At least, when critiques can quite easily be answered from different protocols, a few extra measurements, whatever, it is surely worth doing that?
It’s clear we will never agree.
I don’t consider critiques based in the view that the phenomena is impossible to be valid. You insist in saying “I don’t think is impossible” but you act as if the corpus of evidence is erroneous because it doesn’t align with your worldview.
The evidence is enough for anyone reasonable.
Display MoreEd,
One clear needed addition to your experimental results: you need to specify the area of the Pd cathode. (Maybe I did not look in the right place - but i could not find it).
You and Staker both give excess heat as W/cm^3. This is helpful when considering excess in terms of what is possible chemical energy. But when comparing different electrode treatments etc (assuming the experiment results are believed) between different experiments it is actually the excess in W/cm^2 that is relevant. So both figures are important. (Straker has - for his long-term test - 0.3W over 4cm^2).
The set of metrics:
1. excess power /cathode area
2. excess power / cell heating input power
3. excess power
4. excess power / cell electrolysis splitting power (this with 3. determines electrolysis current)
5. excess power/ cathode volume
6. excess power / cell accuracy (ok - some disagreement over what is needed to measure that)
Are all a useful comparison of different experiments.
Thanks, THH
THH, the production of power is a volume effect because it results from special sites located throughout the volume. The surface area plays no role. This issue was confused by Fleischmann, which has caused problems ever since.
Each piece of Pd is different because each piece contains a different number of active sites. Here is a comparison of 157 samples. The behavior indicates that the number of sites is random with the probability of adding more sites decreasing as the number is increased.
Science has accepted ways to determine the accuracy of measurements that you seem to ignore. All measurements contain two kinds of error, one is the result of random effects, and the other results from the use of incorrect values in the calibration equation. I have addressed both kinds of errors. In addition, normally people test their measurements against known behavior. I have done this. In other words, I have applied all of the methods accepted and used by science to my measurements. Your criteria are not part of how measurements are normally evaluated, yet skeptics continue to use your approach. Clearly, a double standard is applied to LENR, which you are helping.
THH, the production of power is a volume effect because it results from special sites located throughout the volume. The surface area plays no role. This issue was confused by Fleischmann, which has caused problems ever since.
Is that accepted? It is a question how to what extent surface treatment extends throughout the volume. I guess if the Pd is thin enough what you say will be true
Science has accepted ways to determine the accuracy of measurements that you seem to ignore. All measurements contain two kinds of error, one is the result of random effects, and the other results from the use of incorrect values in the calibration equation. I have addressed both kinds of errors.
Well, just to split hairs:
if the calibration constant depends, even by 0.3%, on the heat distribution inside the box, then that constant could be incorrect. (in fact not a constant).
In fact all calorimeters have not-quite-constant calibration constants - it is just a matter of how that is bounded and whether given the experiment can result in variable heat distribution from the calibration.
So surely that bound should be part of the error analysis? It is all I am asking.
THH
Each piece of Pd is different because each piece contains a different number of active sites. Here is a comparison of 157 samples. The behavior indicates that the number of sites is random with the probability of adding more sites decreasing as the number is increased.
Thanks Ed.
Right. That is interesting. It also motivates a lot of effort to do two things:
(1) optimise number of sites (or whatever this is)
(2) ask whether excess heat / volume (for thin material) has got higher in newer experiment than it was in old experiments. Given progress in understsnding this, it should have done that?
THH, as I described in the paper you were tasked to read and critique, active sites can be created many different ways. Each of these treatments has a different collection of variables that determine the outcome. Gaining control over these variables is required but is seldom studied by anyone. I have described one method to control the process in a recent paper. This skill is gradually being mastered. However, it will never be mastered as long as people have the wrong understanding of the active environment. I had hoped this environment could be discussed on the Forum, but sadly, that is not the subject of interest.
Instead of these questions being discussed, we keep addressing issues that have no relationship to how LENR can be made to occur at higher rates. Why??
When you say that a calorimeter does not have a constant calibration constant, you are saying that the constant has a random component. This random error can be determined, which I have done. Unfortunately, you are handicapped by not having a good background in error analysis, which causes this discussion to go off the rails.
