COP 17 Heater | Rosemary Ainslie

[Harvey]

Thanks Harvey. In fact with due respct to DonL - your translation has the edge here.

we both used the same translate.google.com why does my link have the frames?

Edit: nevermind - he linked Pardoxe, I linked Genese

[witsend]

Sorry DonL. I think I was just trying to tell Harvey that his efforts were appreciated here. But clearly my efforts weren't.

There is absolutely no difference between the two and I'm very grateful for your help. I found the link very interesting and have answered you at OU.Com on this.

[poynt99]

It makes sense to me. But then I come from an era before Exabyte storage was within grasp. So when dealing with Kilobyte or even Megabyte storage you only take a sample reading often enough to 'snapshot' the points along the way that allow you to connect the dots. The original signal was ~15µs with ~400µs of off time in between. It would have seemed reasonable to use a 1µs sample interval. This means you take a sample snapshot every microsecond or specifically, every 1.2 microseconds. That should give 15 samples per on time. Even with 200KHz, it should have given a minimum of 5 samples per cycle - plenty.

I think perhaps your confusing the data collection interval with the scope time base.

We'll not know for certain without the actual data from the Excel spreadsheet. In any case, IMO it is not written in such a way as to be absolutely clear. In the one case I cited it is specifically written as 1us/div. Very clearly that is referring to a scope trace.

.99

[poynt99]

SRC is the mnemonic for Source. You are mistaken regarding the impedance, please look at the 'z' symbol. This is the symbol for impedance. It is a necessary function to ensure constant voltage where the source current is varying. The chart I posted is a direct Transient Analysis of the VSRC impedance during a 560µs run. In this case I had shortened the off time considerably so there are many cycles from the 555, and the ring 'tone' is ~360kHz. Where do you think the 80K Ohms is coming from if it is not modeled in the VSRC?

As my head hit the pillow last night I realized the obvious (not at the time) "SRC". Anyway...

There is no "Z" symbol anywhere related to the VSRC or ISRC components. They are pure DC. I know what "Z" means.

I think you are confusing "modeling" with "ideal component operation". Modeling in SPICE requires a subcircuit made up of components, real or behavioral in combination to simulate how a real component would work. In certain cases such as for MOSFETs, there are several mathematical models, each with different approaches which are incorporated into SPICE programs. VSRC and ISRC have no such modeling associated with them at all.

The VSRC and ISRC components work on a pure math relationship, and don't exist in reality. The "ideal" DC voltage source always has an output resistance (not impedance strictly speaking) of zero Ohms. What you were plotting was in fact the "impedance" of your total load, not the voltage source. The voltage source output resistance is always 0 Ohms, and plotting its resistance gives an "apparent" output resistance only.

Being an ideal voltage source it has unlimited current capability, and conversely, the constant current source has an unlimited voltage capability (in theory until the math processor over-runs).

So again, the constant voltage source has no modeling associated to it whatsoever, as it is "ideal" and will deliver whatever current is required as per the load placed upon it. Don't mistake this as having a varying output resistance, because it does not. It is always "0". The resistance (or impedance) you are measuring is strictly determined by and inherent in the load the constant voltage source is attached to.

.99

[FuzzyTomCat]

Circuit Test

Hi everyone,

I'm back and sorry I was unable to answer questions directed at me, presently it's time for me to get together all the results, information, photos and data from the extensive testing of the "Ainslie-Murakami Negative Dominant Waveform Generator" circuit that has been done in that last few days

I'll be posting it later, theres a lot of information to compile ......

Glen

[poynt99]

I've had to re-post this for further comments as the "edit" function does not appear to be working at the moment.

Originally Posted by Harvey
It makes sense to me. But then I come from an era before Exabyte storage was within grasp. So when dealing with Kilobyte or even Megabyte storage you only take a sample reading often enough to 'snapshot' the points along the way that allow you to connect the dots. The original signal was ~15µs with ~400µs of off time in between. It would have seemed reasonable to use a 1µs sample interval. This means you take a sample snapshot every microsecond or specifically, every 1.2 microseconds. That should give 15 samples per on time. Even with 200KHz, it should have given a minimum of 5 samples per cycle - plenty.

