COP 17 Heater | Rosemary Ainslie

[Peter Lindemann]

Excellent Experimental Work

Dear TinselKoala,

Thank you for doing such an excellent experimental reproduction of Rosemary's original circuit. I hope Rosemary, who is watching this thread, will comment on your findings and address your question concerning the duty cycle issues.

Thank you, again.

Peter

[jibbguy]

Imo the differences here are likely more than suggested. Tinsel has reported not getting any heat output in the resistive element to speak of at the same duty cycle (3-4%); although he looks to have a very nice build indeed and his efforts are worth praise

However from my experience, i would state that an inversion error in duty cycle ("97" instead of "3") as he suggests could have happened, is highly unlikely in this case as the independent testing lab in North Carolina would have had to make the same exact simple blunder as the testers in South Africa... Time and again.

BELIEVE ME (with 27 years in the Test & Measurement field): If these great efficiency figures were reported by the first measurement cycle; they would be REPEATED with new eyes looking at them... Probably multiple times, as many would not really WANT to believe them. These guys in the independent lab and their bosses should have been getting very nervous about this test... It put them in the firing line and they knew it. So i doubt that a newbie or the guy who's clothes smelled like pot smoke would have been trusted to do this alone, at least without a repeat by another tester to verify the results

The Quantum article again:

http://www.feelthevibe.com/free_ener...ent_energy.pdf

When calcs are done on a Digital Storage Oscilloscope or data acquisition system, they are not using the same circuitry or calcs used in DMM's (which of course are notorious for averaging errors with complex waveforms at higher F's). For low voltage applications, they are almost always done by using the standard DC amplifier front end of the 'scope for the raw signal, and then doing an "area under the curve" software calculation; therefor they are pretty reliable... Although analog Frequency Response, and more importantly, Digital Sample Rate would of course always be a huge factor in accuracy: Per-Channel Sample Rate should be at least 20 times faster than the fastest transient component for something like this. Since the transient F was described at between 180 to 200KHz, this should not be a problem for the scope that was used since it is rated at "200 MHz" (although the actual sample rate used was not listed).

One nice, and potentially important, thing about this portable model is that it would be "inherently isolated" in the same way all battery-powered scopes or DMM's are, verses most benchtop models with grounded wall-plug power supplies which are "Single-Ended to Ground" and NOT officially "Isolated" (...so putting probe leads across the "floating" resistive element, or across the MOSFET, could likely do BAD things to a Single Ended scope that plugs into the wall directly... Because in the amplifier circuit of those 'scopes, there is then only a resistor between the signal and Ground)... if no damage occurred, then there could at the very least be a "Ground Loop" effect that could seriously skew results (possibly dampening the pulse?).

As far as false triggering goes, it is pretty easy to see when this is happening for those who are familiar with it. With random non-repetitive signals (as Ms Ainslie reported), it is true that they can look the same as being "off-trigger"; however this is a storage scope; and it is presumed they used the capture feature at some point which completely negates false triggering issues... By showing a snap-shot of the signal at any chosen point in time... And again, the on-board measurements made by the scope are done via calcs on the displayed raw signal so they should be accurate for snap shots as well.

Fluke 199C specs:

http://www.transcat.com/PDF/190C_Scopemeter.pdf

They reported that all the data came from the DSO (presumably from its text readouts), so i would generally trust those much more than cheap DMM's for the True RMS averaging issues; although Fluke bench meters are known to be excellent regarding "True RMS" and other pulse averaging issues with complex waveforms too... They are an Industry Leader for good reasons

Frankly, imo it would appear to be very difficult to mix up 97% and 3% when looking at the scope screen.... Especially over and over again (lol but not impossible! Worse things have happened.. The "invert" button could have been on, but you would think SOMEONE would have caught that ).

But it would of course be very helpful to get feedback from Ms. Ainslie or one of her colleagues on Tinsel's well-made points . She stated she is not available for comment on the circuit itself only on theoretical aspects; but perhaps one of the peeps mentioned in the Quantum article could be of help instead (such as Mr. Buckley)?

Also stated in the article is that the resistive element was made by a company named "Specific Heat CC", they may have some insights on the characteristics of this inductor/resistor, perhaps a spec sheet on the item is available.

[Aaron]

self-oscillation

Also stated in the article is that the resistive element was made by a company named "Specific Heat CC"

I searched and almost nothing comes up about Specific Heat CC.

Also, am I correct in what I have read so far in this thread that nobody has been able to get the mosfet into self-oscillation?

