COP 17 Heater | Rosemary Ainslie

[jibbguy]

Hmm sorry to say, i am sadly out of touch with my old contacts, as i left that field about 10 years ago and went into another electronics-related one (high-speed industrial printers for packaging and "RFID"), for several years... Before just up and quitting one day from the "fortune 500" corporate life rat-race, and moving down here to the Florida Keys. I DO NOT regret it, and they can take their cell phones, laptops, and airport security lines and stick them into sunshine-challenged storage cells

Actually i am not old enough to "retire", so i often do consulting/systems integration and install work related to "off-grid living"... Mainly with live-aboard boaters and "out-islanders" seeking renewable energy systems. I also fish, and pretend to be a writer and free lance journalist. And i have been doing EE consulting work for an alternative energy project here that is hoping to be Granted soon with some serious cash so we can go into production of tide turbine generators.

I would probably still be in corporate electronics full time, but my published articles on the excesses and crimes of the bush regime lost me my Security Clearances... And around here any electronics jobs worth having are related to the Navy bases in or near Key West. Ever since i realized i was "blackballed" i've made it my mission to make sure they get their money's worth

So i am afraid i can't help much there... "Loudmouths" have very limited utility when you get down to it.

[Jetijs]

Hi all
I also decided to test this concept out. My setup is a simple PWM circuit that has an adjustable duty cycle and the lowest on times are about 4.8%. The MOSFET is IRFP360. I used a wire wound resistor, it is something like 100w and 330Ohms, I did not have anything else at hand. It is wound on a ceramic tube, the hole in the ceramic tube is perfect to slide in a ferrite rod because this allows to adjust the inductance and thus the efficiency. The pot on the MOSFET base is 10k, but turning it on max does not make the MOSFET self oscillate, maybe I need higher resistance. I used 12v12Ah batteries in series for this test, I have enough of these to go up to 200v. Anyway, here is the setup:



I recorded some scope shots, maybe this will help someone. This is the scope shot across the resistor without any recovery:
http://www.emuprim.lv/bildez/images/..._diode_36v.JPG

This is when the recovery diode is attached to the positive of the battery bank with the same setup settings:
http://www.emuprim.lv/bildez/images/..._diode_36v.JPG

This is closer look at the waveform at 36v and without any core material:
http://www.emuprim.lv/bildez/images/..._diode_36v.JPG

And this is the same setup with a ferrite core:
http://www.emuprim.lv/bildez/images/..._diode_36v.JPG

The higher the voltages, the higher also the spike. But also the ferrite core adjustment changes the amplitude of the spike, for example at 60V I managed to get up to 384v peak to peak with adjusted core position.
http://www.emuprim.lv/bildez/images/...e_position.jpg

Also the duty cycle plays a role on the magnitude of the spike. At 48v I found that 30% duty cycle makes the highest spikes:
http://www.emuprim.lv/bildez/images/...nti_1/48v2.jpg

But the higher the voltages, the smaller duty cycle is needed for maximum spikes.

Now I need to change the MOSFET base pot for higher resistance and to change the PWM for smaller duty cycles. But this is interesting
Thanks,
Jetijs

[Aaron]

metrology

"Metrology" is instrumentation calibration lab work, which explains how he gets all those wonderful toys

I can clearly see why he has problems with the "false triggering" in his mind being not what we actually want. He is used to things working as they are "supposed" to work and the self-oscillation in most applications is a nuisance.

[Aaron]

builders

Thanks for posting that Jetijs!

How many people total have posted that they have at least tried the circuit?

Skywatcher - first one right?
TK
Me
Jetijs

Am I missing anyone?

[Jetijs]

Aaron, I think that gotoluc is also missing in that list although his circuit is a bit different
I want to try something else for this, something from Peters attraction motor circuitry. I will keep you updated

[Aaron]

@Jetijs

Very much looking forward to that!

[Joit]

Jetijs i got pretty the same Shots and Voltage lately with my Coil 12V and a SK2148.

