[NewCandle] Parasitic Cap Charging

[John Winterflood]

Dr Stiffler wrote:
> My site was down for 12 days during power outage as a result of 
> Hurricane Ike followed by 'Denial of Service' and invasion of my old 
> MS Server OS. You are lucky you could find it in operable state.
Yes it appears to be down again.
> Now to your initial topic. Yes and No on ESEG. The current state is; 
> 1) will not start itself. 2) requires initial minimum charge to start. 
> 3) output provides nothing for a secondary device. 4) will run for 
> time periods far in excess of initial start charge, is "limited".
I can't see that any of the limitations you have listed apply to the 
setup I had in mind:  Start with two capacitors charged to say 16V.  Run 
the ESEG from one capacitor while charging the other in a floating state 
- as has already been demonstrated.  Drawing 3mA for say 1 second from a 
3 Farad cap will only cause a voltage drop of about 1mV in its supply 
voltage causing it to drop to 15.999V.  Meanwhile the floating capacitor 
should be charging at up at several (~6) times this rate and will thus 
increase to say 16.005V.  A multi-pole change over relay can swap their 
connections over every second to run from the fully charged one and 
recharge the slightly discharged one.  Some sort of zener arrangement 
across each capacitor could prevent them overcharging and regulate them 
to a suitable maximum voltage.  Such a self-runner would really be 
something to get excited about as it would be obviously and undeniably 
overunity!
> I have not been able to model the exciter in 'MicroCap', I do not know 
> Spice but would welcome help here. If a model could be derived, ESEG 
> may really be something of value.
I'm not surprised you are having problems attempting to model it.  I can 
see no reason for it to oscillate at all - there seems to be no positive 
feedback path that I can discern.  So my guess is that your lack of 
shielding anywhere is allowing feedback to occur by unexpected 
capacitive or inductive coupling.  Until you can find out what this 
feedback path is and quantify it somewhat, it will not be possible to 
model the circuit's operation.

A more conventional approach might be to build such a unit into a 
diecast box with BNC sockets for inputs and outputs and a small metal 
can around the oscillator coil inside, etc.  But then you would probably 
just get conventional results - no oscillation - and nothing new would 
be discovered!

J.