"Enhanced Generator" from JPKBook

[mml]

Hi mml,

Having drawn up the circuit in readiness for the build I have come across an issue that seems to be relevant to the solid state design.

When looking at how to arrange the negative side connection ports on the circuit board I realised that if I connect the capture capacitor to the main negative rail of the generator then I am effectively bypassing the cap dump circuit which switches the negative line back to the batteries in order to release the stored charge. I have shown this in the attached pic and where the cap dump circuit is traditionally connected to an independent coil. In my case the HV pulses are being generated in relation to the battery negative/ground (FET Source) and so I can't connect the negative of the capture cap to that as it will discharge straight away.

I wondered if you can think of a simple solution to this?

Regards,

Jules

Hi Jules,

here is a solution that works, with a second battery.
In red: the circuit of the current when the coil collapses (Note: B1 does not discharge during this time, it just lets the current flow).
You can think of other solutions.



cordially
mml

[JulesP]

A Solution

Hi mml,

Based on your thinking I have one idea that might work that keeps my current battery setup with the swapper. As the problem is around the common ground line for the batteries and the capture capacitor, can't I just switch the positive line instead of the negative with the FETs in the circuit. I attach the basic idea.

Regards

Jules

[JulesP]

Like this . . . Looking back at SG Intermediate, where similar things are discussed, it appears that switching the positive will work just as well. The only downside is the 0.6V loss from the extra diode in the line but given that the capture cap will always be above the 36V battery voltage I might be able to dispense with it.

Jules

[JulesP]

Revised Circuit and Update

I have attached the latest circuit that I think should work based on the earlier designs of others. I had got R5 & R6 connected up wrong (as shown in the attachment in the previous post) and which would have shorted out D2 so have corrected that. This circuit also switches the positive line since that dovetails much better with my generator design. As I understand it the two pots, R10 for the lower and R13 for the upper threshold, are adjusted once the circuit is operating via scope measurements.

My second attachment indicates some suggested values for these upper and lower thresholds in the charge and discharge cycle to use with a 36-38V battery system but if these are not reasonable then please let me know.

The last two pics show the cap dump circuit build so far. The two FETs and the BJT are located on a clip heatsink under the circuit. When I have got my HV output back to where it has been I will see how the cap Dump Circuit behaves. What are the odds on it working first time? lol

My overall aim here is to improve on the capture of ZPE at higher frequencies since my measurements indicated that the battery was not able to absorb the energy as the HV pulse frequency increased (best was 150-200Hz). If the capacitors are able to respond much better to the higher frequencies and then transfer that energy as regular 'hot' current pulses to the batteries, I might get closer to a linear response of energy absorbed with frequency. While some theory suggests that invoking ZPE is directly related to the voltage gradient of the HV pulses and that the amount of current flowing in them is not that important, I have noted that the higher the circuit/coil current the better the charging so current must have some role to play.

If using the cap dump circuit makes no significant difference then I will have done all I can over the last year to produce an electronic generator based on induction and ZPE and will move on to something else.

Regards,

Jules

[JulesP]

Problem

Well I was hoping for a smooth operation of the cap dump circuit but something is not right in the way it is integrating with the rest of the generator. When I have the generator working and showing HV pulses via a potential divider board (20:1 ratio - see pics of pulse) everything is fine but as soon as I divert the HV pulses to the capture capacitors the coil FET sometimes stops working. This has happened a few times but maybe due to some intermittent fault. One of the advantages of using capacitors is that they present a low impedance to the HV output from the FET Drain so might this result in an overload? Interestingly the HV pulses shape is much sharper than before having lowered the resistor value between the Gate and Ground from 15K to 2.2k.

What I believe I have built is shown in the third pic.

Any thoughts would be appreciated.

Regards,

Jules

[JulesP]

Observations and Thoughts

Hi all,

While I haven't yet found out why the cap dump circuit isn't discharging to the batteries, I seem to have stopped the FET from being damaged (not sure how or why) and have instead been able to measure the voltage rising on the 45,000uF capture capacitor and how quickly that occurs. As previously indicated the hope is that the capture capacitor and circuit might be able to 'process' the HV pulses/ZPE faster than a battery, as there are no heavy chemical ions to shift, and therefore respond better to an increase in the HV frequency. So far observations suggest that the issue is more complicated in that there are other factors that can make a significant contribution to the energy transferred to the caps/batteries. For completeness I have included an updated and simplified schematic for the whole generator including the cap dump circuit. The battery swapper incorporates a heavy duty relay which switches the battery stack being used to power the inverter and the main circuit every two minutes (adjustable) and at the same time directs the output from the cap dump circuit to the other battery stack.

