Meyer's Resonant Charging Circuit Analysed

[Farrah Day]

Hi Guys, Merry Christmas.

Firstly I'd like to agree with you Zippy about the wfc being effectively a resistor in 'parallel' with a capacitor - can't think now why I'd thought it to be a series set-up... but, I was of course wrong.

That Meyer cct equivalent looks fairly good, though I'm not sure what the additional resistors RP1 and RP2 are meant to represent. Any ideas?

Here's something interesting. I've managed to get a few goes on an Electronic Lab Circuit Simulator.

Firstly, I made a few assumptions for my control cct.  The wfc was a 50nF capacitor in parallel with a resistor of 1 Mohm.  I had a 10mH inductor with a 5 ohm resistor in series.  I then applyied a square wave of 10 volts amplitude.

Now with the blocking diode in place, the results were very discouraging. The scope showed a 20volt ish flat dc after the capacitor initially charged up. This was over a frequency range of 100Hz - 50KHz. All this was to be expected really.

Taking out the diode out does indeed make things more interesting. At around 7KHz, the voltage jumped off the scale, and I had to alter the scope to a 1Kv per division in order to see the full waveform. I reckon I was seeing about 5Kv pp. As I reduced the value of the R-C (wfc) resistance, the waveform voltage reduced, but was still well into the kilovolts at much lower resistances. At 10Kohm I was still getting a 1kv pp waveform.

Altering the capacitor value to just 5nF gave pretty well the same results, albeit at a higher frequency (21.7KHz). So even with 5nF capacitance and 10Kohm resistance I still had a 1Kv pp waveform across the R-C component.

With 5nF and just 1Kohm resistor, I only had about 40 volt pp. 

I never altered the inductor from its 10mH value or the 5 ohm resistor, as these are things that we can be in control of ourselves.

From this it would appear that things show far more promise without the blocking diode. But of course this now brings new problems, in that we do not have a bipolar signal, hence the ploarity across our wfc is continually changing. But, just how important is this?  It doesn't matter to us that hydrogen and oxygen will be evolved from both electrodes as we don't want then kept separate, but will this be detrimental to the formation of the dielectric layer during conditioning, or can we condition both electrodes this way?

Two things are for sure, if it is neccessary for the pulses to be bipolar, hence if we need the blocking diode, then we need to use some form of transformer to up the voltage pulses.

Secondly, the greater resistance of the electrolyte and/or the better the capacitance of our wfc, the greater the voltages provided across the cell.

We do however have to bear in mind that these electronic simulators use ideal components and of course do not take into account the non-linear resistive properties of the electrolyte, nor indeed the somewhat mysterious properties of water and the effect of ioniation.  So, though they can be useful in getting you on the right track and perhaps throwing up ideas, only practical tests will be ultimately conclusive.

[z_p_e]

That Meyer cct equivalent looks fairly good, though I'm not sure what the additional resistors RP1 and RP2 are meant to represent. Any ideas?

Yes, it is my understanding that the parallel Rp in an inductor model represents the core losses. You will notice Stan hints in Fig. 7-8 that there are "Rp" core losses in the transformer as well.

I had a 10mH inductor with a 5 ohm resistor in series.

I had asked the question in a previous post, is there reliable information as to the construction/value of these chokes? Where did you get the 10mH and 5 Ohm values from?

This simple LR model is a good start, but in reality it will yield very poor results. I have fairly extensive experience modeling inductors, and I can say with confidence that unless all the parasitic elements are included, the simulation results will be far from reality. All the intrinsic components shown in the above Fig. 7-8 choke must be determined, AND the core type known and implemented in the model to get realistic outputs.

Inductors are one of the most difficult devices to model, and unfortunately, ou research always seems to involve their use. The values of Rp and Cp must be determined on the bench. Failing the availability of reliable construction notes, one can estimate these values, but as a minimum, the inductance should be known.

Taking out the diode out does indeed make things more interesting.

From this it would appear that things show far more promise without the blocking diode.

IMHO, the diode must be there. I do not feel this was a ruse on Stan's part. As you found, there is quite a difference in the resulting voltage and wave form across the cell. He also goes into the specifics of why it is there. Albeit it is still confusing, by its presence alone, how can resonance as we know it occur?

Two things are for sure, if it is necessary for the pulses to be bipolar (you mean unipolar?), hence if we need the blocking diode, then we need to use some form of transformer to up the voltage pulses.

It would appear from the patents and brief, that indeed a step-up transformer is/was always used. The primary is being pulsed through a transistor switch and a variable DC supply, anywhere from 0V to 12VDC. If Stan grabbed an OTS 12V transformer, then we can expect the secondary voltage to be about 100 to 120 Volts, before the chokes. I would want to simulate all this to be 100% certain.

Regarding the two chokes, in the earlier patents, they were not only drawn separate from the step-up transformer, but one was variable. In the later drawings, these two chokes were shown formed on the same core as the step-up transformer, and both chokes were of a fixed value. This would appear to be a refinement, but evidently the "older" way of doing it also worked.

