Rosemary Ainslie Quantum Magazine Circuit COP > 17 Claims

[MarkE]

I have data from three very important trial runs.

These runs were designed to compare the performance of the system with 1) the full waveform showing Q1 fully on and full amplitude Q2 oscillations during their respective period portions; 2) the exact same settings on everything except with the Snubber installed to eliminate the Q2 oscillations without affecting either the Q1 current or the computed mean Q2 current (as shown in the latest video); and 3) the "Figure 3" configuration as in the Sneak Preview, with the FG offset turned down so that there is "complete restriction of positive current flow" i.e. no Q1 current indicated.... except in my version, there really IS no current shown and no current flowing because my measurement is correct, not fabricated as Ainslie's was.

Each run used 4 x 12V, 5 A-H batteries, and a frequency of 1 kHz and 10 percent HI duty cycle +/- square wave, with no offset except in the last "Fig3" oscs only run. This frequency and duty cycle are similar to what the Ainslie mob chose to use in their recent demonstrations.  Each run lasted an hour and took the load to thermal equilibrium and the sample interval was one minute, as usual, and there is a video record of all raw data, recorded and logged by the process I have detailed above. The raw data videos are available for inspection by anyone at any time and my definition of "thermal equilibrium" is well defined and known and can be seen in the raw data. In some cases the cell comes to equilibrium in 40 minutes or less but I still run for the full hour. In all cases the cell temperature changes by less than 0.05 degrees C/minute at the end of the trials.

Since the DC calibration power-temperature relationship is nicely linear, as are the power-temperature relationships for equal parameters of the DUT, an "efficiency" or COP can be determined simply by the ratio of the final temperature of the experimental trials, with the equivalent temperature of the calibration runs at the same power level. Since it is the temperature _over ambient_ that is used, it is not necessary to convert all temperatures to degrees Kelvin in order to calculate the ratios, I think.

The "bottom line" results are as follows:

Trial 1, full waveform with both Q1 current and Q2 oscs: 28.7 W input shown on DMMs, 39.7 degrees C rise above ambient.
The DC calibration temperature for 28.7 W DC input is about 51 degrees C over ambient. Efficiency COP = 39.7/51 = 0.78 or 78 %.

Trial 2, Snubber in place, no oscs only Q1 current + real Q2 current: 25.6 W input shown on DMMs, 33.1 degrees C rise above ambient.
The DC calibration temperature for 25.6 W DC input is about 47 degrees C over ambient. Efficiency COP = 33.1/47 = 0.70 or 70 %.

Trial 3, FG offset enough negative to make "Figure 3", Q2 oscs only and no Q1 current: 10.8 W shown on DMMs, 9.6 degrees C rise above ambient.
The DC calibration temperature for 10.8 W DC input is about 23 degrees C over ambient. Efficiency COP = 9.6/23 = 0.42 or 42 %

I will also be presenting these data in graphical plots a bit later on. Note that the measurements of the input power to the DUT system are actually likely to be somewhat _low_  for various reasons I will discuss later. If it really took more power than indicated, this of course moves the COP numbers even lower.

Some conclusions are immediately obvious. Not only were efficiencies over 1 not encountered at all or even hinted at.... the Q2 oscillations especially are demonstrated to be very wasteful: more than half of the applied power is wasted, spent heating the mosfets and the internal load in the FG !! It never makes it to the "element resistor" at all.

Other conclusions are also clear. I'll leave the further discussion for later, I really need a cup of coffee right now.

That's interesting data. 

[TinselKoala]

Wow!  She really wrote that and put it on her web site.  If Steve sees it he will not be happy.

Steve has seen it, and he has received the email(s) she sent, and you are right... he is not well pleased.

That's interesting data. 

Yes, it is interesting. The trends I identified earlier at lower power and different frequency/duty cycle combos still hold.
I have also done a run with just the FG connected directly to the load, with no intervening circuitry or other power supply _at all_. The temperature of the load cell actually rises by a few degrees C. Not quite as much as the "Figure 3" oscs only runs, but still a measurable and significant amount, which at least indicates that the FG IS INDEED capable of contributing power, meaning heating, to the load, even when used with no other power supply.