From the Quantum article: "Reducing the gate current of the mosfet results in an oscillation that overrides the predetermined frequency and duty cycle."

So basically increasing resistance to the base until it self-oscillates just like in any Bedini type circuit or similar with a transistor - to my understanding of her explanation.

The self-oscillation is said to have this difference:
From 3.7% duty cycle @ 2.4 kHz to 1.3% duty cycle @ 143 kHz to 200 kHz

[TinselKoala]

@jibbguy:
I don't know anything about other independent replications or their chances of error. All I can say is that I built the circuit shown in the Quantum article that you linked, from that diagram right there on the page (or rather a cleaned up version of it), and it produces the long ON duty cycle that I have demonstrated, and with the specified component values it CANNOT be made to produce the short one.
I found it strange and interesting that, in the EIT paper describing this experiment, when making the power calculations she nowhere describes the process as "integrating the instantaneous power waveform over time" to derive the energy, which the oscilloscope used is surely capable of doing, but rather describes a different process done in the spreadsheet--which may or may not be equivalent, depending on how the duty cycle was figured into the spreadsheet calculation. Also strange and interesting, in that article, is the omission of the problematic 555 circuitry.
(I see now that the 19oC doesn't do integration directly on board, but the ScopeMeter software, IIRC, can do it in the computer. Did they use this feature for power calcs?)

"Frankly, imo it would appear to be very difficult to mix up 97% and 3% when looking at the scope screen.... Especially over and over again (lol but not impossible! Worse things have happened.. The "invert" button could have been on, but you would think SOMEONE would have caught that ). "

So you would think. Of course, those silly Fluke-o-scopes have an annoying habit of showing you what it thinks you want to see, especially if you acccidentallly hit the "auto" button at any time during measurement.

Did you watch my second video? The inverted waveform looks pretty good.

I wonder why no scope shots from her experiments have been published, showing the waveforms.

@Aaron:
You will note that in my build, since I did not have the original article at first and did not know the value of the attenuating potentiometer in the EIT article's diagram, I used a 200K ohm pot. So I can indeed reduce the gate drive, much farther than can be done with the 100R pot specified in the Quantum article.
BTW, this 100R pot does almost nothing at all in my build. It changes some higher-order features of the waveform, which perhaps could have induced false triggering. But it certainly isn't able to induce any kind of strange behaviour in the mosfet .
Will the IRFPG50 mosfet behave differently than the 2sk1548 that I am using? There's one way to find out. But I'm not going to spend any more of my money on this until somebody can make a better case than just saying "it's the wrong transistor." So if somebody wants to send me their precious IRFPG50 for testing and comparison in my circuit, PM me and we can make arrangements.

"The self-oscillation is said to have this difference:
From 3.7% duty cycle @ 2.4 kHz to 1.3% duty cycle @ 143 kHz to 200 kHz"

Said by whom? How was this determined? Can a scope shot be found?

The load resistor as described is an ordinary high-wattage wire-wound hollow ceramic resistor, as far as I can tell, so that's what I used. The physical dimensions are right, the inductance is close.
If anybody can make a case for using something different I'd be glad to do so.

[DrStiffler]

@jibbguy:
I don't know anything about other independent replications or their chances of error. All I can say is that I built the circuit shown in the Quantum article that you linked, from that diagram right there on the page (or rather a cleaned up version of it), and it produces the long ON duty cycle that I have demonstrated, and with the specified component values it CANNOT be made to produce the short one.
I found it strange and interesting that, in the EIT paper describing this experiment, when making the power calculations she nowhere describes the process as "integrating the instantaneous power waveform over time" to derive the energy, which the oscilloscope used is surely capable of doing, but rather describes a different process done in the spreadsheet--which may or may not be equivalent, depending on how the duty cycle was figured into the spreadsheet calculation. Also strange and interesting, in that article, is the omission of the problematic 555 circuitry.
(I see now that the 19oC doesn't do integration directly on board, but the ScopeMeter software, IIRC, can do it in the computer. Did they use this feature for power calcs?)

"Frankly, imo it would appear to be very difficult to mix up 97% and 3% when looking at the scope screen.... Especially over and over again (lol but not impossible! Worse things have happened.. The "invert" button could have been on, but you would think SOMEONE would have caught that ). "

So you would think. Of course, those silly Fluke-o-scopes have an annoying habit of showing you what it thinks you want to see, especially if you acccidentallly hit the "auto" button at any time during measurement.

Did you watch my second video? The inverted waveform looks pretty good.

