GM East Burst Heater Circuit

[MarkE]

Greg has updated his slide-show presentation:  http://www.youtube.com/watch?v=0fYCZsIsi7c&list=TL9a9XJC8CKzXE4C9ZZ2Ig-B2F-rpd-eX4

Things that Greg did well: 
Meticulously collected data over extended periods of time using equipment he believes to be reliable, and which should be.
Devised means to calibrate his measuring equipment:  Specifically the resistance of his CSR.
Devised null tests to compare the results of known or reliably determined measurements against measurements with his DUT.
Devised more than one null test to verify consistency in his results.

And his bottom line: 

1) For a given amount of battery voltage depletion using similar average current, Greg measures delivered battery energy using a pulse circuit that is 125% that of a DC circuit that drains the batteries by a like voltage over a similar period as a DC load.
2) For a DC current that generates the same rate of battery voltage depletion as the pulse circuit, Greg measures a significantly lower temperature rise:  29.7C versus 34.5C on the inside of his heating element.

So, what has Greg observed?  Is it:

1) A pulse circuit generates over unity results?
2) A pulse circuit does not generate over unity results?
3) A pulse circuit delivers more heat to a resistor than the Iaverage^2*R*DutyCycle?
4) A pulse load of some average power Pave_pulse discharges a lead acid battery at a lower rate than a DC load with the same average power?

Is it none, one, or some combination of the above?

What tests can Greg run that will either reinforce or dispute his conclusion that he has a 125% efficient configuration?

[gmeast]

Greg has updated his slide-show presentation:  http://www.youtube.com/watch?v=0fYCZsIsi7c&list=TL9a9XJC8CKzXE4C9ZZ2Ig-B2F-rpd-eX4

Things that Greg did well: 
Meticulously collected data over extended periods of time using equipment he believes to be reliable, and which should be.
Devised means to calibrate his measuring equipment:  Specifically the resistance of his CSR.
Devised null tests to compare the results of known or reliably determined measurements against measurements with his DUT.
Devised more than one null test to verify consistency in his results.

And his bottom line: 

1) For a given amount of battery voltage depletion using similar average current, Greg measures delivered battery energy using a pulse circuit that is 125% that of a DC circuit that drains the batteries by a like voltage over a similar period as a DC load.
2) For a DC current that generates the same rate of battery voltage depletion as the pulse circuit, Greg measures a significantly lower temperature rise:  29.7C versus 34.5C on the inside of his heating element.

So, what has Greg observed?  Is it:

1) A pulse circuit generates over unity results?
2) A pulse circuit does not generate over unity results?
3) A pulse circuit delivers more heat to a resistor than the Iaverage^2*R*DutyCycle?
4) A pulse load of some average power Pave_pulse discharges a lead acid battery at a lower rate than a DC load with the same average power?

Is it none, one, or some combination of the above?

What tests can Greg run that will either reinforce or dispute his conclusion that he has a 125% efficient configuration?

Hi MarkE,


Thanks for your input. It is clear you put some time into reading what I have done and it is also clear that you are sincere. For that I thank you.


But I'd like to correct you on one point. My temperature measurements did not involve only the temperature of the 'inside' of the element. The temperature measurement was not a temperature 'rise' as you referred to above, but a more valid 'temperature differential' ... the constantly maintained temperature between the inside of the element and the ambient environment tending to bias the element's temperature proper. This is a much more reliable technique than attempting to measure a temperature rise in these tests. Things are able to establish an equilibrium.


As I have indicated, my YouTube slide show presents an "exploration". In the accompanying description I acknowledge possible 'problems' in the collected data when using batteries as a power source. Those problems are multi-pronged and include battery temperature, chemistry, battery age and number cycles the batteries have been subjected to (maybe again ... age here). In general, these 'problems' are associated with, what is referred to as, "The Battery Effect" ... a term used here and in other forums.


My next slate of tests will be to replace the batteries with a capacitor bank comprised of caps having low ESR. To my surprise (at first) it is neither super nor ultra capacitors. Aluminum and Tantalum caps seem to be the choices. However, because of a capacitor's discharge curve, my cap bank will need to be HUGE both in series and parallel size. My goal is to construct a bank of caps that will give me the same 8-hour test as in my slide show and with a drop in voltage from 35VDC of only 1VDC under a 3.5 Watt load. Upon doing the math, I see it is financially impossible for me. But something like this needs to be done to eliminate "The Battery Affect".  So the answer is to scale the tests down to perhaps 1 or 2 hour tests ... at that, the cost is still in the $1000's.