So what I have noticed so far is that the choice of MOSFET makes a big difference to the battery charging. For most of the work I have used an IRF840 but out of curiosity I switched to the IRFP260 (see spec comparison pic). With the former FET, when the HV (at 150Hz) is connected to a single battery in common ground mode, the voltage would rise from 12.2 to 15.8 in about 10 seconds. However, when the same HV is directed to the three battery 36V stack then the voltage would rise from 36.8 to 37.1V over 10s. When I switched to the IRFP260, charging the single battery was similar but with a higher circuit current and with the three batteries the voltage went from 36.9-38.1V over 10s. In other words charging a set of three batteries works much better with the 260 FET. I surmise that this is in part due to the HV pulse width being wider so there's more energy 'under the curve' (compare pulse pics) even though the peak pulse voltage is lower (both measured with 20:1 voltage divider). I am also aware that the best energy transfer occurs when the coil impedance matches that of the batteries being charged but apart from changing the coil configuration I don't know what can be done to better match them. My seven coils in parallel come to 0.5 Ohms which is probably quite close to the batteries but FET's characteristics are also relevant it seems. RS suggested that the 'flyback' diode built into the FETs will try to dampen HV pulses and so using BJTs should give better results.

These observations and the measurements I have done to date raise various questions about the optimum conditions for inducing ZPE into our circuits. From what I have read one of the prime theories for ZPE influx is based on John Wheeler's theory of vacuum polarisation. I have shown this in the attached drawing and where getting ZPE to cohere and flow into a circuit is a direct result of a high voltage gradient of the sort that Bedini type devices produce. It is common to have a high dV/dt with little or no current and while a few forum members have had a go at measuring the current component of the HV pulses flowing into a battery, it seems from my own and others' observations that the higher the circuit current the better the battery charging which suggests that either the charging involves 'hot' components as well as ZPE or that an HV spike with little current component is less of an inducement for ZPE to enter the circuit than one with a bit more electron flow.

Measurements I made a month or so back showed that as I increased the HV frequency with the signal generator, the overall circuit current reduced and the rate of battery charging did not get better and better but actually went down after an optimum value. While I expect the small digital voltmeters have a harder time keeping up I can't assume they are completely wrong and the lower current may well explain why the charging levels fall with increasing frequency. On the face of it, delivering twice as many pulses to the battery should produce twice the rate of charging and the fact that it doesn't may be related to the reduced circuit current at the higher frequencies, quite possibly due to the increasing coil impedance that rises with frequency, or some other factor. We must remember that when a coil is energised there is the opposite of the HV pulse occurring at the front end of the energising process and perhaps this starts to dominate at higher frequencies.

As ever some useful questions arise out of play to date and which any contributions would be welcomed.

1. Which BJTs have been seen to give the best results as coil drivers instead of using FETs?

2. Does the influx of ZPE to a circuit depend just on the voltage gradient of the HV pulses or is the current component of the HV a significant factor?

3. What is the largest net power output or COP that anyone has achieved so far with either rotary or solid state induction generators?

Onwards - as ever.

Jules

[JulesP]

Inductive Reactance

The correct term for the increasing impedance of an inductor (coil) with increasing frequency is Reactance. This is why just ramping up the frequency gives diminishing returns. To some extend one can compensate by using a higher duty cycle with the square wave generator.

When I have the cap dump circuit operational and found the optimum coil driver BJT or FET, I will start to do tests of rate of capacitor charging and direct battery charging against frequency and consequently available net output power. All results will be posted here in due course.

Jules

[JulesP]

Coil Driver Comparisons

Hi all,

I have been experimenting with different coil drivers to obtain the best HV spikes and battery charging and partly inspired by the suggestion that using FETs can severely limit the spikes due to the 'flyback diode' incorporated in many of them between the Drain and Source.

I did a comparison of 4 drivers, two FETs and two BJTs but my tests were interrupted by some component failures so I will have to complete them again when all is back up running smoothly but I thought it might encourage some useful discussion to show what I found so far and that might come up with some more suggestions.

I have presented the data I have so far in the table below and invite feedback via some questions and comments.

1. There seems to be a distinct advantage to using a BJT over a FET with regard to the HV spike output however, those rotor systems using a Hall sensor will need to add some current to its output as, while they can drive a FET, they have little power to drive the base of a BJT. I am assuming that the 'flyback diode' in many FETs serves to short the HV spike to ground when the Drain goes positive however, the from the spec sheets, the diode would seem to be the wrong way round so I am unclear how this clipping occurs.

2. The circuit current using the MJE13009 was surprisingly small and yet overall it performed the best in battery charging. As I understand it this was in part due to the much higher voltage, and hence inducement of ZPE into the circuit, but I can imagine that a higher current would support the battery charging even better. The question is how to increase the current when it is largely determined by the combined coil resistance, in my case 0.5 Ohms, assuming that the Base is fully turned on. As mentioned previously I have noted that as the drive signal increases in frequency the overall circuit current drops and have suggested that this might be due to increasing inductive reactance of the coil. However, it has been pointed out that such reactance is only relevant to full since wave ac and not pulsed DC despite the latter having sharp rise and fall gradients that would surely invoke an inductance based impedance.

