RESONANCE EFFECTS FOR EVERYONE TO SHARE

[TinselKoala]

Another thought: it might be instructive to compute the energy involved in a one volt drop from 50 to 49 volts in a 6000 mFd capacitor. Let's see, energy on the cap at 50 volts is (0.006000 Farad x 50 x 50)/2 = 7.5 Joules, and the energy on the cap at 49 volts is 7.203 Joules, for a difference of 0.3 Joules, about. You'd not even notice a brief drop of one volt on this cap as you actuate your system.
Now let's figure the energy on a 6000 mFd cap at 3.5 volts. (0.006 F x 3.5 x 3.5)/2 = 0.03675 Joule, or about one-eighth the input energy, assuming the 1-volt momentary drop on the supply cap.

[Groundloop]

@TinselKoala,

Moving the ampere meter to the DC side will make it impossible to get a good reading.
Then you will need a expensive true RMS meter capable of measuring high frequency
at complex waveforms. The way Luc has placed the amp meter is correct. The meter
itself will display the correct AC ampere because the meter is designed to operate with 60Hz.
The ampere meter itself is a low resistive shunt and you are measuring the AC voltage over
that shunt. It will be the same as if Luc put in a low ohm resistor and then measured the
voltage over the resistor.

Groundloop.

[gotoluc]

Luc,

Thank you for making the video. It is a great video and it clearly demonstrate what we where talking
about in private mails. It seems to me that your circuit (two coils) is outputting more Watt than
you are providing to the circuit. I have one question, it is possible to do the same test at 12 volt
input level? Then you can use the input voltage to also deliver power to the switched side. This will
give us an idea on how much the circuit is using at total compared to the total output. I ask because
the two diodes at the IR2103 is connected to the switch output side and may inject power to your
output, thus adding to the result. It may not be so but I think it is important to check that also.

You are doing a great research with this circuit and you should keep posting your results even if there
are few replays to you posts. Keep up the good work. Now that there is a replica (Tinsel Cola) I bet
he can confirm many of your findings.

Regards,
Groundloop.

Okay,  just now I tested the 12vdc feed side to the logic of the circuit using a 12v 90ma car dash instrument bulb in series on the positive lead and started the circuit to see how much current is going through the bulb. First test is noting connected to the switch side of the circuit and the bulb just glows. I know this bulb well and I would say it is about 25ma of current going through. I then connect the switch input 50vdc cap and nothing changes in the bulb intensity. I then connect the coils and again nothing changes in the bulb intensity and the 2 capacitors with load fill up as usual. The next test I did is connect the switch input to the same battery as the logic is using and nothing   I think I fried something by doing this, since it no longer works   bad idea to have the logic and the switch feeding from the same battery.

Anyways, I think my first tests are good enough to prove the power is not coming from the battery feeding the logic side of the circuit.

I'll try to fix the circuit now.

Luc

[Groundloop]

@gotoluc,

The switch should not have any problem with using 12VDC for both the switch side and the logic side.

Groundloop.

[TinselKoala]

Well, OK... I guess. I'm used to having the fancy equipment at my fingertips and sometimes I forget others aren't as privileged. Sorry. I have the dream job...

But:

First, what about the energy calculation.

Second, what about the surge-filtering ability of the big input cap.

Third, I run my unit on the same battery supply to both logic and bridge with no difficulty. I even accidentally hooked up the logic supply voltage backwards once, for a second, and it survived. My initial tests last night were driving the LT solenoid (16 ohm electromagnet coil) as a load, with the 11.4 volt LiPo used as both supplies simultaneously. It worked from fractions of a Hz to 200 kHz (input clock). And with no load, but with the battery hooked to the bridge input and logic side, I let it run for 4 hours this afternoon.

Btw, there is a bit of power coming from the logic thru the diodes. You can see my output leds glowing very dimly, when there isn't any bridge supply connected. But this leakage is only a few milliamps and will be swamped by the operational current. So it isn't the answer to the question.

Why not do one cycle with the power supply to the input cap disconnected? Too bad you don't have the 2104. It turns off all output mosfets with a single pin, no matter what the clock is doing, so it would be easy to set things up, charge up the cap to 50 volts, disconnect its PS, trigger the bridge to start the circuit, stop at the normal time, then look at the voltage on the input cap. Then you'd need to correct slightly for the leakage, but you'd have a handle on the input power, even if your meters won't measure it while it's rushing by.