I wonder why no scope shots from her experiments have been published, showing the waveforms.

@Aaron:
You will note that in my build, since I did not have the original article at first and did not know the value of the attenuating potentiometer in the EIT article's diagram, I used a 200K ohm pot. So I can indeed reduce the gate drive, much farther than can be done with the 100R pot specified in the Quantum article.
BTW, this 100R pot does almost nothing at all in my build. It changes some higher-order features of the waveform, which perhaps could have induced false triggering. But it certainly isn't able to induce any kind of strange behaviour in the mosfet .
Will the IRFPG50 mosfet behave differently than the 2sk1548 that I am using? There's one way to find out. But I'm not going to spend any more of my money on this until somebody can make a better case than just saying "it's the wrong transistor." So if somebody wants to send me their precious IRFPG50 for testing and comparison in my circuit, PM me and we can make arrangements.

"The self-oscillation is said to have this difference:
From 3.7% duty cycle @ 2.4 kHz to 1.3% duty cycle @ 143 kHz to 200 kHz"

Said by whom? How was this determined? Can a scope shot be found?

The load resistor as described is an ordinary high-wattage wire-wound hollow ceramic resistor, as far as I can tell, so that's what I used. The physical dimensions are right, the inductance is close.
If anybody can make a case for using something different I'd be glad to do so.

@All
I have a problem in the accuracy of the diagram of the 555 PWM that is shown in the Quantum 2002 Article, my primary concern is Pin#6, the Threshold. It is connected to nothing but the capacitor as indicated.

555 PWM's can indeed get to low ON duty cycle rates. I use them all the time and they work quite fine if you select the values properly, as any good spec and sample sheet will provide.

For one of the simplest PWM's I have seem out there one can see this link.
DPRG: A Simple PWM Circuit Based on the 555 Timer

I doubt very, very much that a number of people made the same Electronics 101 mistake.

[jibbguy]

Small Clarification: When i mentioned "area under the curve" i was not suggesting they did a proper "Integration" as a specific measurement... I was simply "Differentiating" ( ) between the usual complaints of possible error we see with DMM's using mainly hardware-generated calcs... Which should not be a factor with DSO's that use the more reliable "soft" calcs done on board in nearly "real-time" (... RMS based mainly on calculating "Area" then further massaging the result), on the original signal trace.... Verses running the signal through an actual circuit to convert it into a conditioned DC representation before sending it to the A-to-D as most DMM's do.

Regarding the scope set-up errors you suggest could be involved here (such as letting the dreaded "Auto Setup" have it's head, which i totally agree is always a bad idea, lol...): Again we would have to conclude that the North Carolina lab people made the same exact error (and it's no "little" one, either; it's a "duzie")... Lol and one thing i do remember about instruments' "auto setups" are, that expecting them to work the exact same way twice is very dicey

But the problem of not getting the 555 to duty cycle that low is very interesting for sure.

[Armagdn03]

I find it a tid bit odd that such duty cycles are chosen for the frequency at hand. In my opinion, the duty cycle should reflect inductive time constants dictated by the inductor. The frequency then should take into consideration the time it takes for the inductor to discharge into the capacitor, which would be around (.5 * fn) of the would be tank that is created by the inductor and cap (Fn is the natural freq, one half is used because we are only interested in one half of the cycle, the discharge) . From what I see here, there is a ton of "dead" time in between cycles, this directly relates to efficiency.

Maybe things were drawn out to have measurements over a very long period of time?

Not sure, but if it takes such a small time to charge the inductor, why would it take 33times longer for it to discharge.....it doesnt, the ideal proportions are not exact to the 50 / 50 standard duty cycle, but much closer than 3% - 97%.

just a thought.

Fantastic work guys, This is a thread to be envious of. No BS, brings a tear to my eye.

[Aaron]

quantum article quote

"The self-oscillation is said to have this difference:
From 3.7% duty cycle @ 2.4 kHz to 1.3% duty cycle @ 143 kHz to 200 kHz"

Said by whom? How was this determined? Can a scope shot be found?

This is a quote directly from the Quantum magazine article.

[TinselKoala]

@All
I have a problem in the accuracy of the diagram of the 555 PWM that is shown in the Quantum 2002 Article, my primary concern is Pin#6, the Threshold. It is connected to nothing but the capacitor as indicated.

555 PWM's can indeed get to low ON duty cycle rates. I use them all the time and they work quite fine if you select the values properly, as any good spec and sample sheet will provide.

