Jed, it is helpful that we now know these two runs were not that same, with one heated inside and the other outside.
We wrote that in the paper.
Jed, it is helpful that we now know these two runs were not that same, with one heated inside and the other outside.
We wrote that in the paper.
We can now perhaps say that these published 2017 excess heat results are unsafe. If the reactor body is much cooler in the control than the active test, aspects of the airflow, possible waste heat transfer direct to the RTD, etc could determine these results.
No, that is complete bullshit. The reactor went for a year or so producing no excess heat, with glow discharge inside, and sometimes with resistance heating inside. It was much hotter than during the calibrations with the external heater, but the heat balance was zero. It was indistinguishable from a calibration. That is how all calorimeters work. I have pointed this out many times, but you keep coming up with this same nonsensical, physically impossible assertion.
I can't tell if you really believe this, or you are just trolling. But anyway, you are wrong. But I am sure you will keep repeating this. It is tiresome. I should ignore your postings, but I hate to see you introduce bullshit into a serious discussion.
Jed, you still have not answered, is the 2017 paper or the 2019 paper correct wrt the reactor the 2017 100% excess power results come from:
2017 (20kg new-style)
2019 (50kg old-style)
I may have misunderstood it but have given quotes on which I base this above.
However the time constant of the active run is 7000s
THHnew is unwilling or unable to show the fit of 7000 s to the actual data.
any engineer, or competent experimental physicist
Any engineer or competent experimental physicist would know that there is a poor fit
of a single time constant to the actual data.
Of the over 4000 data points only ~10% would fit a 7000s time constant curve.
This means that 90% of the data point would constitute noise in THH terms.
If I bought a capacitor with a similar charge/discharge curve
I would seriously suspect it was not a capacitor.
We wrote that in the paper.
About masses of reactors
This was written in the papers too. Papers are more reliable than mobile phones.
Cylindrical R20 2019
"
The reactor vessel and flanges are stainless steel. The vessel is 600 mm long, diameter 114
mm, wall thickness 3.2 mm (1/8 inch). The flanges are 150 mm in diameter. Weight 20 kg "
Cruciform 2017
"
The reactor is made of SUS 316. Its volume is 2740 cm3 and its weight is 20.3 kg. "
Ascoli,
I'm not sure we will get a definitive answer on this, but in the interest of accuracy:[...]
We can now quite easily make progress because we can estimate the thermal capacity and therefore the time constant for given cooling.
Well, it seems we made progress in the meanwhile. JR has finally admitted (1) that, in the active run of May 2016, the heating was internal and not external, as in the control run. Therefore, he confirmed that the cause of the greater time constant – that he called latency – was the deliberate choice of heating the active reactor from the inside, "mainly with glow discharge".
The use of glow discharge suggests another possible trivial explanation - in addition to a possible mistaken value of shunt resistance entered into the data system - for the apparent excess heat: the spiky current on the DC side of the power supply could have been heavily under-measured. This would fit with your concerns about the waveform, wouldn't it?
However, even in this case, the Yokogawa power analyzer, positioned at the AC inlet side of the PSU, would have correctly measured all the absorbed power, but its values have been removed from the spreadsheet of the 120 W active run. This is the main red flag, which JedRothwell has not yet explained.
QuotePaper B says these (old method) came from (old method) calorimetry using 50kg reactor. Paper A 2017 says these same results came from calorimetry which is described as identical to the new method calorimetry in paper B: dual 20kg reactors side by side in the calorimeter.
A paper presented at ICC18 in 2013 (2) describes more extensively both the 20 kg and the 50 kg cruciform reactors. It seems that the reactor shown in Figure 1 of the paper A 2017 is different from the 50 kg cruciform reactor shown in Figure 7 of (2). For example, the reactor in paper A 2017 has 16 holes in the upper main flange, whereas the holes in the larger flange of the 50 kg reactor are more than 20.
In any case, all photos of the experimental setups with cruciform reactors show two identical devices. Therefore, the differences between the active and control curves can't be explained by a difference in the mass of the two tested reactors.
THHnew is unwilling or unable to show the fit of 7000 s to the actual data.
Perhaps I am being presumptious. Perhaps THHnew is willing to show the fit.?
I used a time constant of 8203 seconds
Perhaps THHnew has a better fit with 7000s?