Now, what _else_ should I do in an attempt to reproduce Ainslie's claimed OU of >17 or INFINITY? Is there some preferred frequency and duty cycle, some preferred waveform that I should be trying to use? I am happy to consider suggestions as to further parameter sets to explore. It takes about half an hour to set up, an hour to perform the experiment and another 90 minutes or so to cool down for another run. During the cooldown I collate and enter the data into the spreadsheet for evaluation. So I can reasonably do two or three data runs in a day, without too much strain or boredom.

My only guidelines so far have been what Ainslie has herself "published" or demonstrated, and my experiments have shown rather mundane and ordinary and indeed expected results, results predicted by conventional understanding of electronics. There is no hint of any extra power sources in my experiment. All load heating is fully accounted for by the power supplied by the conventional source, battery or power supply; if the FG is supplying significant power (it is, barely) this would actually_reduce_ the COP values I have calculated using only the main power supply values.

Since the Great Scientist's latest pronouncements.... is the "Ainslie effect" now reduced to being merely a claim about battery charge lasting a bit longer when discharging on a pulsing schedule? Well... this is a known phenomenon and does not require the participation of zipons or a new "theory" to replace Faraday, Maxwell, and QED. Will the battery, under _ANY_ discharge schedule, give out more energy than was originally used to charge it? What do you think?  I'm not going to explore this issue .... unless and until Ainslie herself actually reports some valid data on the subject. Which could be another four years, until she finds some other poor innocent unsuspecting victims to twiddle her knobs and put up with her twaddling bullying obnoxious ignorant rantings.

Or is it that she still believes her "negative power product" calculations based on spurious data inputs are somehow valid? Yet the batteries discharge normally, yet anyone can reproduce the "negative power product" values by reproducing her bad technique, yet it has been explained and demonstrated just where the spurious values come from. If someone can claim OU based on fabricated and error-ridden data, and ignore all evidence to the contrary including good measurements at the same operating parameters.... that is not Science, it is religion or some other delusional system.

[MarkE]

Steve has seen it, and he has received the email(s) she sent, and you are right... he is not well pleased.



Yes, it is interesting. The trends I identified earlier at lower power and different frequency/duty cycle combos still hold.

Now, what should I do in an attempt to reproduce Ainslie's claimed OU of >17 or INFINITY? Is there some preferred frequency and duty cycle, some preferred waveform that I should be trying to use? I am happy to consider suggestions as to further parameter sets to explore. It takes about half an hour to set up, an hour to perform the experiment and another 90 minutes or so to cool down for another run. During the cooldown I collate and enter the data into the spreadsheet for evaluation. So I can reasonably do two or three data runs in a day, without too much strain or boredom.

My only guidelines so far have been what Ainslie has herself "published" or demonstrated, and my experiments have shown rather mundane and ordinary and indeed expected results, results predicted by conventional understanding of electronics. There is no hint of any extra power sources in my experiment. All load heating is fully accounted for by the power supplied by the conventional source, battery or power supply; if the FG is supplying significant power (it is, barely) this would actually_reduce_ the COP values I have calculated using only the main power supply values.

Since the Great Scientist's latest pronouncements.... is the "Ainslie effect" now reduced to being merely a claim about battery charge lasting a bit longer when discharging on a pulsing schedule? Well... this is a known phenomenon and does not require the participation of zipons or a new "theory" to replace Faraday, Maxwell, and QED. Will the battery, under _ANY_ discharge schedule, give out more energy than was originally used to charge it? What do you think?  I'm not going to explore this issue .... unless and until Ainslie herself actually reports some valid data on the subject. Which could be another four years, until she finds some other poor innocent unsuspecting victims to twiddle her knobs and put up with her twaddling bullying obnoxious ignorant rantings.

The rant Ms. Ainslie posted raises the bar on insane even for her.  In one paragraph she admitted and then demonized her own recorded observations.  She made Steve the bad guy for reminding her of those observations in private.  One sentence later she went all Tasmanian devil on some "negative wattage" diatribe.  Just for the insane woman's benefit, let's review again her own written statement on the Aug. 11, 2013 matter:

Reference measurements taken at new sense points directly at the battery bank indicated average net positive battery drain of 14W to 15W.  Maximum heater temperature rise during these experiments was 21C.From our electrical DC power to temperature rise tests conducted in 2011 and appear as Table II in this paper, a 21C heater temperature rise corresponds to an equivalent power of between 2.4W and 3.4W.  We therefore obtained heat output that was only a fraction of the input power.