I think perhaps your confusing the data collection interval with the scope time base.

If the assumption is made that the stated 1.2us is indeed the sampling interval, then I would add:

There was some discussion early on about the sample rate used to acquire the data. It was pointed out that 5 samples per cycle was not adequate to obtain an accurate measurement of the signals at hand, and I agree. With the transient and complex nature of the wave forms present on the shunt, 5 samples across the whole cycle won't capture a sufficient portion of the wave form. The 200kHz may also be the oscillation frequency only, and may not include the resonant ringing present as well, which with Aaron's circuit has exceeded 5MHz. Not capturing the necessary nuances will leave the measurement unuseable. As a minimum, a sample rate of 10MHz (100ns sampling interval) is required to reliably and accurately capture the data. A 1.2us sampling interval is not adequate here and could skew the results.

This is important from the point of view that this data was used to obtain the "net power used" figure of 1.13W. I contend that this power will indicate a much higher value when the data to calculate it is acquired properly. When I test the circuit, the input power will be demonstrated.

.99

[Harvey]

As my head hit the pillow last night I realized the obvious (not at the time) "SRC". Anyway...

There is no "Z" symbol anywhere related to the VSRC or ISRC components. They are pure DC. I know what "Z" means.

I think you are confusing "modeling" with "ideal component operation". Modeling in SPICE requires a subcircuit made up of components, real or behavioral in combination to simulate how a real component would work. In certain cases such as for MOSFETs, there are several mathematical models, each with different approaches which are incorporated into SPICE programs. VSRC and ISRC have no such modeling associated with them at all.

The VSRC and ISRC components work on a pure math relationship, and don't exist in reality. The "ideal" DC voltage source always has an output resistance (not impedance strictly speaking) of zero Ohms. What you were plotting was in fact the "impedance" of your total load, not the voltage source. The voltage source output resistance is always 0 Ohms, and plotting its resistance gives an "apparent" output resistance only.

Being an ideal voltage source it has unlimited current capability, and conversely, the constant current source has an unlimited voltage capability (in theory until the math processor over-runs).

So again, the constant voltage source has no modeling associated to it whatsoever, as it is "ideal" and will deliver whatever current is required as per the load placed upon it. Don't mistake this as having a varying output resistance, because it does not. It is always "0". The resistance (or impedance) you are measuring is strictly determined by and inherent in the load the constant voltage source is attached to.

.99

.99

I don't know how you ever got any degrees arguing with your elders this way, but I'm going to tell you flat out - you are wrong. Furthermore, you are digging a hole here that you can't possibly get out of. I posted the information regarding the Impedance plot, and if you look very carefully just above the information window you will see 'vbatt2(z). Now, go back and look at the plot itself in the upper left corner - same thing. Now go back and read what you have to do to measure the circuit impedance.

Now go look at your bench power supply. Is it voltage regulated? How does it regulate that voltage? By changing the impedance of the regulating medium.

Now - go back to school, and quit trying to get free tutoring from me - I am not your professor and you certainly have not paid the required tuition.

[poynt99]

Rose, regarding the above response and our Skype chat today (Friday), I rest my case. Strengthened even more by a few subtle yet un-called-for and un-provoked signature indignities thrown in that are truly only Harv's.

Need I say more?

.99

[Harvey]

I've had to re-post this for further comments as the "edit" function does not appear to be working at the moment.



If the assumption is made that the stated 1.2us is indeed the sampling interval, then I would add:

There was some discussion early on about the sample rate used to acquire the data. It was pointed out that 5 samples per cycle was not adequate to obtain an accurate measurement of the signals at hand, and I agree. With the transient and complex nature of the wave forms present on the shunt, 5 samples across the whole cycle won't capture a sufficient portion of the wave form. The 200kHz may also be the oscillation frequency only, and may not include the resonant ringing present as well, which with Aaron's circuit has exceeded 5MHz. Not capturing the necessary nuances will leave the measurement unuseable. As a minimum, a sample rate of 10MHz (100ns sampling interval) is required to reliably and accurately capture the data. A 1.2us sampling interval is not adequate here and could skew the results.