All of this hints that any useful system would need to be 'cyclic' ... charge the cap bank (with something conventional) and then discharge the cap bank through the heater and circuit. And this mode of operation, unfortunately,  will introduce additional losses that will, at least, need to be accounted for.


Thanks MarkE.  Sincerely,


Greg

[MarkE]

What you call a temperature differential I understand to be the temperature rise of the inside of the heater over ambient.  In any event, from a purely qualitative standpoint, it showed that under the two test conditions there is reasonable evidence that more heat is evolved from the pulse circuit configuration than the DC condition that leads to an equal rate of battery voltage decay.

I can suggest tests that you can run that will not require the up front expenditure that you contemplate for capacitors.  Of course if you really have found an over unity condition a self-running test will be an acid test proof.  But if you can generate strong evidence without the expenditure then it might make it easier for you to find help covering the expense for that ultimate test, if you really have over unity.

If we return to the set of questions, the big one that you ultimately want to get to is whether or not you have more energy evolved from the heater than energy you have drawn from the power source.  The proxy that you have so far used for your power source is battery voltage decay.  If you have access to a 30V regulated power supply that can output at least 0.5A, then you can proceed by substituting a couple of configurations of the power supply for the battery bank.

In one configuration, you would connect the power supply in place of the batteries, and perform an initial test to qualify droop during the pulse on-time.  The idea is to insure that there is enough bulk capacitance to limit droop to less than 1%.  I did not see mention of the pulse width or rate in your presentation.  Your peak current is less than 500mA, so you can measure the power supply droop during the pulse on period.  If it is more than 0.25V, then add a 50V electrolytic capacitor that is at least:  2uF * Ton of your pulses, where Ton is in us.

From the measurements that you presented I derived a value for the current sense resistor of 49.86mOhms, and an average peak current for pulse operation of: 467.9mA, and an RMS current of 229.2mA.  I solved from: 27.59V/(CSR+245.6)=5.6mV/CSR, and a stated duty-cycle of 24%.  Among the assumptions that we are making is that the 5.6mV read by the DMM is a faithful average of the voltage across the CSR.  As I have previously posted, as long as the pulse frequency is more than 100Hz, DMM DC voltage readings are generally within 1% of the true average.  1% is certainly accurate enough for what we are looking at.

A first test would seek to determine whether the calculated RMS current produces the same heating effect as the pulse circuit.  If it does that suggests that the current through the heating element is what existing circuit theory suggests for a periodic rectangular waveform:  Irms = Iave/(Duty-Cycle)^0.5.  Because of the losses in MOSFET and the recirculation diode, existing circuit theory predicts that you will need slightly less current to reach the same temperature rise over ambient using DC than the pulse circuit. 

A result close to or under 229.2mA suggests that what you are seeing has to do with the battery being loaded by DC versus pulses.
A result close to 256.2mA reinforces the idea that the pulse circuit is producing 125% as much heat as existing theory predicts.

A second test would seek to calibrate heat output more rigorously.  This would give you a direct measure of the amount of power the heater resistor dissipates.  For that you want to measure the amount of heat conveyed to a thermal mass that is much larger than the thermal mass of your wirewound heater resistor.  A thermos that is large enough to hold the resistor and at least five times the resistor's weight of mineral oil is a good choice.  You would drill a hole in the lid for your thermocouple probe.  The probe would contact only the mineral oil.  You can manually agitate the oil during tests.  The thermal time constant will depend on how much oil you have, IE how big the thermos is.  You should be able to reasonably get through one test in half an hour or less.

[gmeast]

What you call a temperature differential I understand to be the temperature rise of the inside of the heater over ambient.  In any event, from a purely qualitative standpoint, it showed that under the two test conditions there is reasonable evidence that more heat is evolved from the pulse circuit configuration than the DC condition that leads to an equal rate of battery voltage decay.

.....................................................................................  If you have access to a 30V regulated power supply that can output at least 0.5A, then you can proceed by substituting a couple of configurations of the power supply for the battery bank.

In one configuration, you would connect the power supply in place of the batteries, and perform an initial test to qualify droop during the pulse on-time.  The idea is to insure that there is enough bulk capacitance to limit droop to less than 1%.  I did not see mention of the pulse width or rate in your presentation.

Hi MarkE,


I have a 30VDC regulated power supply. It's the one used and shown in the presentation. In the first graph showing the battery voltage plot, the text pointing at the middle of the curve states a 24% Duty Cycle. I did not post the frequency anywhere in the presentation so it would be hard to determine the PW, so here it is now: the frequency was 434,000Hz. So the pulse period was 2.3x10-6sec making the PW=5.53x10-7sec @ 24%D.C..