3. The HV pulse waveforms are markedly different between the FETs and the BJTs. Is this indicative of coil behaviour that is significant in any way?

4. Any suggestions of other BJT or configurations that have resulted in good battery charging or COP>1?

I will repeat these tests when my system is back running fully again but that is going to be some weeks now.

Regards,

Jules

[TruthInZero]

I am assuming that the 'flyback diode' in many FETs serves to short the HV spike to ground when the Drain goes positive however, the from the spec sheets, the diode would seem to be the wrong way round so I am unclear how this clipping occurs

That spike acts like high pressure area and kind of travels to all places provided there Is conductive path. I tried charging high voltage cap with some small SG oscillator and tried putting such protection diode over CE on BJT , and basically without diode, cap charged over 300 volts, with diode about 220 volts.

[JulesP]

That spike acts like high pressure area and kind of travels to all places provided there Is conductive path. I tried charging high voltage cap with some small SG oscillator and tried putting such protection diode over CE on BJT , and basically without diode, cap charged over 300 volts, with diode about 220 volts.

Yes that seems to concur with my findings but, as I have shown in the diagram, there is a good path for the HV to take and the position of the 'flyback diode' does not seem to interfere with that as, when the Drain/Collector goes +, the diode is in reverse bias. On that basis it shouldn't make any difference?

Julian

[JulesP]

Coil Driver 2

Hi all,

After some repairs and modifications to allow me to compare coil drivers more easily, I have returned to test the same four ones that I reported earlier. When I first tested the BJTs, I overloaded my square wave signal generator and so have now added some amplification to make up for its limited output current capacity. The relevant circuitry for the BJT is attached as 'Driver Circuit (BJT)' as this is more complicated then the one for the FET.

My aim here is to see which driver is looking to produce the best battery charging and then move on with that driver to see the effect on charging of ramping up the input square wave frequency.

The HV waveforms of the 4 transistors (attached) are the same as before except that now the waveform from the MJE13009 BJT is showing an interesting feature, an oscillation on the decay curve.

Has anyone has seen a similar effect and got any thoughts on whether it is significant, either positively or negatively, and what might be causing it?

Jules

[JulesP]

Finding the resonant frequency of coils

Hi all,

Having sorted my coil driver circuits to give enough power when using a BJT, I have been making measurements of battery charging at different HV pulse frequencies and so far not found one that gives a NET positive battery charging. Whilst I believe that the batteries are receiving ZPE via the HV pulses, the net effect with the circuit demand takes the system just below CoP=1.

The key component of most types of devices that extract energy from the 'environment' is the use of some sort of pulse that 'shakes the tree' and it occurred to me that there is going to be a resonant frequency or 'sweet spot' where the coils perform much better than normal.

Has anyone found a way to determine or estimate the resonant frequency of their coils? Alternatively, If I could measure the capacitance and the inductance of my seven coils in parallel then I can calculate the theoretical resonant frequency.

Jules

[kwag]

I've never been able to get to a COP over 1 with any Back EMF circuit. No matter what coil, MOSFET, transistor, diode I try or combinations. My goal of doing a self charged battery circuit has always failed. Even while pulsing back to the source (or single battery) via SCR or switched capacitor, it simply is always less than what was put in. Tried with NiMh batteries, NiCad and Lead Acid. Same results.

[JulesP]

I've never been able to get to a COP over 1 with any Back EMF circuit. No matter what coil, MOSFET, transistor, diode I try or combinations. My goal of doing a self charged battery circuit has always failed. Even while pulsing back to the source (or single battery) via SCR or switched capacitor, it simply is always less than what was put in. Tried with NiMh batteries, NiCad and Lead Acid. Same results.

I believe it's possible as the original South African developer that I based my rotor based system on uses his to run his house lighting. I just haven't yet found out what the tweek is!

[JulesP]

Solid State Generator Summary Report

Hi all,

At last I have completed my tests on this generator and so have written the attached 3 page summary report giving an overview of what I have done and my conclusions.

The bottom line is that I could not get the generator to perform with a CoP>1 and, although there was battery charging taking place, when the current demand of the whole circuit is taken into account then the net battery charing is negative. Raising the HV pulse frequency made no difference, possibly for the reasons I state in my conclusions.

While I believe that some researchers have been able to get a CoP>1 with a rotor based system I have not yet heard of any doing so with a solid state system and, even if a few have, it is likely that there are particular design features that have enabled that.

As a scientist I am bound to accept my findings however disappointing they are at a personal level but I am not deterred from continuing to explore. I have started on a flywheel project, based on the theory of Lead Out Gravitational energy from Lawrence Tseung and also the study done by the astrophysicist Bernard Haisch that inertia has its roots in ZPE. That being so then a flywheel generator is just another way of accessing this universal energy.

I am starting with a scale model to test the principle and, assuming I can get that to work, then I will post my findings, and how to replicate it, on this forum in the appropriate place.

Julian