COP 17 Heater | Rosemary Ainslie

[MileHigh]

Quote:
Originally Posted by MileHigh
No energy discharged from the coil goes back to the souce battery at all. In Rosemary's case, all of the current + voltage discharge from the coil goes through the resistive part of the coil and the diode.

Is this what is "supposed" to happen or did you actually verify this with this circuit?

>>>

When the MOSFET is off, and is an open-circuit, there is no closed circuit loop that would allow current to flow into the battery. If the current could flow, it would have to go from the discharging inductor then into the positive terminal of the battery, out the negative terminal, through the shunt resostor, through the MOSFET and then back into the inductor. Oops! The MOSFET is open-circuit. There is only one place the current can flow, and that's through the diode.

[poynt99]

As you know MH, nothing is ever a perfect OPEN.

With extremely fast transients and the huge parasitic capacitances around a MOSFET of this size (D-S, D-G, G-S) it is quite easy to see a path across the MOSFET, even when it is in an OFF state.

I've just probed the MOSFET Source pin and indeed it has a 400mA negative current spike on it.

.99

[Aaron]

duty cycle

_|-|________________________________|-|_______
..<------------------------------------->...<------
.................PERIOD(T) (100%)..............3.7%

Not quite Aaron. It's 3.7% of the PERIOD (T), not 3.7% of the time the MOSFET is ON.

.99

Right - that is what I meant. It would be 100% of the 3.7% the mosfet is on.

[MileHigh]

.99

Chances are your PSpice program is doing a simple modeling of a switch-on time for the diode and a switch-off and resstance time for the MOSFET, both on the order of nanoseconds. Hence you get your 20 ns spike back into the battery in your model. For those that may not be aware, a 20 nanosecond spike is very very short, and typically would contain only micro or nano-Joules of energy.

I assume that you probed your virtual MOSFET source pin. Congratulations on the report again.

[gotoluc]

I ran a quick test of Rosemary's circuit in PSpice, and it does appear that current (and power) spikes do run backwards into the battery.

These appear to be rather narrow, about 20ns in width and depending on the presence or not of the flyback diode, are: -120mA and -3W (with diode), and -1.6A and -40W (without diode).

This is with using the proper 3.7% duty switching, and without any strange high frequency oscillation. Just simple switching.

Don't jump to conclusions please. Let me play with this a bit and I will report my findings.

.99

Hi .99,

looks like you're having fun here ... I think I should get myself a PSpice may save time

I have a question regarding Zoltan Szili test document posted by FuzzyTomCat. In the document he says: the signal generator pulse must be a square, positive, with a rise time of 10 nanoseconds. The simulation shows very clearly that if the rise time of the square is longer than 10 nanoseconds, the extraction decreases very quickly and completely canceled between 50 and 100 nanoseconds.

How do you get a 20KHz pulse to a rise time of 10 nanoseconds

Luc

[Aaron]

return path wire not needed

When the MOSFET is off, and is an open-circuit, there is no closed circuit loop that would allow current to flow into the battery.

These high voltage potential spikes can be delivered and accepted to a battery with one single wire connection. That isn't necessarily what is happening here but it is a fact that a battery can charge from a single terminal without a return path for current to enter the battery.

Electrostatic potential hitting a single terminal of a battery can get the lead ions to start moving in charging mode. It is like getting an elephant to move by whipping it with a wet noodle but it will do it.

There are countless demonstrations of this and here is one simple old one I re-posted a while back:
YouTube - One Wire Tesla Transmission

Here is a cap charging with a single wire:
YouTube - Single Wire Capacitor Charge (from battery terminal)

So anyway, a wire or switch connection to the battery is not mandatory. And yes, the batteries will power a load with that potential that the battery sucks up.

edit: way more than nano-joules as you mention.

[Aaron]

overunity thread

This thread is undoubtedly over 1.0 COP. This thread has generated more heat than energy I contributed to it!

[poynt99]

.99

Chances are your PSpice program is doing a simple modeling of a switch-on time for the diode and a switch-off and resstance time for the MOSFET, both on the order of nanoseconds. Hence you get your 20 ns spike back into the battery in your model. For those that may not be aware, a 20 nanosecond spike is very very short, and typically would contain only micro or nano-Joules of energy.

I assume that you probed your virtual MOSFET source pin. Congratulations on the report again.

Indeed,

It may be the overlapping turn-ON/turn-OFF of the devices. Likely those parasitic caps are contributing quite a bit though.