Maybe you try to place 2x12V and the Element in Serie with the Resistors, adjust it to your best Spikes, Diode in, and another 12V or 9V for the Trigger and see what it does?
I dont think, that anyone will get the same Oscillation at the Scope. i did not have it too.
For tuning the Gate i think, its maybe enough, when you get a good Point there, where it just suppress the Signal close.

And yes Aaron, you 4 are, and wonder, why not more did make a build to play around with it.
I can slap my Head, that i care to much about other Things, got mostly only few hours a Day to play with it.

[MileHigh]

Rosemary: Here are the answers to your questions about the proposed measurement system:

>> The shunt resistor should be moved to the battery postive terminal, that way you can record the voltage drop across the shunt and derive your source voltage and current with a single recording channel.

> Are you proposing to measure the resistor at the same time as the shunt? I would prefer that there is a reference to current flow remaining at the negative terminal. Maybe both?

There is only the "shunt resistor", one resistor. There is no need for two shunt resostors. Putting the shunt resistor on the battery positive is more "elegant", with the shunt resistor on the battery negative you are "disturbing" the potential at the gate input of the MOSFET because it's source pin is "wiggling" do to the shunt resistor being in this place.

This all assumes that your battery voltage will be constant throughout the test, which is a reasonable assumption. It should still be spot-checked a few times during the test anyways with your multimeter. Personally I would still go with a "real" variable bench power supply and forget about the battery.

> Since the dissipation of heat is the critical factor for power dissipated I would recommend platinum based probes? Not sure if they're more expensive but they're not influenced by applied high frequencies. But you can then insert them into the hollow body of the resistor and hold it with plastic putty or somesuch. That way no interference with drafts that really do influence temperature.

I don't know anything about thermocouples so I can't help you there. I don't see any high frequency issues here affecting the probes.

To slightly refine my setup, if you are using a hollow tube-type coil-resistor then it should not hang horizontally between the wires, it should be vertical or at least hang at 45 degrees. This will generate a stable, repeatable chimney effect that will lower the thermal resistance and reduce the settling time a bit.

You should should not insert the thermocouple inside the tube, it should be on the outside.

Convection drafts around and through the coil-resistor are not bad for the setup as long as they are let to form naturally and you don't interfere with them. These convection drafts are an integral and very important part of the setup.

> Excel analysis really only needed to confirm displayed measurements across the shunt and to prove aliasing if applicable. Once the spreadsheet and the machine numbers marry - you've got a dependable voltage. But you also need more than 1 sample per cycle. I would recommend between 5 to 8 especially if you're using a Fluke. Tektronix or somesuch can take even more samples to establish a fair average.

You are not getting it here Rosemary. The Excel analysis is the heart of your power measurement. If you can't figure that out get someone to explain this in more detail for you. There should be no aliasing as long as you sample at a high enough rate.

> Once the spreadsheet and the machine numbers marry - you've got a dependable voltage.

This does not make any sense in the context of this discussion.

> But you also need more than 1 sample per cycle. I would recommend between 5 to 8 especially if you're using a Fluke. Tektronix or somesuch can take even more samples to establish a fair average.

Again, in the context of what I am proposing, this does not make any sense. If you are talking about sampling the shunt waveform over one full cycle, at least 1000 points would be nice. The more points the better your data set is to work with. The oscillator frequency is 2.4 KHz, therefore if you want 1000 points per cycle your data acquisition system has to sample at 2.4 MHz, which is not an excessively high sample rate at all to get good data. If you have ringing and you want to make sure you don't subsample that, you may have to acquire at 10 or 20 MHz. You have to scope the waveform to see the highest frequency components. I have to assume that most PC-based scope systems can easily sample at the 2.4 MHz rate.

> Control on heating is more readily established using a variac or somesuch. But I may be wrong here.

A variac is not used here. You want a bench power supply with a 20-turn or at least a 10-turn voltage adjustment pot. What might be even better would be to use an operational amplifier tied to a voltage reference IC and make your own voltage source with "coarse" and "fine" adjustment trimpots. You may be able to get much better source voltage control this way and it would be ridiculously cheap compared to a bench power supply.

> But it must be understood that the measure of energy out must be in line with the difference in voltage over a sample range so that the power in can be compared to the power out.

