OU/COP>1 switched cap PS cct like Tesla's 'charge siphoning'

[nul-points]

if you can make use of the energy dissipated in charging the capacitor, the switched-charge circuit is very efficient

is there any way of drawing on energy from the 'environment' to achieve a COP > 1?

i have another long-term experiment running which builds on some of the principles learned from the switched-cap experiment

it's looking hopeful at the moment, but it's very early days yet:-

a) the test circuit is VERY low-powered in order to gain sufficient relative influence by the environment

b) due to the low-power nature of the experiment it will probably need months of elapsed time to confirm if the total energy in the system is increasing or decreasing

the circuit charges a small capacitor from a very large cap; then a discrete SCR cct (for Nerzh!) triggers a current pulse which flashes some LEDs which feedback the current to the rechargeable battery (any flyback current from the Xfr also gets fed back to the battery)

the question is:  can this circuit cause the battery to stay charged at the same (or greater) level - or will it discharge?  i can tell you now that with NO battery, the input cap just discharges - so no pressure! 

here's the circuit...

[nul-points]

if the circuit 'captures' sufficient  environmental energy then the battery voltage will not decrease

preliminary results look interesting:

the graph below compares 8-minute long traces of the off-load voltage (blue) of the battery with two other conditions:

  - battery loaded with the operating pulse circuit but with feedback disconnected (red)

  - battery loaded with the operating pulse circuit and with feedback connected (yellow)

as expected, the battery voltage falls from its no-load value to its 'loaded, no feedback' value
(the start of the 'offload' and 'onload with no f/b' traces actually capture the slight fall in battery voltage when disconnecting the relevant part(s) of the circuit whilst running)

however, when the onload circuit has its feedback connected to the battery, then the battery voltage rises

results so far show a very slight increase of battery voltage from offload to 'onload with feedback'


at the moment the DVM reading for the battery is 1.260V (with the final digit occasionally flicking to '1')

place your bets now, ladies & gentlemen...will we be in the Red - or in the Black? 

[nul-points]

for you guys who get as excited as i do by the sight of a bunch of wires sticking out of 'breadboards', here's a photo of the setup...

[exnihiloest]

...
It's simply charge separation, and since the net charge has neither increased nor decreased in the charging process, no net work has been done to charge the capacitor's dielectric.

Not exact. Before charging a dielectric, the electric dipoles are randomly distributed. When you "charge" a dielectric, positive charges are attracted to the negative plate and negative charges are attracted to the positive plate. Therefore there is a torque that globally separates positive from negative charges inside the dielectric, at the cost of a work (W=1/2*C*U²).

[nul-points]

hi all

there's been an interesting development with the latest version of my low-power experiments here...


the simplified circuit config. at present looks like this:

  <<see simplified schematic below>>

(in case it's not obvious from the schematic, this is a fully-looped circuit - energy from the LED pulses is fed back into C2 and Vb1//C1)

the supply battery is an old rechargeable 8.4V NiCad which has been used occasionally, over a period of several months without recharge, to supply low-powered experiments - the off-load voltage has gradually reduced to approx 7.5V

after connecting the battery to this pulse circuit the terminal voltage increases by approx. 1 volt! (so the on-load voltage across Vb1//C1 is approx 8.5V)

  <<see graph below>>

(graph shows voltage across Vb1//C1 against hours of operation)


usually, when a load is connected to a battery, the terminal voltage will fall as current flows through its own 'internal resistance' first before entering the load

this effect is more evident when the battery is discharged because its internal resistance increases significantly

however, this experiment is definitely showing the opposite effect - the voltage across the battery is rising

somehow this circuit is operating at a higher steady voltage than the off-load battery voltage - the battery/circuit combination 'appear' to be in some kind of high-impedance 'self-sustaining' operation mode


please note that i'm NOT saying that the battery has become fully charged to 8.5V (if i connect a scope probe across Vb1//C1 the voltage starts to decrease, so the battery obviously still has high internal resistance)

BUT - if i let the circuit run and take 'spot' readings (say every 30 minutes), with scope or DVM, then the voltage across the battery terminals is being sustained at a higher value than its off-load voltage

this circuit is providing a 15us pulse (off-period approx 30s) to some LEDs, so one possible explanation might be: 

  "just for the duration of the pulse, the battery voltage drops to an 'on-load' value (which is less than its off-load value)"

this pulsed 'on-load' voltage drop on the battery WAS happening with a previous config. of this circuit but NOW the input is filtered by Q1 and C2, where C2 is approx 0.3 of a Farad - and close inspection with a 'scope confirms that the pulse energy does not produce any significant voltage or current spikes at the battery terminals

  <<see trace below>>

(trace shows voltage across Vb1//C1 (RED channel, 0.4V/div);
   and input to LED1 (BLUE channel, 2V/div))

i believe that the immmediately previous load history on this battery can rule out that this effect is just 'battery relaxation' - and i've been able to repeat this 7+V to 8+V increase several times over the last week just by stopping, waiting and then re-starting the experiment

the voltage increase measurements have been recorded from both a DVM and PC-based scope/datalogger, which would rule out one of the usual explanations behind this sort of result: ie. that the DVM battery needs replacing

i've also run tests with the circuit operating inside a shielded container (aka. a powered-off, grounded, MWO) - the same battery voltage increase still occurs


something unusual is happening here - comments welcome (conventional explanations, or otherwise!)

thanks
sandy


(i'm in the process of gathering the experiment/circuit details together which i'll post here when ready)