One other thing though: In some of Tesla's stuff I recall that some of his devices would not work if the system was grounded. I believe that may hold true in these sorts of systems where rapid edge-transitioning electromagnetic waveforms are characteristic. I honestly believe I have witnessed this phenomenon in my tests. It's possible that these sorts of systems are required to be isolated during OU operation. It's also possible it's the reason you can produce Heat-Equivalent OU performance and NOT Electrical OU performance. Thus it's the reason you can't use a regulated power supply in place of the batteries ... even if it's sitting on an isolation transformer. My beliefs are that the increased performance is not centered in the batteries, rather it is centered in the heater element itself. Allot of the Cold Fusion and LENR research use Noble Metals and metals having similar attributes. Nickel is one of those metals. The generation or presence of magnetic domains is also important in CF and LENR research. My heater element is air-core and the winding is NiCrFe wire, and the Iron provides support for the magnetic domains ... or however it is to be properly stated.


I'm pretty happy with the direction I'm taking my next slate of tests.


Regards,


Greg

[MarkE]

Hi MarkE,


I have a 30VDC regulated power supply. It's the one used and shown in the presentation. In the first graph showing the battery voltage plot, the text pointing at the middle of the curve states a 24% Duty Cycle. I did not post the frequency anywhere in the presentation so it would be hard to determine the PW, so here it is now: the frequency was 434,000Hz. So the pulse period was 2.3x10-6sec making the PW=5.53x10-7sec @ 24%D.C..


One other thing though: In some of Tesla's stuff I recall that some of his devices would not work if the system was grounded. I believe that may hold true in these sorts of systems where rapid edge-transitioning electromagnetic waveforms are characteristic. I honestly believe I have witnessed this phenomenon in my tests. It's possible that these sorts of systems are required to be isolated during OU operation. It's also possible it's the reason you can produce Heat-Equivalent OU performance and NOT Electrical OU performance. Thus it's the reason you can't use a regulated power supply in place of the batteries ... even if it's sitting on an isolation transformer. My beliefs are that the increased performance is not centered in the batteries, rather it is centered in the heater element itself. Allot of the Cold Fusion and LENR research use Noble Metals and metals having similar attributes. Nickel is one of those metals. The generation or presence of magnetic domains is also important in CF and LENR research. My heater element is air-core and the winding is NiCrFe wire, and the Iron provides support for the magnetic domains ... or however it is to be properly stated.


I'm pretty happy with the direction I'm taking my next slate of tests.


Regards,


Greg

Greg, so if your pulses are only 553ns wide, at a peak current of 0.5A it will only take a 1uF capacitor to keep the voltage droop under 0.25V.  There is certain to be at least that much capacitance at the output of your bench supply, so no supplemental capacitor should be needed.

I do not doubt that you've witnessed what you have measured.  I think the task now is to see what it means by conducting some additional tests.  For example the heat transfer test that I suggested would allow you to directly measure the heat output of the resistor, so you can compare that against the measured input energy over some period of time.  If you have a power gain for any reason that will show up as more thermal power output than the RMS power input.  You can perform the thermal test using batteries, so there is no change that would affect isolation.  Mineral oil has a uR of 1.0, and the eR is only 2.1.  That means the parasitic capacitance will be about twice what it was in free air, still a very small value.

Other things that you could do before you go out and buy a giant phalanx of capacitors is to use a relatively small capacitor, say 10uF @ 50V and a Schottky diode.  The Schottky diode would go in series with the battery red lead, anode to the battery+, cathode towards the test circuit and the capacitor would go across the test circuit power input side of the wiring.  You could then put your scope across that capacitor, and look for evidence of recharging.  If you do that test, make sure to disconnect any other oscilloscope channels. 

Another thing that you can do that will help is that if you are using 4" ground clip leads that are standard with most scope probes, then buy some 75 Ohm resistors and solder a resistor right at wherever you normally clip the scope probe in your circuit, and then clip the scope probe hook onto the free end of the resistor close to the resistor body.  Little 1/8 W to 1/4 W resistors are good because they are small.  That will largely damp out the inherent resonant circuit formed by the scope probe input capacitance and the inductance of the 4" ground clip.  Also, use the 10X setting on the scope probe. I have attached a couple of scope captures that show the improvement in waveform quality that results before and after adding a 75 Ohm resistor when measuring a 10MHz, 1ns rise / fall time signal using 200MHz probes and a 200MHz bandwidth oscilloscope.  The green traces are using Chinese probes I bought for $12.50 each, and the blue traces are using probes that cost $150. each.