For all members.
I am starting this new thread as a form of placing what I think is a very real possibility of a working TPU.
More time is needed to be spent on this, time which I do not have, so I am throwing this open to all to give input to what I have found and designed to date. The theory is sound, the problem is doing it in the right way.
Below is the basis of the current TPU, it is a cascaded boost circuit with current and voltage feed back into the coils from the magnetic loop. P1 is a type of class E transmitter feeding a carrier via a transformer coupling into the zero voltage part of the loop, 180º from the tuning capacitor. Voltage and current comes from the cascade, P2 to P3 and on to P1, note the estimated voltages on the schematic and final feed voltage to the loop.
The loop is made of copper, 1/4” or 3/8” copper tube would be good but I expect at the low frequencies we will be running at (up to 30kHz “VLF” rf) the skin effect will not be great and a solid 4mm copper wire would be good. The total length of the circumference (including the coil of the transformer secondary), must be more than 1/8 of the wave length and less than 1/3, so as to have a fairly uniform circulating current in the loop. High voltage will be present at the Cr loop capacitor, where as at the feed point at 180º to the capacitor, the voltage will be near 0v.
The idea is to extract some of this current and voltage from around the loop, through magnetic induction into the coils. The a, b, and c coils will change polarity depending on whether the mosfets are on or off, but the a1, b1 and c1 coils will not. It is envisaged that the latter coils “1” are the ones that will gain current from the loop.
It is hopefully found that no feedback is needed, C3 and 4 will charge from internal feed back, the mosfet drain source diodes doing a dual function along with coil b to feed back into the capacitors.
There is no fixed frequency which has the magic, in my opinion, but I do believe we need to start with a clock frequency, the 4047 ic, pass it through a flip flop, 4013, to create an exact 50% duty, then divide this frequency into 3 phases of 120º apart. The 4017 ic is up to a X 10 divider of which we will use 3 with a down time between each so as the cascade works between each phase, the 3 phases will add up to the clock frequency between 0º and 360º.
I have the outputs going to buffered 4049's which drive my high voltage and current mosfets. It would be better to use dedicated mosfet drivers from this point on, but the mosfets I am using seem to work ok without. The IRFPC50 only needs a simple gate drive and is what I had on hand at the time. C2 in the schematic gives you the frequency range the 22k pot will give, from 100nF to 10pF will give low KHz to MHz. You can tune from dc to 3MHz with the 4047, but I would try to work from 30kHz to 300kHz (LF band) and tune the loop to the clock frequency, you can then play with the clock frequency to tune the whole TPU to run as required.
The output is DC with a spike from the down time between phases, now where have I heard that before!!! I started out using iron as the loop, iron seems to stop the inter reaction, only use copper, it is probably because iron will cause the magnetic effect within the iron and not outside, a reaction is needed with the near field and not the internal field. One learns from one's mistakes, seems that maybe the aether will give up some energy within this near field which the coils will pick up and recycle.
Regards
Mike
"All truth passes through three stages. First, it is ridiculed, second it is violently opposed, and third, it is accepted as self-evident."
Arthur Schopenhauer, Philosopher, 1788-1860
As a general rule, the most successful person in life is the person that has the best information.