The calibration was performed with the heater wrapped around the outside of the reactor, and the reactor did not get very hot inside. The excess heat run was heated inside, mainly with glow discharge.
Good progress. You have finally confirmed what I've been saying for many weeks and that you have always denied so far (1).
QuoteThat is incorrect, as I said. It is not plausible that several experienced scientists using multiple meters (including ones brought by me and other outsiders) would fail to notice this problem for many years.
What is incorrect? Are you meaning that I was incorrect in attributing the cause to a "mistake"? OK, now you told us it was instead a deliberate choice of the experimenter. But how could I knew it? This fact was not reported in the 2017 paper, which describes the May 2016 tests (2). The author of this paper didn't tell readers that the active run was done by heating the reactor from inside and by means of glow discharge. He mentioned the glow discharge procedure only in section 2.5 "Preparation of reacting material". However, the author pointed out at page 18 that active "tests behave differently from the calibration test", but he didn't explained the true reason, ie that the heating process was completely different. By doing so, he induced the readers understand that the "different behavior" depended only by the activated status of the Ni mesh. Let me say, this is not a correct way of reporting extraordinary results.
Well, by positing unknown endothermic and exothermic processes you can match any temperature curve under the sun!
This is true.
But certainly THHnew must admit that matching the 2017 120W active reactor data with a single time constant of 7000s
is force fitting at best.
LENR theory and experiment suggests that both exothermic and endothermic processes
occur contemporaneously.
We know from chemistry that in any nonspontaneous reaction there is an endothermic phase
which is described as overcoming the activation energy barrier.
LENR theory and experiment suggests that both exothermic and endothermic processes
occur contemporaneously.
How the energy of of the big exothermic deuterium-deuterium fusion is transferred to atoms in the NAE
to appear as low grade heat cannot be revealed by looking at overall calorimetry.
Kasagi has made a start by looking at the transfer to Nickel atoms
"
For a thin Ni foil with Q = 1 MeV, a hot spot is formed only for the case of 500 < Ne <1000; corresponding radius 38 > R > 16 nm and temperature 145 < T < 2150 K. When the excess heat is 1 W, hot spots are repeatedly generated in the sample at a rate of 6.3×1012 / sec. Although very rough, 10-3 to 10-4 of the total number of Ni are included in a hot spot portion. "
Narita and co have also recently results suggesting the exo/endo combo in Ni/Pd
"
Short-period temperature fluctuation was often observed, lasting for 2-4 h at the beginning of the desorption phase for the Pd-Ni sample with fine structure at the interface. It is possible that deuterium diffusion from Pd to the membrane and from the membrane to Pd occurred frequently in the period, and that endothermic and exothermic phenomena associated with the heat of solution repeatedly occurred owing to deuterium transport between the two metals. "
Takahashi et al report endo/exo here
"
Data of 50 to 140 W/kg level excess thermal power was repeatedly obtained by CNZ-type samples with H-gas at elevated temperatures after the saturation of H-gas absorption (endothermic) by sample. Excess thermal power of ca. 50 W continued for more than two weeks by 505 g CNZ7r (re-calcined) sample, with very strange behavior of the "cooled-flat and oscillating" TC4 RC upper flange temperatures "
Has there ever been a spreadsheet posted of the raw data for R20? There are a lot of disagreements that seem to be partly occurring due to a lack of data. If the spreadsheet needs cleaned up a bit, then it seems like that would take a lot less time than all the posts defending the final calculations. I don't see much resolution happening until raw data is posted and/or more data from replication attempts come in.
? There are a lot of disagreements that seem to be partly occurring due to a lack of data
Not because of R20...
a lot of the recent discussion appears to be in 2017
did Jack look at the 2017 120W active reactor sheets..?
they were posted.
THHnew is unwilling or unable to show the fit of 7000 s to the actual data.
Perhaps you'd like to substantiate that? 7000s is a pretty good approximation to the time constant? perhaps a bit low, but I was only eyeballing. What would you estimate it as?
But certainly THHnew must admit that matching the 2017 120W active reactor data with a single time constant of 7000s
is force fitting at best.
Well, on the one hand, since it is clear that a more accurate model of the thermal behaviour here would have more than just one time constant, I can predict that a single time constant cannot exactly fit the data.
On the other hand the reactor is so very massive that its thermal inertia dominates the dynamics. In addition, the system is nonlinear due to the fact that the calorimeter is nonlinear (see data in paper). These two effects, higher order poles + higher order amplitude response, have some effect on the exact curve. Which matters not at all if what you want to do is estimate the magnitude of the dominant pole.