As we are unable to replicate our earlier reported results, we respectfully withdraw this paper in both of its parts.

Details of the test protocols are available as August 11 Demonstration Outline_draft_05.pdf.  Test Phases 1 - 3 were conducted during the live demonstration.  We ended the demonstration after Test Phase 3 when it became clear that the net battery power drain was far in excess of the possible heater output power.

Now, she's off to the races with the gem you copied, declaring that what she wrote is: duplicitous, a blatant lie, immoral ... .  The woman has gone cuckoo for Cocoa Puffs.  Add Steve to the long list of respected professionals that she's declared are part of some evil cabal.  This is going to cost me some beer.

You've already done the 3.7% at 2.4kHz tests before.  I suppose that you could repeat them now so that is all done on this gear.  Can the F43 generate such timing? 

Other tests that you could do might include characterization of the current regulation behavior during Q2 phases with and without the snubber.  For instance you could set the function generator low swing to:  -4V -8V and -12V and show what that does to the current and heater.  You might sneak in a drain voltage capture in the process.  There is other fun that you can have with the test fixture, like comparing the uncompensated and compensated current sense waveforms side by side.

[TinselKoala]

The workhorse F43 Fg that I have been using lately can only go as short as about 5 or 6 percent HI duty cycle. However I have the DataPulse DP101 fast-risetime (last century "fast") pulse generator which can do much shorter duty cycles without difficulty. It has separate positive- and negative-going outputs though which may make it unsuitable for the "Q-array" circuit, but since the 3.7 percent at 2.4 kHz applies to the original single mosfet Quantum Magazine claims it should be no problem to use it for that. The output is also limited to 10 volts max in either direction and the "offset" is controlled just by setting the two output amplitudes differentially. I usually use the negative output as the monitor and the positive output to drive whatever load I'm running. I will have to see if it will work properly. How ever (tm RA)... she has always claimed that there were "oscillations" in her apparatus at that time too. The only such behaviour I have ever seen in the one-mosfet circuit was caused by severely underpowering the 555 timer circuit if used, and/or burned spots on the "gate" potentiometer. It is easy to burn this pot and an ordinary half-watt carbon pot will not last very long, so experimenters should use a "type J" or better pot for the "Gate" pot in the Q-17 single mosfet circuit. The "oscillations" claimed by the folks at Energetic Forum always looked like bad scoposcopy use to me and I even showed the same fake "random aperiodic Hartley resonance" as she called them, things on the Fluke 199 scopemeter I used (among other instruments) back in 2009/2010.

[MarkE]

The workhorse F43 Fg that I have been using lately can only go as short as about 5 or 6 percent HI duty cycle. However I have the DataPulse DP101 fast-risetime (last century "fast") pulse generator which can do much shorter duty cycles without difficulty. It has separate positive- and negative-going outputs though which may make it unsuitable for the "Q-array" circuit, but since the 3.7 percent at 2.4 kHz applies to the original single mosfet Quantum Magazine claims it should be no problem to use it for that. The output is also limited to 10 volts max in either direction and the "offset" is controlled just by setting the two output amplitudes differentially. I usually use the negative output as the monitor and the positive output to drive whatever load I'm running. I will have to see if it will work properly. How ever (tm RA)... she has always claimed that there were "oscillations" in her apparatus at that time too. The only such behaviour I have ever seen in the one-mosfet circuit was caused by severely underpowering the 555 timer circuit if used, and/or burned spots on the "gate" potentiometer. It is easy to burn this pot and an ordinary half-watt carbon pot will not last very long, so experimenters should use a "type J" or better pot for the "Gate" pot in the Q-17 single mosfet circuit. The "oscillations" claimed by the folks at Energetic Forum always looked like bad scoposcopy use to me and I even showed the same fake "random aperiodic Hartley resonance" as she called them, things on the Fluke 199 scopemeter I used (among other instruments) back in 2009/2010.

The pulse generator sounds like it will do the job.  Or you can hook up a 555 to generate the timing.  My personal preference in that regard is to use two 555s, but it can be done with one.