For one of the simplest PWM's I have seem out there one can see this link.
DPRG: A Simple PWM Circuit Based on the 555 Timer

I doubt very, very much that a number of people made the same Electronics 101 mistake.

Please take another look at the diagram in the article. The pin names are out of register with the wire leads. The pin 6 threshold is connected normally, to pin 2 and to the diodes D1, D2 and capacitor C3.

If there is an error in the circuit, it isn't this one. I doubt very very much if anyone would make that Electronics 101 mistake. Especially since the diagram is pretty clear.

[TinselKoala]

This is a quote directly from the Quantum magazine article.

Thanks--I see it now. It was split over two pages and I missed it.
Unfortunately my MOSFET doesn't behave this way.
It sure would have been nice to see a scope shot though.

[TinselKoala]

Small Clarification: When i mentioned "area under the curve" i was not suggesting they did a proper "Integration" as a specific measurement... I was simply "Differentiating" ( ) between the usual complaints of possible error we see with DMM's using mainly hardware-generated calcs... Which should not be a factor with DSO's that use the more reliable "soft" calcs done on board in nearly "real-time" (... RMS based mainly on calculating "Area" then further massaging the result), on the original signal trace.... Verses running the signal through an actual circuit to convert it into a conditioned DC representation before sending it to the A-to-D as most DMM's do.

Regarding the scope set-up errors you suggest could be involved here (such as letting the dreaded "Auto Setup" have it's head, which i totally agree is always a bad idea, lol...): Again we would have to conclude that the North Carolina lab people made the same exact error (and it's no "little" one, either; it's a "duzie")... Lol and one thing i do remember about instruments' "auto setups" are, that expecting them to work the exact same way twice is very dicey

But the problem of not getting the 555 to duty cycle that low is very interesting for sure.

Umm.
I have located a FLuke 199 Scopemeter to play with. And as I had recalled correctly, the "trace invert" function isn't a real button on the panel, rather it's a "check box" buried 2 levels down in a software menu and its setting is not displayed during signal measurement.

Is it possible to get access to the original report of the North Carolina lab's replication?

And I have checked, yet again, the timing component values and hookups of the 555 circuit, just to make sure my caps weren't marked wrong or misconnected or something.

Can't somebody just breadboard up the 555 circuit from the Quantum article and see if I am right about the duty cycle? I hope I have made a mistake, because this is an interesting project and I'd like to proceed to the output measurements...but if the whole claim of excess energy is based on an erroneous duty cycle setting there isn't much point, is there?

[poii]

Cap charge

I find it a tid bit odd that such duty cycles are chosen for the frequency at hand. In my opinion, the duty cycle should reflect inductive time constants dictated by the inductor. The frequency then should take into consideration the time it takes for the inductor to discharge into the capacitor, which would be around (.5 * fn) of the would be tank that is created by the inductor and cap (Fn is the natural freq, one half is used because we are only interested in one half of the cycle, the discharge) . From what I see here, there is a ton of "dead" time in between cycles, this directly relates to efficiency.

Maybe things were drawn out to have measurements over a very long period of time?

Not sure, but if it takes such a small time to charge the inductor, why would it take 33times longer for it to discharge.....it doesnt, the ideal proportions are not exact to the 50 / 50 standard duty cycle, but much closer than 3% - 97%.

just a thought.

Fantastic work guys, This is a thread to be envious of. No BS, brings a tear to my eye.

I did a SIMetrix simulation starting with a 3% duty cycle @2.4khz and there was a huge current spike on the first cycle and <10% of that on the following pulses "if " they were in phase. The current wave form on all pulses was like:

l
l
l
l
l
l
l
ll
ll
ll
l l
l l
l \___----
l



I think to the left of the flat part of the curve is the cap charge and to the right is the inductor.
It seemed like power in was minimum if the pulse was shortened to just after zero slope with little effect on the strength of the ring.


Ditto on the:
Fantastic work guys, This is a thread to be envious of. No BS.