If every scientist whose results did not exactly agree with some (known) simplification of the the system is viewed as "force fitting" then no experiment could ever be trusted, and no engineer could ever design anything.
The art of modelling the real world is to accept that nothing is perfect, but good approximations where the errors are understandable (though too complex to model) can make predictions.
I think you are confused about data modelling. Trying to model the data too accurately (as you did with your "random noise due to NAE") is usually a big mistake, leading to the serious sin of overfitting.
Any engineer or competent experimental physicist would know that there is a poor fit
of a single time constant to the actual data.
Of the over 4000 data points only ~10% would fit a 7000s time constant curve.
This means that 90% of the data point would constitute noise in THH terms.
If I bought a capacitor with a similar charge/discharge curve
I would seriously suspect it was not a capacitor.
RB - this indicates that perhaps you have not looked carefully at the way that physicists (and engineers) work out how to model data.
It is OT for this thread, but I'd recommend (googled randomly):
https://www.sciencedirect.com/…hematics/hypothesis-space
and for a much more insightful treatment
https://bayes.wustl.edu/etj/prob/book.pdf
or, for a more practical guide, just google overfitting underfitting hypotheses.
You are claiming that I'm underfitting the data, by using a single time constant, when that is probably not true, but even if it were true the higher order corrections that could possibly be predicted here don't alter the needed results, which come from the low frequency information.
Has there ever been a spreadsheet posted of the raw data for R20? There are a lot of disagreements that seem to be partly occurring due to a lack of data. If the spreadsheet needs cleaned up a bit, then it seems like that would take a lot less time than all the posts defending the final calculations. I don't see much resolution happening until raw data is posted and/or more data from replication attempts come in.
One of the interesting things (that we should have realised earlier) is that the dynamics of the output temperature curve, combined with the reactor case temperature, give us genuine insight into the combined power required to generate those dynamics. That is a truly independent cross-check of all the output-side power estimation.
It still does not preclude input power mismeasurement, but it would short-circuit most of the R19 arguments.
Ascoli's useful observations on the 2016 data (that is the 100% excess power 2017 published results that preceded the R19 and R20 results):
On investigating the dynamics issue we find that the active data has an internal heater, The control data has an external heater. This explains the dynamics if we assume that the external heater does not make good thermal contact with the reactor and therefore heats up the reactor casing much less than the internal heater. It also explains the different resistance. There remains no explanation for the different input power measurement between active and control runs, which is unfortunate, although not in itself a problem. The exact methodology of these results (old-style or new-style) has still not been clarified; the two papers contradict each other.
There is nothing here that necessarily invalidates these results. However, the poor methodology (very different systems used for active and control runs) and poor documentation - e.g. the difference in power measurement is not made explicit on the spreadsheets - is unfortunate and makes it more difficult to accept extraordinary results as real rather than some mistake caused by poor methodology and record-keeping.
As one example. It would normally be obvious from reactor temperature whether control power out was more or less than active power out (at least at the 2x level reported). In this case that cannot help, because the active reactor, with external heater, would get much hotter even with the same power as the control and no mesh inside.
It should also be possible to do something similar (compare dynamics) with the R19 results given detailed spreadsheet data.
On the positive side: we can see that looking at the dynamics of these runs, together with the reactor case temperature, allows an independent measurement of the total (output) power, as an excellent cross-check on the output calorimetry that should be accurate +/- 20%.
Display MoreRedux
Ascoli's useful observations on the 2016 data (that is the 100% excess power 2017 published results that preceded the R19 and R20 results):
- The input power is measured differently between the control and active run spreadsheets. In the control case with a mains power analyser. In the active case with figures computed from V*I. This is not evidenced on the spreadsheets where the power column is used and filled with V*I or other measurements which by inspection would come from a power analyser.
- The resistance of the heating element for control and active runs differs by a factor of 2.
- The dynamics of the control data are 10X faster than the active data
- The mass of the reactor used for these results is inconsistenctly stated as 50kg old-style (from the 2019 paper) and 20kg - seemingly new-style though not called that (from the 2017 paper!). It would be good to have some confirmation of which reactor was used, and also which methodology. the old-style reactors were measured individually by the calorimeter. the new-style reactors used two reactors, control and active, together at the same time.
