrs.  Is this not the theoretical potential of this PM?
In your MEG writeup you describe getting 25 watts output using a 5 watt input, and that the input switches the PM  flux from leg to leg of the MEG.  If this COP ratio of 5.0 is valid, why not input 1,000 watts and get 5,000 watts output, etc.?   Are you not switching only a tiny fraction of the immense field of the PM (about 0.04%)?  But, if the PM is loaded more heavily will it not begin to reflect the loading?  Perhaps by losing some of its magnetic strength due to counter EMF?  Or, will the field simply be diverted back into the air and not affect the magnet at all?
Item II.  If my electric bill is $240/mo., at $0.08/hw-hr, my use was 3,000 kw-hrs total, or about 4 kw-hr/hr, on average.  My peak use may be several times that and my low use may approach zero. Yet my MEG may deliver 6 kw at a constant rate.  Do  you envision a (from the beginning) contract with the El. Co. to accept the excess of the MEG output and to supply my peak needs?  Or will you counsel me to smooth out my usage to keep the MEG capital costs at a minimum, and then setting the MEG to automatically supply my needs on demand, up to its limit?
Item III.  I plan to set up my  MEG with the coils mounted in bobbins on the legs of AMCC-320 core (not the center regions) so that they may be manipulated a bit.  Once I have a working model, I plan to optimize it, as built, an to then increase the drive and test my "more input-more output" premise.
Your advice and comments will be appreciated, and I will certainly keep you updated.