[Allcanadian]

@All
Does anyone find it odd that so many well educated people could have so much trouble with a circuit that in it's basic form requires only five simple components? A source of potential(a low voltage DC source battery), an inductance (resistive wire wrapped in a solenoidal form), a capacitance to store energy, diodes to direct current flow and a means to disrupt or produce change within the circuit(a switch). A man by the the name of Victor Schauberger once stated that we should comprehend and copy nature but it seems we have little if no understanding of how nature charges her atoms or how such an incredible range of motions found in nature would seem effortless, why is it that we suppose to understand everything from the quantum world to the cosmos yet I am willing to bet few can tell me exactly how a simple tree can transport large volumes of water upward to it's extremities in a single day, why the wind blows, why no two snowflakes are alike?.
Concerning the circuit of Rosemary Ainslie, I have built countless circuits which utilize exactly the same effects and it should be understood that there are two distinct currents at play in this circuit, each having distinct qualities and these qualities can produce very different effects in each component at any given time. I have another question for some of you, why do you spend so much time and effort to disprove Rosemary Ainslie's circuit? It would seem many have made up their mind before they have even started. Personally I can tell you that I found success when I stopped making excuses for my failures, my personal failure to understand---as there is no one else to blame. If you can do this then you may find yourself looking back and wondering what in the hell you were thinking--this is easy, you may also find yourself wondering what everyone else could possibly be thinking, LOL.
Regards
AC

[DrStiffler]

@All
Does anyone find it odd that so many well educated people could have so much trouble with a circuit that in it's basic form requires only five simple components? A source of potential(a low voltage DC source battery), an inductance (resistive wire wrapped in a solenoidal form), a capacitance to store energy, diodes to direct current flow and a means to disrupt or produce change within the circuit(a switch). A man by the the name of Victor Schauberger once stated that we should comprehend and copy nature but it seems we have little if no understanding of how nature charges her atoms or how such an incredible range of motions found in nature would seem effortless, why is it that we suppose to understand everything from the quantum world to the cosmos yet I am willing to bet few can tell me exactly how a simple tree can transport large volumes of water upward to it's extremities in a single day, why the wind blows, why no two snowflakes are alike?.
Concerning the circuit of Rosemary Ainslie, I have built countless circuits which utilize exactly the same effects and it should be understood that there are two distinct currents at play in this circuit, each having distinct qualities and these qualities can produce very different effects in each component at any given time. I have another question for some of you, why do you spend so much time and effort to disprove Rosemary Ainslie's circuit? It would seem many have made up their mind before they have even started. Personally I can tell you that I found success when I stopped making excuses for my failures, my personal failure to understand---as there is no one else to blame. If you can do this then you may find yourself looking back and wondering what in the hell you were thinking--this is easy, you may also find yourself wondering what everyone else could possibly be thinking, LOL.
Regards
AC

@Allcanadian
At least from my view I 'am not' trying to prove anything wrong with her work. I will say that the diagram in the Quantum article must have been copied into the article as it is not going to do what is stated in the text. Some one that is versed is reading cryptic diagrams or reading between the lines may go for it, but my direction is to meet the text description.

Now having said that I might jump onto using a generator and not the 555, but in 5 minutes you can build a 555 PWM that meets the requirement, sans the article. I have seen some strange thing happen when you drive isolated gate circuits from a 555, therefore when I first saw this thread I constructed the coil to meet spec as far as L and R, the difference is I wound it on Pyrex so it would work in mu calorimeter.

I have included a few pictures, scope and coil. So far I see no oscillations, but the turn off time of the MosFet is very long, strange.

[TinselKoala]

It's rather amazing to see how this is running. A worker publishes an article with a circuit in it, in black and white. The article is presented over and over during several years, and astounding claims are made for the behaviour of the published circuit--which has appeared on sites endorsed by the original researcher.

Now, when another researcher innocently tries to build and test THE EXACT CIRCUIT published and apparently endorsed in the article, he finds that a critical feature does not perform as stated, and it has nothing to do with the MOSFET or its characteristics.

So what happens? "The circuit was inserted in the article, it's cryptic, it leaves something out..." or "It's a misprint" or something else.

But strangely, the specified "misprint" cryptic circuit is easy to build and check, and strangely, produces the exact specified frequency range and the exact INVERTED pulse width waveform...accidentally...from an inserted misprint...

OK, I agree: it's hard to argue with logic like that.

(And Dr Stiifler your scope shots are missing important details. I don't see any ringdown or inductive spikes on the trailing edge. Try turning off "bandwidth limiting."
My circuit's mosfet pulses track the input pulses well, I don't see the long turnoff delays that you have shown, at short or long duty cycles. I'm sure you will see heating since your mosfet is staying on half the time.
Have you tried building the EXACT 555 circuit listed in the Quantum article? There is a cleaned up version on overunity that may be easier for you to read, since you seem to have trouble with the Quantum version...)

And still, the issue, as far as I am concerned, is this: did the testing reported in the Quantum article use the circuit specified in the Quantum article, or not? If so, what about the inverted duty cycle? If not, what circuit WAS used and how did the wrong one get in there?