On investigating the dynamics issue we find that the active data has an internal heater, The control data has an external heater. This explains the dynamics if we assume that the external heater does not make good thermal contact with the reactor and therefore heats up the reactor casing much less than the internal heater. It also explains the different resistance. There remains no explanation for the different input power measurement between active and control runs, which is unfortunate, although not in itself a problem. The exact methodology of these results (old-style or new-style) has still not been clarified; the two papers contradict each other.
There is nothing here that necessarily invalidates these results. However, the poor methodology (very different systems used for active and control runs) and poor documentation - e.g. the difference in power measurement is not made explicit on the spreadsheets - is unfortunate and makes it more difficult to accept extraordinary results as real rather than some mistake caused by poor methodology and record-keeping.
As one example. It would normally be obvious from reactor temperature whether control power out was more or less than active power out (at least at the 2x level reported). In this case that cannot help, because the active reactor, with external heater, would get much hotter even with the same power as the control and no mesh inside. It should be possible to do this with the R19 results given detailed spreadsheet data.
On the positive side: we can see that looking at the dynamics of these runs, together with the reactor case temperature, allows an independent measurement of the total (output) power, as an excellent cross-check on the output calorimetry that should be accurate +/- 20%.
Thank you for the clear and concise summary of the issues that have emerged so far from a more in-depth survey on the 120W spreadsheets and on the other documents reporting the May 2016 tests. It shows that these tests are so far the most significant of the whole Mizuno's activity, even more important than the recent R19 and R20 tests, because the numerical data allow us to estimate the first and most important requisite for scientific research, that is the reliability of the information which are provided to support the claimed results. What emerged by looking at these data is that this reliability is very low, especially when compared to the extraordinary nature of the claimed results.
In your summary, you said "The input power is measured differently between the control and active run spreadsheets". Well, as I already told you (1), this is not exactly my position. We can only say that the input power is "reported" differently, but the circumstances and the information from JR strongly indicates that the power was measured in the same way and by the same instruments in all cases. So IMO the real big problem is that the data coming from the Yokogawa power analyzer have been removed from the spreadsheet of the 120 W active run, and this modification can't be done unintentionally.
You, me, and others have asked many times to provide an explanation for the differences between the data in the "Input power" columns of the two available spreadsheets. So far, JedRothwell has always ignored this crucial question and I hope he understand that the lack of any alternative plausible explanation raises doubts that go beyond the mere unreliability of the released information.
THHnew is unwilling or unable to show the fit of 7000 s to the actual data.
I guess THHnew is unwilling to show the fit of 7000s to the actual data.
But he asserts it is 'good' fit
based on what measure?
" Lack-of-fit sum of squares ??· Reduced chi-squared statistic"???
The fact is that 7000 s ,8203 and any single time constant for
the whole reactor are poor fits.
In particular THHnews single time constant for the reactor does not explain
1. the difference between the shape of the heating curve and the shape of the cooling curve
2. the differential scatter in the 4000 temperature reading,
The heating and cooling profiles of the reactor are not due to any simple assumption
of one thermal mass and first order heat transfer.
Not because of R20...
a lot of the recent discussion appears to be in 2017did Jack look at the 2017 120W active reactor sheets..?
they were posted.
I did. It was not helpful. I'm not really interested in that. The 2017 data posted was not complete. Maybe I'm thinking of the wrong data though. I'm referring to a link that you posted some time ago.
In particular THHnews single time constant for the reactor does not explain
1. the difference between the shape of the heating curve and the shape of the cooling curve
2. the differential scatter in the 4000 temperature reading,
The heating and cooling profiles of the reactor are not due to any simple assumption
of one thermal mass and first order heat transfer.
I'm sure others are as bored as I am by repetition, but let me say again:
The noise has various possibilities (e.g. ascoli turbulence) and is irrelevant to understanding the system dynamics, since it is noise
NOISE NOISE NOISE
The noise THHNew propose in his vagueness is indeed noise.
Electrical noise?? Audible noise?
Ascoli turbulence? is this a scientific term.???
is there a paper on Ascoli turbulence?
Is it related to Ascoli foam?
Perhaps by noise THHnew simply
means deviations from THHnews simplistic
mathematical assumption of a single time constant.
Such noise amounts to more than 80% of the readings
Unless defined and accounted for
noise is just typical THHnew noise and blather