Sorry Rosemary, but I have to keep on being tough on you here. I can't make heads or tails of what you are saying here. It makes no logical sense in the context of what we are trying to accomplish here. This is a critical test to make - to see if your circuit is OU or not, and you have to understand what is going on. Perhaps other people in the forum will help you because I am not guaranteeing any replies. You have to fully understand how the power in and power out are being measured.

> And the diode must be attached to the positive terminal of the battery with no other resistance in its path. We don't need that heat to be added to the heat value.

Note the positive of the battery is the same node as the other side of the coil-resistor. Not to fight, but truly, the coil-resistor discharges across the diode and back into itself. You will just have to keep exploring this circuit and one day you may arrive at the same conclusion. Your comment about "heat being added to the heat value" makes no sense.

> In any event there is the likelihood that it will then be heated to a greater degree as a result of its proximity to the load resistor thereby detracting from rather than adding to the heat.

I think that someone should walk you through how the test works step by step, explaining the rationale behind how it works. Again, the statement above makes no sense in the context of what you are trying to achieve here.

Your questions about this critical test have been answered Rosemary. I am sorry again for being tough on you but this is the "real thing" here. You have to understand what's going on and how it works, there is no room anymore for playing around, you are in the "big leagues." I hope that your friends in the forum can help get you up to speed and good luck.

MileHigh

[Aaron]

Batteries Are Required

Personally I would still go with a "real" variable bench power supply and forget about the battery.

Battery is required to prove the gain. For the fact my battery went up is a gain. For the fact that the shunt shows the reverse spike going to the battery is a gain.

Take out the battery and you will not have the means to soak up the gain.

Everyone, please use ONLY BATTERIES as they are the only thing that will show you Over 1.0 COP.

There may be a gain in the heating element and if there is, what is the point of allowing other gain to be dissapeared into nothingness?

The tests VALIDATED by Fluke, BP, etc... used BATTERIES. If anyone wants the same kind of results. Use Batteries.

Also the analogy given before about comparing the coil charging to charging a cap is misleading because when charging the cap, you just charge the cap. With charging a coil, WORK IS DONE in creating HEAT, then you still get the discharge back.

[Joit]

The tests VALIDATED by Fluke, BP, etc... used BATTERIES. If anyone wants the same kind of results. Use Batteries.

So true.

But seems we got the same Situation as you did describe with your Friend and the Magnet.
After you show it someone, he runs to the Table, calculate something, and says, No, that is not possible.

[poynt99]

Also the analogy given before about comparing the coil charging to charging a cap is misleading because when charging the cap, you just charge the cap. With charging a coil, WORK IS DONE in creating HEAT, then you still get the discharge back.

That is just plain incorrect.

Heat is generated while charging a capacitor in MUCH the same manner as when energizing an inductor, but inductors are more lossy due to their finite DC resistance.

.99

[MileHigh]

Aaron: I will rebut your points just once. We are probably going to have to agree to disagree. The last thing we want is for the thread to descend into another whizzing contest.

Re: Battery required

The shunt resistor will record any current going back into the battery or power supply by going voltage positive. One school of thought says there is no current or perhaps a very brief spike of current going into the battery. Another school of thought says there are regular pulses of current going back to the battery or power supply. The grab of the data points across the shunt resistor will definitvely resolve this issue one way or the other.

From the paper:

>
The temperature rise above ambient stabilized at 52 degrees Celsius
when the applied dc voltage from the variable power supply was set at 13.32
volts. This represents
R
V 2
, 13.32 x 13.32/10 = 17.74 watts. Results indicated
that an average of 17.74 watts was dissipated at the start of the test period.
>

You can't change horses in mid stream and bet on the battery. You are testing against the claim made in the paper.

