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

[nul-points]

hi Poynt

what level of energy dissipation will you be expecting to occur in the DC resistance (mostly sum of ESRs for L, Cs & D) of the circuit when i run your proposed test?

thanks
s.

[poynt99]

hi Poynt

what level of energy dissipation will you be expecting to occur in the DC resistance (mostly sum of ESRs for L, Cs & D) of the circuit when i run your proposed test?

thanks
s.

Sandy,

I ran a sim to get some numbers.

First it is difficult to find numbers for capacitor ESR so I used 100m Ohms total for both. This seems to jive with data I was able to find on the web. I also added 50m Ohms for the switch resistance. Total circuit R then I used is 0.65 Ohms, which includes 0.5 Ohms for the inductor. The 10 Ohm R1 is removed.

From the scope shot you'll see that the diode appears to continually use energy throughout the switching process and uses 268uJ after 100ms (blue). The energy lost in the total circuit resistance of 0.65 Ohms levels off quickly at 208uJ (yellow).

The total energy loss after 100ms run time is about 476uJ.

C1 (red), C2 (green) final voltage is about 5.39V.

After about 80ms, the capacitor voltage begins to slowly decrease due to the continual switching loss in D1. From this the optimal run time would be about 80ms, which is about where the two capacitor voltages appear to level off. The total Joule loss in this case decreases a little from 476uJ down to 460uJ.

Regards,
.99

[nul-points]

thanks, Poynt, that's an interesting prediction from the sim

in the past test where i removed the 10ohm Rload and replaced it with a 1ohm current sensing resistance the measured results showed a charging current which resulted in an energy dissipation in the inductor and sensing resistance around 2x that stored in C2, 196uF (which was 2.8mJ approx)

this division of output energy between stored and dissipated has been a common feature of the results for these tests, as Dave (CTG Labs) mentioned previously

so, on a previous test, the effective removal of Rload didn't cause a greater proportion of the output energy to accumulate in the o/p cap

it reduced the effective load resistance on C1, which increased the peak pulse current, but the dissipated energy was still several times greater than that stored in C2

so unfortunately i don't think this proposed test is going to provide the simple, single-value measure we'd hoped for

i can still go ahead with the test but obviously the dissipation will also have to be monitored, we can't just note the final stored energy

nailing jell-o to a tree must be a breeze 

all the best
s.

[poynt99]

Sandy.

I'm not sure what else I can say.

Removing R1 will significantly increase the transfer efficiency of energy from C1 to C2. I don't think you would disagree with that. If you want to do the same test but with the 10 Ohm load present, then go ahead. It makes no difference really because you will account for all the energies in the circuit anyway. All that will happen with R1 present is more of the energy will get wasted as heat instead of being transferred to C2. The final capacitor voltage will be lower with R1 present.

At the end of the day you would measure the voltage on the two capacitors when they level off (they will be the same voltage) and add up all the losses and see how it all comes out. All the energies should equal (or better) that which started in C1.

.99

[nul-points]

hi Poynt

i was just following-up on your original intention for a simple test which relied on reading only the single final voltage on the two caps:

> "The circuit is operated until VC1=VC2, and the voltage noted.
> I don't think it can get much easier or simpler."


i agree this test idea is easy & simple, i'm just noting it also appears from previous results that this single measurement won't be sufficient on its own:

  ie. we'll still need to try and get a measure of any dissipation - that doesn't mean we need to use a 10ohm load to do it - and hopefully my previous post doesn't suggest that we should (i mentioned using a 1ohm current-sensing resistance previously for this purpose)

all the best
s.