It is interesting to note though that the negative current spikes back to the battery are much more powerful without the diode installed

Will do some quick analysis later to see how much RMS power there is in them anyway.


.99

[poynt99]

Hi .99,

looks like you're having fun here ... I think I should get myself a PSpice may save time

I have a question regarding Zoltan Szili test document posted by FuzzyTomCat. In the document he says: the signal generator pulse must be a square, positive, with a rise time of 10 nanoseconds. The simulation shows very clearly that if the rise time of the square is longer than 10 nanoseconds, the extraction decreases very quickly and completely canceled between 50 and 100 nanoseconds.

How do you get a 20KHz pulse to a rise time of 10 nanoseconds

Luc

Luc,

If your generator is a "real good one", it will have about 10ns rise/fall already. If not, you find or create your own design. 10ns is not impossible, but it does require some knowledge of high speed design and the right components.

I wonder if he built it for real. I had a quick glance at his figures and since it is a simulation, I highly doubt he is actually getting overunity. I will run it myself in PSpice and see what falls out from that.

If you don't use SPICE simulation correctly, it will give you incorrect results...imagine that?

.99

PS. The selection of emoticons here is refreshing LOL

[MileHigh]

.99

> It is interesting to note though that the negative current spikes back to the battery are much more powerful without the diode installed

You are back in "clobber the MOSFET" mode again, and your PSpice model may or may not be able to handle it. ALL of the inductor energy should show up in your model, but I doubt it will.

For the real-world circuit, the stray capacitance in the whole setup is going to turn the 20 nanosecond spike to mush, and the diode and resistive component of the circuit will dissipate essentially all of the coil energy. The MOSFET will not likely see a true 20 ns spike, just a pale semblance of one. The way to test for that would be to run the setup for a long time at a chosen frequency and measure the temp differential of the MOSFET, do a thermal profile of the MOSFET, etc. It will probably be less than a 1/10th of a milliwatt at 1KHz. These are just ballpark guesses.

Supplement: To address Aaron's point: In a Bedini motor the discharging inductor has a full circuit path so that current can be pumped from the inductor into the charging battery's positive terminal, and then out the negative terminal, and then back to the inductor. There is a full circuit loop that allows the current to flow so that the charging battery can be charged. That's in contrast to Rosemary's circuit discussed above where the lack of a circuit path for the discharging inductor means that essentially all of the current flows through the diode that's across the inductor. The inductor discharges through the diode because that's the only place that it can go.

[EgmQC]

YouTube - Electric OU: Critical Thinking Supplement

It cant be more clear how a Mosfet work now.

Best Regards,
EgmQC

[Aaron]

battery gets the spike

It is interesting to note though that the negative current spikes back to the battery are much more powerful without the diode installed

In the real world even the basic Bedini school girl motor with no recovery diode has an extended running time because the spike is absorbed by the battery when the transistor is shut off.

[poynt99]

The voltage is not a problem.

Spikes are up to about 540V. IRFPG50 is 1000V MOSFET, which may in fact be why it was chosen.

I can see now the possibility that with the Quantum circuit, they may have intentionally left out the flyback diode. If what you want is large spikes back into the battery, then it would appear leaving the diode out maximizes this effect.

A power dissipation measurement on the load and shunt resistors showed there is very little change with or without the flyback diode connected. Power from the supply had little change as well (not a precise measurement).

An analysis of the power spikes from and to the power supply is as follows (without the flyback diode):

RMS forward power from supply ~ 1.82W (15.2us/50W pulses)

RMS reverse power to supply ~ 27mW (277ns/-40W pulses)

PERIOD ~ 416us

That's about 1.5% power returned to the battery.

.99

[gotoluc]

Luc,

If your generator is a "real good one", it will have about 10ns rise/fall already. If not, you find or create your own design. 10ns is not impossible, but it does require some knowledge of high speed design and the right components.

I wonder if he built it for real. I had a quick glance at his figures and since it is a simulation, I highly doubt he is actually getting overunity. I will run it myself in PSpice and see what falls out from that.

If you don't use SPICE simulation correctly, it will give you incorrect results...imagine that?

.99

PS. The selection of emoticons here is refreshing LOL

Thanks for the reply .99

Looking forward to your SPICE results

I'm uploading a new video to get some opinions on a different testing method I thought of and the video should be available in about 10 minutes.

Luc

[gotoluc]

Hi All,

I have just uploaded a new video for your evaluation and opinions.

Please post your comments.

Link to Video: YouTube - Effect of Recirculating BEMF to Coil test 9

Luc