Selfrunning Free Energy devices up to 5 KW from Tariel Kapanadze

[IWD]

Dear Wattsup ,
Please find , that the pin numbers on TBC110ЛА starts from 2 after 1 then 3 and so on ....
The winding exsist between pin6 and pin 7 . I think aproxsimately 140-150 turns . ( Voltage 2.36 V on pin7,
then 10V sinus signal was applyed to pins 5-9 . By documents transformation rate 0.232 +- 5% .
I have added photos of TBC110ЛА . New . Never used .
Regards

Hi Osiakosia, pls, can you measure the inductance of windings? and post it.  I am unable to find it

[ronotte]

Hi Roberto,
Trying to estimate the pump input energy your circuit needs,  let’s say your average voltage input (waveform D) is around 6V (12V_pp) and let’s say your average input current (waveform G) is around 1.5 - 2mA (8mA_pp) so to maintain the loaded oscillations in L1-L2 tanks you need about  9-10mW input power  (rough and rigorous estimation of course).  This is what roughly the xtal oscillator as its output power  should produce instead of the Wavetek gen and the xtal oscillator is to be fed from the tank circuit of course.  It was a very good idea to use xtal oscillator to get rid of the supply voltage dependent frequency stability of  CMOS RC oscillators and reduce the self-consumption of the oscillator itself. Here I would refer to an also micropower xtal oscillator with bipolar transistor, maybe its idle current consumption is lower than that of the FET oscillator, this way convertion efficiency may improve, see this link:  http://discovercircuits.com/DJ-Circuits/gatextl.htm   (the Gate Control Function with the CMOS buffer may be omitted of course and supply voltage can be towards the 5-10V range what the rectifier diodes give, instead of the suggested 3V.)
I understand the definite need for the 50% duty cycle the pump oscillator should produce, sorry for my suggestion on lower duty cycle, my thought was to reduce the ON time for the CMOS RC oscillator, to reduce its power consumption.
Regarding the 5cm OD coils, it is ok but perhaps the coupling coil L3 could be made on a separate card board support of an OD=5.1cm and you could adjust the coupling of L3 to L1 by sliding L3 over L1, maybe a more optimal position could be found, meaning less load on the tank.  This variable coupling could not be made in case of ferrite pot cores, the solution in that case is to use a multitap coupling coil. 
The forward voltage loss of the diode bridge can be reduced not only by 4 Ge diodes but by using a full wave voltage doubler which needs only two diodes instead of 4, and if these two diodes are Ge types, you may gain (for free) about 1.2 to 1.3V in voltage across the 200uF puffer cap versus the 4 diode bridge. (OF course the voltage doubler in itself is not a power multiplier.)  See this link which doubler I mean: http://www.play-hookey.com/ac_theory/ps_v_multipliers.html
It would be interesting to learn whether you changed the negative bias voltage up and down from its 10V value as shown or you found it as optimum at 10V for the varicap diodes?  (This 10V source was loaded by the reverse current of the varicap diodes, meaning only a maximum of some 100 nanoAmper load all the time.)  Also I would be curious on the DC resistance of L1 and L2 coils.

Your kind and detailed report on this circuit, perhaps combined with my humble suggestions, could inspire some further interest towards achieving a selfrunner at some mW power level. One would ask what usefulness has a mW power circuit? and the answer is: enormous if it selfruns...   
rgds,  Gyula

