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

[HEYDUDE]

"eg. text book would say that to store 2 Joules in a cap requires a total of 4 Joules - 2 end up stored in the cap, 2 get dissipated as work - therefore, the efficiency of charging cap: 50%"
hmm, never realized this, thanks for pointing out.

This statement holds true only for a voltage source doing the charging.

it should read:

"eg. text book would say that to store 2 Joules in a cap requires a total of 4 Joules - 2 end up stored in the cap, 2 get dissipated as work in the external resistance- therefore, the efficiency of charging cap: 50%"

[nul-points]

hi alan

i'm sure you're aware by now, having followed this thread with your customary attention to detail, that our recent discussion, on relative energy conversion between charging a cap and the resultant stored energy achieved by it, applies to the test circuits shown in the thread - ie. circuits supplied by voltage source (constant or capacitive)

the 50% efficiency stated by text books, under the context we've been considering, is stated as being regardless of whether the dissipation occurs in external resistance or internal

what if there's absolutely no external DC resistance (improbable)?  Then we'd still get 50% energy loss in the internal ESR of the cap

of course, in the real world we'd always get at least some loss in the wiring resistance as well

however, as shown in the schematics of the test setups here, this is an RLC circuit with an intentional external resistive load whose value is chosen to ensure that a significant majority of the output energy gets converted in the load, not lost in wiring or the cap ESR

of course, you're well aware of all the above, so i only need to answer:

"eg. text book would say that to store 2 Joules in a cap requires a total of 4 Joules - 2 end up stored in the cap, 2 get dissipated as work - therefore, the efficiency of charging cap: 50%"

shorter - to the point - and still true 

all the best
s.

[HEYDUDE]

External resistance includes ESR and wiring.......textbook considerations of capacitors are usually as "ideal" with ESR modeled as a external series resistor. Same for Spice models of capacitors.

[nul-points]

a real capacitor has real internal ESR -  dissipation of real energy in its internal ESR can contribute to temperature increase in the capacitor

i'm sure each of us is aware of this

all the best
s.

[poynt99]

Doc Ringwood Analysis

The biggest problem facing all FE enthusiasts (of which I am one) is obtaining accurate and reliable measurements. Even with the simulations I've been running lately, a fair bit of "eye-balling" is necessary to perform energy calculations. Not being happy or satisfied with this, I set out today to solve this problem once and for all.

In SPICE I have developed a Joule/Watt meter that can be used to directly measure the energy dissipated in any device by measuring the voltage across it, and the current through it to obtain instantaneous power. Then the result is sent through an ideal integrator to provide direct readout of Watt-seconds, or Joules. Sampling before the integrator shows power in Watts if desired.

With this handy new tool in my arsenal, I set out to perform my own simulation test of Sandy's charging setup to see what gives. See the three attachments for the voltage wave forms, the Joule readouts, and the circuit diagram.

From the scope shot "voltages.gif" it's evident that the sim wave form results are very close to Sandy's as posted on his website. The C2 end voltage was very close as well. The time in interest is from 0s to about 9.12ms.

The Joule Meter outputs are shown in "Joules.gif". All dissipative elements are accounted for (except losses in the caps themselves). The green trace is R1 (948uJ), red is D1 (104uJ), and purple/blue is the small coil resistance R2 (47uJ)  (in my diagram). The sum of all three losses from the Joules scope shot is 1.1mJ.

The final voltage on C2 is about 3.28V, which computes to 1.05mJ for a 196uF capacitor.

The sum of the total dissipated energy and that left in C2 is 1.1mJ + 1.05mJ = 2.16mJ.

The energy used from C1:

Starting Voltage = 8V; E= 6.08mJ
Ending Voltage after 9.12ms = 6.42V; E= 3.916mJ
Energy used = 2.16mJ

So the energy used equals the energy dissipated and transferred.

Sandy has stated that in his test C1 was discharged from 8V down to 7V, but from my test this seems unlikely if C1 really was 190uF. In order to achieve the same results in the sim, C1 had to be increased a full 50% to 300uF.

As we don't have the scope wave form of C1 discharging in Sandy's test, I have to assume that this is where the error lies. 7V to 6.42V may not seem like a significant difference, but because these voltages are squared to compute energy, it DOES make quite a significant difference, and emphasizes the need to perform careful measurements.

Regards,
.99

PS. Many may be thinking that this Joule Meter would be a great tool to have on their bench Well, I hope to have a design done in the new year for all to build