This is important from the point of view that this data was used to obtain the "net power used" figure of 1.13W. I contend that this power will indicate a much higher value when the data to calculate it is acquired properly. When I test the circuit, the input power will be demonstrated.

.99

I for one (and I am sure there are many other silent readers here) will be very interested in seeing how the "resonant ringing present as well" makes through the "OFF" HEXFET, across the 'shunt' resistor and into your data collection that you intend to 'acquire properly'.

Your raising grains of salt issues of timing regarding a measurement process that was flawed to begin with. Surely the subtraction of power should turn on some red lights in you alarm system - especially since I have already tipped my hand to it. The only way we can possibly subtract 'power' from the equation is if somewhere we show a negative resistance.

But since you raise the matter of timing, lets look at the possible pulse width of one of these "OFF" period reverse conduction events. Nine Hundred Fifty Nano Seconds. That's the diode recovery time. 950ns. Just a little over 1.05MHz. Think you can get anything faster through there and be meaningful? And if you do, where do think its being dissipated?

I'm feeling a bit facetious - please show us how its done


BTW - use 'go advanced' if you want your edits to stick.

[witsend]

Rose, regarding the above response and our Skype chat today (Friday), I rest my case. Strengthened even more by a few subtle yet un-called-for and un-provoked signature indignities thrown in that are truly only Harv's.

Need I say more?

.99

Poynt - you're on your own here. I think Harvey has been excessively tolerant and have told him so. I think the term is that you're 'punching above your weight?' Something like that.

EDIT - 'unprovoked signature indignities'? Never heard you mention this - certainly not to me.

[poynt99]

you are the dude harv


.99

[Harvey]

First Video

Negative Current in the Positive Circuit Leg

It has been a bit of a process to get this done. I was going to use Fraps, but evidently I never paid for it. I then found CamStudio. This free open source program worked surprisingly well, but it hammers the processor and makes dynamic recording of a simulation being processed almost impossible. So I had to prepare the simulations before hand. Then of course there were Audio to Video timing issues that had to be worked out. Then, I made the first recording full of details and it ended up being 600MB and 23 minutes. Attempting to split it into three 8 minute segments, failed in windows movie maker and Virtual-Dub. So I re-recorded the video, watching the time to keep it under 10 minutes. Finally got it right (sort of) after 3 takes and still had to add some post processing edits in You-Tube. Life would have been so much simpler if I had just stayed a Janitor.

[Aaron]

negative current

5 Star video Harvey!

Just to clarify in my own mind, the ringing in the coil has big voltage swings up and down but the current is basically canceling itself out.

That current that appears to not be flowing in the coil, would you say that would be showing that no current is leaving the battery to being dissipated in the coil, but just voltage is in the coil?

There is one thing about resistive wire and I realized this with the very first Ainslie experiments I did. Two other people with a lot of experience here described this concept as well and confirmed my belief, which is this and I posted this before.

The resistive wire transforms voltage to current and this current doesn't have to come from the battery. I saw this first with the Ainslie circuit version with the flyback diode. The spike enters the coil and produces heat because the wire manifests current in itself from the voltage and again, this current doesn't come from the battery and this is how the inductive resistor gets more heat than can be accounted for by what is leaving the battery.

What do you think about that concept?

Also, I tried some differential measurements with the Tektronix TDS3054C. The concept of measurement was straight forward and someone walked me through the functions. On this particular test, the readings weren't net negative but were positive in the very small range of about 1/10 of 1 watt, which the person walking me through the measurements believes it was unreliable because that power appears to be way to low to be running the circuit.

I tried to repeat the differential measurements with the Fluke 123 Scopemeter but found out that this particular model is one without the differential ability.

I have about 1.5 weeks left with this Tektronix and perhaps we can talk on Skype or the phone and you can just walk me through what you want me to test - that way you can "be there" while I'm doing it.