OK folk,
I'll cut it short as the following rumblings are completely outside this thread's target, just allow us to complete only a preliminary check as the matter's OU implications are to be considered worth of interest.
Hi Gyula,
I've seen that in order to maintain the down-conversion mode in tank circuit the power needed could be less than you have assumed. In fact the waveform in TestPoint G due to the small on-screen view is not good. For your use I report here a much detailed waveform that I captured using a wide-band I probe on the pump oscillator return wire. In the upper trace you can appreciate the down-converted 0.5MHz wave (pump = 1MHz) and the corresponding current pulse delivered by the pump oscillator itself. The +/- spikes have 100mv amplitude and the pulse width is only 20nsec: the required equivalent power is very, very low (The current pulses themselves pass capacitively only trough the back-biased varactor's capacity). You have also to add the power delivered by the back bias circuit...that in any case is in the nA range (yes, back bias is a very nice method to change the varactor's operating mode: easy to change from: down conversion to up conversion...I did  both). The bad side is that even if pump (you see it) should be almost decoupled from the real output captured by the link, nonetheless I noted an increase in pump current when really driving a final 100Ohm resistor load in open-loop condition...(I do not remember now the increased range...I should consult my hand written work-notes).
In my case the tank has L= 2.3mH and his resistance is 2.3 Ohm. The 1/2F0 resonance coil reactance XL =2pifl=2*3.14*500000*0.0023 =7222Ohm hence, considering a possible useful 100Ohm load, we have Q= XL/R=7222/102.3=70 that is not bad.
I agree with you about Ge diodes consideration...at the time I considered it but not tested it. As you may imagine I did an extensive study and actually tested many different kind of micro power oscillator: best solution has been with fet (2n7000) XTAL solution...much better than CD4007 always with XTAL connection.
I tried in every way the selfrunner but for a reason or another did not reached it. I am interested in case of  new ideas to restart that experimentation.
Thanks for your help.
ronotte

[gyulasun]

Hi Roberto,

Thanks again for the finer details on your tests.  Yes I felt my 10mW input power figures sounded too high but the current waveform scope shot let it figure that way.  Now your latest input current scope shot surely suggests an order or even less input power necessity, probably less than a mW.
It is interesting the output load still reflects back on the pump input, maybe higher than 1:2 ratio in the down freq--pump freq this reflection could become less? (I mean your mentioning the pump current increase when the 100 Ohm load is connected vs the unloaded case.)
IF you are interested in even newer ideas for improvement, I suggest pondering on using a current transformer instead of the 50 turn coupling coil, let's suppose a few turns of the current transformer would be inserted into the tank circuit in series, just cut the wire at the upper end of L1 for instance and insert the few turn long primary coil of the current transformer there and the secondary coil of the CT would feed the two Ge diodes of a full wave voltage doubler.  What do you think? I have not tried this, sorry but such CT output power coupling method is not 'unknown' in a thread like this on Kapanadze... certainly not off topic. 

rgds,  Gyula

[ronotte]

Hi Gyula,

the fact I mentioned (oserved increase in pump current when driving 100Ohm load) should have been given much more attention. The observed increase if I remember well about 20-30% so not so big. It's the fact in itself that sounds weird as the mechanism ...how to say...is at least too  'far' to be explained only in terms of reflected loss. Of course the breadboard's prone to many problems hence it could also have been stray cap coupling...In the next try I should consider to build a better designed layout possibly getting rid of the breadboard and building it directly 'in air on Cu plane' and very near the 2 tank coils: a RF approach.

I'm sorry I did not tough about a CT solution to pickup the output power: it would have been interesting even if for the comparison to standard methodology.

Here come the bad news, for the moment I'm to much committed to Kapanadze like devices so not much free time slots to give another try to this particular circuit. There is also another consideration to do: the varactor diodes are delivered in a microscopic package: it is very, very difficult to solder them without a stereo binocular microscope and a micro iron-solder stick and ancillaries tools (micro pliers,...). I remember to have had lot of problems...When proper time I'll keep you informed. It goes without saying that that you may contact me whenever you want.

Thank you for your time

ronotte

[gyulasun]

Hi Roberto,

Ok, I understand and thanks for the further comments.  Unfortunately,  nowadays the old types of varicap / varactor diodes with TO92 casing like BB112, and with SIP5 casing like BB113 or BB313 etc from ex-Siemens are rather rare, although still available,  and I know what it is like to handle SMD or similar microscopic packages: cumbersome to say the least...
Thanks for you kind offer on contacting you, I will review my resources and possibilities to attempt a replication when my time permits and will send you a message.

Greetings,  Gyula