So the step-up transformer we can readily obtain, the chokes however, we need to find out what their value is.

We do however have to bear in mind that these electronic simulators use ideal components and of course do not take into account the non-linear resistive properties of the electrolyte, nor indeed the somewhat mysterious properties of water and the effect of ioniation.  So, though they can be useful in getting you on the right track and perhaps throwing up ideas, only practical tests will be ultimately conclusive.

Indeed. However, as I mentioned, if the time is taken to bench test the chokes (to find all the model values) and the (conditioned) WFC, it would be rather easy to see what is really going on in this device.

My understanding so far regarding the water-resistor is that its V-I characteristic is somewhat similar to that of a silicone diode. The V-I curve is non-linear up to a certain value, at which point it is fairly linear from there upward.

[dutchy1966]

Hi Darren, Farrah day,

I have been playing with the thought of introducing a transformer into the D14 circuit for a while now. The idea is to make the voltage powering the mosfet adjustable (as suggested earlier in this thread). Just an ordinary adjustable PSU would suffice I guess. I rem,ber reading somewhere that Stan used a transformer that upped the voltage about five times. As I do not know what input voltage he was using this is pretty useless. My plan is then to take an ordinary 230V/12V tranformer and stick in there in reverse. This would make my pulsetrain adjustable from virually zero up to about 230 volts. Can I just do this or are there things i should take in consideration before doing this?

I have my controller cct ready, tubes ready, just searching for proper casing to be able to adjust things easily. I have a acrylic tube but it makes disassembly quite a task, so looking for a larger container. If anyone has any suggestion.... The tubes are 20 cm high and there are six of them.

Also what I would like to mention is the presentation of Moray King. He states that using a around 40 khz square wave and a gap space of 1mm or less is enough to get over Faraday levels of Browns gas production. He stressed that a gap of 1mm or less is important to see the effect. This matches with Ravi's gap of less than 1mm.....

regards,

Robert

[Farrah Day]

Hi Zippy

Yes in my haste I wrote 'bipolar' when as you correctly pointed out I meant 'unipolar.

The electronic simulators are good in some ways, in that you can adjust variables at will in seconds and see instant results, but one thing that no simulator I know of will do is simulate the results of ionisation taking place.

I simply pulled the 10mH inductor 'out of the hat', as the one thing that we can make ourselves that won't vary is the inductor.  It doesn't matter what it's value is precisely as you simply need to alter the frequency to compensate for a different resonant frequency.  That said nothing resonates once we include the diode.

I don't think it is - or will be - possible to get accurate bench readings on the capacitance of our wfc's as I would imagine them to be constantly changing, with voltage, dielectric breakdown and resulting ionisation.  I think, knowing the inductor value and frequency is enough to allow use to calculated the mean capacitance. I think that is as close as we will get with that.

The diode is a real problem though as without being able to simulate ionisation, the cct with the diode just produces twice the supply voltage as a straight dc on the scope. But is this the case if ionisation takes place??

OK, I see now, Rp1 & 2, are loses due to eddy currents within the transformer and inductor cores.

Bit of a pain that Meyer's diagrams varied so much.  I've not yet experimented with more than just the one inductor in the simulator as of yet. Neither is there the option of bifilar winding, so simulation experiments are of limited use.

Purarich never used a blocking diode in his ac electrolysis, and Meyer's Resonant charging cct is a copy of this, so I would not totally discount the diode as being there to lead us on a wild goose chase!

It's crazy insn't it. The more you delve, the more complicated and mysterious this ridiculously simple little circuit becomes when you add the 'water' into the equation.  Practical experiments with a few 'knowns' will, I think be the only reliable way forward.

Robert, I don't see any problem with reversing the transformer, but bear in mind that the primary winding will have a much lower current capability than the secondary, so you will easily burn out the windings if you are not careful.

[Garfield]

FarahDay:

When you were referring to kilo-volts I thought I was reading it wrong. Did it really go from a 10 volt square wave to 5 kilovolts? Thats amazing.
    Here is what I found:
Using a 2.5Mh coil and a .01 mfd (10 nano) cap.
Fed 12.5 v p-p sq.wave to this circuit and measured 45v p-p across the cap. Not nearly as impressive as your circuit. But my coil had a resistance of 42 ohms. Resonance was at 33kc.
  The other thing is that once resonance was achieved the resonant circuit changed the square wave to a sine wave. This, I was expecting. What I did NOT expect was. When inserting the series diode, I got zero output. (this is what you were expecting). I tried reversing the diode, using a silicon diode, a germanium diode but to no avail. So much for that theory.
    Notice that in Meyer's circuit, he uses 2 resonant chokes inductively coupled through the common core. Lawton's 2 chokes being wound in bifilar style are both inductively coupled and capacitively coupled. Being one on each side of the wfc, are they trying to achieve some type of balance?
     I'm really puzzled on how that blocking diode is supposed to work.
Oh well, will just have to put the old brain into a higher gear.

Garfield