If you did go the battery route in a SECOND round of testing, you can not use the battery voltage as an indicator of the amount of energy in the battey. That is simply not true. I just said the other day how the battery can simply heat itself up as it drives a load and therefore increase its output voltage because the chemical reactions in the electrolyte run more actively. The voltage in a battery is essentially constant over 90% of its charge state. Voltage is not in any way an accurate meaure of the battery's energy state, nor does it mean that the battery energy went up just because the voltage went up. An extensive round of load testing would have to be done. Cycling the batteries, measuring how much energy they contain, establishing a base line, and then running the setup and then running a suite of before-test and after-test battery charge and discharge cycles to see if any energy went into the battery. It is totally beyond the scope of what was presented in Rosemary's paper.

> Also the analogy given before about comparing the coil charging to charging a cap is misleading because when charging the cap, you just charge the cap. With charging a coil, WORK IS DONE in creating HEAT, then you still get the discharge back.

When you charge a cap, work is done. The work is done by the battery. The cap stores that work. Charging a cap is identical to filling an air tank with air.

When you charge a coil, work is also done. The work is done by the battery. The coil stores that work, and no heat is created. Heat x time represents lost energy. Charging a coil is identical to compressing a spring.

If you disagree, that's fine. I am not going to sweat it. The real issue is that Rosemary's device as presented in her paper produces more heat output power than battery or power supply input power. Any specuilation that the battery is gaining in energy can be verified by the shunt resistor waveform capture. You will be able to compute the power going back into the battery with the Excell spreadsheet. If you think that the battery is "multiplying" that power or using it to somehow capture energy from the environment, all of that discussion will have to wait for a whole second round of testing after the first claim is tested.

MileHigh

[Aaron]

self oscillation

Plus, since the mosfet isn't turning fully on before you turn it off and start the ringing, the current draw goes down and the batt voltage goes up.

Now turn the gate up or the duty cycle longer. The mosfet turns on fully so the current goes up and the batt voltage goes down. And the scope's trigger sees a cleaner rising portion of the wave and locks it in.

Ramset,

Is TK willing to show what his shunt shows during the "false triggering" or not?

TK is claiming in the "false triggering" mode that the mosfet can't turn on completely because it is being turned off too fast.

What kind of analysis is this?

If what TK says is true, then the mosfet can't conduct. If the mosfet can't conduct, then no current can leave the battery. If no current leaves the battery, there can be no waveform on the shunt.

Ramset, I think this needs to be posted at ou.com on the thread to
debunk the disinformation. Ramset, why not post the below picture into
the heater thread at ou?

Please ask him to do the same test.



For everyone else, this should give you confidence that when the noise is
on the scope, the battery is delivering exactly what it is predicted to do.
It is delivering high frequency pulses.

[Joit]

@Milehigh
Sorry, to disagree but
"Charging a cap is identical to filling an air tank with air."

I read lately, when you want to store Energy efficient, then store it as Air or Water in a Tank, not in a Cap.

And surly, the Voltage is not the right Indicator in a Batterie, the Amps are, what is.
But, when you put away the flyback Diode, ie, as Aaron did, and Energy runs out as like from a Pot with no ground inside,
isnt that a logical clean Proof?
Or your Battere is are drawn down after 2 Hours without the Circuit.

But yes, at J. Newman was this no Proof too, that he got still full Batteries beside let his Device running.

[EgmQC]

Battery is required to prove the gain. For the fact my battery went up is a gain. For the fact that the shunt shows the reverse spike going to the battery is a gain.

Take out the battery and you will not have the means to soak up the gain.

Everyone, please use ONLY BATTERIES as they are the only thing that will show you Over 1.0 COP.

There may be a gain in the heating element and if there is, what is the point of allowing other gain to be dissapeared into nothingness?

The tests VALIDATED by Fluke, BP, etc... used BATTERIES. If anyone wants the same kind of results. Use Batteries.

Also the analogy given before about comparing the coil charging to charging a cap is misleading because when charging the cap, you just charge the cap. With charging a coil, WORK IS DONE in creating HEAT, then you still get the discharge back.

Why not use a capacitor for the input ? A battery act differently at each plateau , a capacitor discharge curve is predictable so at any moment you will know exactly how much power is in it, no more question about fluffy charge. If that dosent work with a capacitor then its a chemical process who make the magic and not the circuit directly.

Best Regards,
EgmQC