[Harvey]

Hi Aaron,

I'll PM you with my Skype info - I only turn it on when I am using it which is rarely. You can share it with Rosemary as well.

From a classical point of view, those large voltage swings (+/- 24V) seen in the ringing contain very little energy. In my imaginary circuit, the values were chosen to specifically exaggerate the power dissipation effects during the first spike cycle. After the body diode ceases conduction, the vast majority of the stored energy in the inductor is already dissipated. The residual energy seen is the sum of those minor ripples times the voltage swing, low power by comparison to the enormous 6kW pulse present in the center of the first spike after HEXFET turn-off.

I personally haven't read any data as of yet regarding power transformation from voltage to current or vice versa within a single wire due to its resistivity - I suppose it could be possible, I just haven't considered it. However due to certain physical geometries, self inductance can lead to power transformations where the process reacts similar to that of a standard transformer. A single wire acts as a one turn coil, and eddy currents in a core each act as a single turn. In order for a transformation to occur, a turns ratio must exist. When the ratio is other than 1:1, a voltage for current transformation occurs.

In the Ainslie test, there are other factors to consider - e.g. emissivity, incandescence, radio-luminescence, radiative cooling etc. Just because we do not see the photons being exchanged between the wire and the enamel does not mean they are not there. Also, we must consider the possibility that they do exist at invisible frequencies. Tungsten for example may emit red photons at low power levels, yellow at higher levels and full spectrum (white) at maximum levels. We tend to confine 'lumen' prefixed words to the visible spectrum, and 'radio' to that spectrum - but really, the process is basically the same across the board. Magnetolumenescence is also a distinct possibility. Of course the classical view is that all of these conditions represent an energy exchange and that the energy must pre-exist prior to the exchange. Static voltage, of itself with no current flow, is an energy source. The power can be as small as the energy stored on your comb after passing it through your hair, to as great as that stored in a cloud ready to produce lightning.

Lets say that none of the classical laws are violated in the Ainslie test. Can we still have a COP > 17? Sure. Just because the 'heating' efficiency is shown to be very high, does not mandate an OU condition. So we will have different camps, 1. the skeptics who simply refuse to accept such a high value and immediate jump to the conclusion that some error in measurement must have been involved; 2. the OU enthusiast who readily accepts that classical laws are flawed and immediately jumps to the conclusion that the test demonstrates a proof to the case; 3. The true scientists who recognizes that the universe is still being mapped and demands hard evidence to substantiate the observations and claims. 4. The experimental physicists who delve into the fringe areas of the unexplained and often witness first hand the bizarre actions that seemingly contradict the classical laws as we currently know them. 5. And then there are those who simply understand that we don't know everything yet, and that hard evidence is not always available at the present but are confident that in time the truth will always become known. And there may even be other camps that I haven't thought of. But regardless of what our beliefs are, the truth never changes. Our perspectives change - and sometimes the 'facts' change with them. But the truth always exists there, waiting to be uncovered. Remove all the possibilities and we are left with the truth.

[Harvey]

Imagine for a moment that we own a Watt Meter. And that we have an Adjustable Voltage and Current Power Supply . Now imagine we take that Ainslie resistor and stick it across that output of that supply, and adjust those values until we get a temperature in the resistor of 50° over ambient and we read the value on the watt meter. That value represents all of the power consumed by the supply and the resistor - entirely. Now we simply insert the Ainslie circuit in between the power supply and the resistor. And without changing the power supply settings we are able to get the same 50° increase over ambient - but the watt meter shows a decrease in overall consumption. Then there can be no doubt that the circuit does show an improvement in COP.

For more precision in the experiment, one could baseline the no load power supply draw and then compare that to the load power draws for both direct and circuited tests. Also, tests could be compared for different thermal settings.

The advantage of this arrangement is that the watt meter is being used in the way it was intended to be used, at the frequency and voltage envelope it was designed for. And its measurement becomes insulated from the aperiodic nature of the Ainslie circuit.