COP 20.00 (2000%) Times, Reactive Power Energy Source Generator,

[hartiberlin]

Hi Jason, you are absolutely right !
They are just switching one cap only in reverse and back again !

I installed Java again and then went to:
http://www.falstad.com/circuit/

where the online simulator is hosted and copy and pasted the code you posted into the
import window and ran the simulation.
(For those not familar with this online simulator)

Yes, that ´s it...

Now we only need to fine tune it and design the right switching circuit to efficiently
electronically turn the cap back and forth connected.

Any idea how to do this switching electronically with a few MOSFETs and some driver circuits ?


BTW, Jason, how do you synchronize your toggling circuit inside the simulator ? How can one control,
when the switches are toggling ?

[hartiberlin]

Hi Barry, great devices !
You and Jason are brilliant ! Yes, they are only using one cap that is switched back and forth
in its polarity.
I have to learn this simulator to see how it can be tuned further.

Now we only need to fine tune it and design the right switching circuits.

Many thanks for your brilliant ideas !

Regards, Stefan.

[listener191]

Hi Stefan,

Not brilliant for sure but perhaps persistant!

Attached is a zero cross circuit that produces edges for positive and negative crossing points. Ignore the other circuitry, as this was used as part of a capacitive voltage multiplier.

For delays and pulse width timing that does not use a microcontroller, suggest Linear  LTC6993-1 and -3 devices which are cheap and are  programmed via external voltage control, but are based on internal clock counter. Use the positive and negative zero cross edges to trigger  a -1 and -3  device is series. The -1 sets the delay the -3 the pulse width.
A lot less components than 555 timers more accurate and they can all be voltage controlled from a couple of pots.

Barry

[listener191]

Attached is a bidirectional switch example with a driver built from standard parts. Photo voltaic drivers would be a good alternative for simple implentation, but can be expensive.

A floating 15V supply is required for each switch. I dont favour charge pump devices as they can be easily damaged but as this is a constant duty cycle requirement, they would work OK.

Use the same circuit for MOSFET's

If just switching caps into resistive loads there should not be a big problem with voltage transients.

When switching inductors, the most important point is to recover the inductive energy at the device terminals, as interconnecting track or wire inductance can result in large transients that exceed the peak voltage rating of the device.

Also IGBT's rated at 1200V 150A pulsed will only typically handle 10A continuous, before you depart from the Safe Forward Biased Operating Area. When switching low voltage but at the same time allowing very high voltage transients to be developed, devices need to be paralleled if 10A is to be exceeded.

Babcocks switching scheme which uses SCR's to turn on and IGBT's to turn off, has the downfall of relatively slow switching speed due to the SCR turn off speed. Carefull selection of the SCR type is needed to keep switch off speed optimised. The SCR is used to handle the large voltage transient, which the IGBT doesnt see, as its is already turned on.

If your intention is to recover a large voltage due to a high rate of Dv/Dt, then the MERS scheme would work better for you. 

Barry

[popolibero]

Hi Stefan and all,


I have been working on this for the last few days. I have tried both the single cap and the parallel/series setup. The circuit works well but I don't see the low input power at the source.


I'll explain what I'm doing. Since I'm not an electronics genius I didn't build a circuit for the control signals. But I'm a sound engineer so I recorded the signals on digital multitrack recording software on my mac. A sine wave and four 20-25%duty cycle square waves. This way the control signals are always in sync with the sine wave. The sine wave output goes to a power amplifier which powers the transformer.
The 4 control signals go to a breadboard where I control and do the switching.


It goes like this:
The output of the audio interface is low voltage so I send it to MOSFET drivers(4 channels), to sharpen the signal and because its input threshold is very low. This way I get a sharp 15V signal on the driver output. The drivers power the input of 4 (or 6 depending on setup) H11D1 optocouplers. On the output of the optos I have 4 (or 6) discrete micro DC-DC converters to do the independent switching of the power mosfets. Each switch is made of 2 mosfet (search mosfet AC setup).
For the parallel/series setup I used the Bedini scalar battery charge setup, but replaced each of the 3 transistors with an AC mosfet setup (this is where I need 6 mosfets thus 6 optos), so it can switch parallel/series in the pos. and the neg. half cycle.


The one cap setup is the same but you only need 4 mosfets, that's one AC mosfet setup on either side of the cap.


I'll try and draw a crude basic schematic...


Normally, in a non switched setup, when I run the transf. with my amplifier and have a tuned cap on the secondary the amp meter on the primary of transf. goes down almost to 0 like it should. But here even though the circuit works well I don't see this (yet).


Since my signals are recorded and looped I can't change duty cycle in real time but I can shift them back and forth in real time very precisely by nudging the track(s).


I just thought it could be an option for someone who doesn't know how to make the signals needed.


I'll keep working on it...


regards,
Mario