THH, as I described in the paper you were tasked to read and critique, active sites can be created many different ways. Each of these treatments has a different collection of variables that determine the outcome. Gaining control over these variables is required but is seldom studied by anyone. I have described one method to control the process in a recent paper. This skill is gradually being mastered. However, it will never be mastered as long as people have the wrong understanding of the active environment. I had hoped this environment could be discussed on the Forum, but sadly, that is not the subject of interest.
Instead of these questions being discussed, we keep addressing issues that have no relationship to how LENR can be made to occur at higher rates. Why??
When you say that a calorimeter does not have a constant calibration constant, you are saying that the constant has a random component. This random error can be determined, which I have done. Unfortunately, you are handicapped by not having a good background in error analysis, which causes this discussion to go off the rails.
Ed, forgive me, but I am very familiar with error analysis (and have a good maths and stats background which helps with analysis generally) and the basics are quite transferable.
Not constant is not the same as "has a random component". For example, if the calibration constant varies by 1% depending on which side of the box heat is emitted on (for example, the fan is not 100% effective in equalising temperatures) that is not properly a random change. It is a systematic change depending on cell conditions.
You may choose to call this a random component - but detecting for it is different. For an independent random component you can just repeat calibration. For variation in constants with cell conditions you need to check all the relevant cell conditions - or at least extreme points within the parameter space of possible cell conditions.
What you say is only true under the assumption that there are no such changes, so that all change is independent random.
That assumption may be (effectively) true for your calorimeter. But it is not necessarily so.
THH
Gaining control over these variables is required but is seldom studied by anyone. I have described one method to control the process in a recent paper. This skill is gradually being mastered. However, it will never be mastered as long as people have the wrong understanding of the active environment. I had hoped this environment could be discussed on the Forum, but sadly, that is not the subject of interest.
Well it interests me. And it is a very practical matter. If you can control this (assuming your hypothesis is correct which I'm happy to do for purposes of investigation) then you can increase the excess heat generation in these experiments anyway - and increase that enough all these skeptical caveats become moot.
you are handicapped by not having a good background in error analysis,
but makes up for it with expertise in wordspiel in spades
as with the mcroanlysis of Mizuno' calorimetry speadsheet thread
another longwinded thread
where THH dismissed in your face statistics as 'noise'
Display MoreNot constant is not the same as "has a random component". For example, if the calibration constant varies by 1% depending on which side of the box heat is emitted on (for example, the fan is not 100% effective in equalising temperatures) that is not properly a random change. It is a systematic change depending on cell conditions.
You may choose to call this a random component - but detecting for it is different. For an independent random component you can just repeat calibration. For variation in constants with cell conditions you need to check all the relevant cell conditions - or at least extreme points within the parameter space of possible cell conditions.
What you say is only true under the assumption that there are no such changes, so that all change is independent random.
That assumption may be (effectively) true for your calorimeter. But it is not necessarily so.
THH
In the case of LENR, a person might measure .5 watts but the actual value would be 0.4 watts. You always assume the value is always zero regardless of how much power is measured.
If the calorimeter calibration constant changes because the cell is not in the same place or because some condition changes, this is not random but results in a change in the calibration constants. In other words, the wrong information is being used. On the other hand, the random effects are superimposed on this kind of change. A good calorimeter does not change its calibration constant. If such a change were to occur, this would be obvious. You keep assuming that I or other people would not know that such a change had taken place. In any case, I have demonstrated that my calorimeter does not have this problem.
Yes, all measurements contain errors. But this error does not manifest as a reproducible behavior. In fact, we now have hundreds of such measurements showing nonzero power. You assume each of them is wrong by an amount that would give zero power if the truth were known. When are you going to stop beating the dead horse and work to understand this amazing discovery?
If such a change were to occur, this would be obvious. You keep assuming that I or other people would not know that such a change had taken place. In any case, I have demonstrated that my calorimeter does not have this problem.
Let us agree to disagree about the demonstrated part of that. There is not however much difference between us.
In a published paper - those statements would - for best credibility - need to be substantiated in the case at least that I have mentioned here. Below is how you do this, and how I think you would need to do this to strengthen your results.
Consider:
That is what any objective reviewer of your results would say - and it is not difficult to do. You have resistor in electrolytic cell. You need also resistor on top (or side at top) of recombinator. Those two give decent worst case distributions for this experiment, since the two main heat sources are the electrolytic cell and the recombinator.
Also - it is easy to do this. So any non-LENR community reviewer, if you did not do this, would be suspicious why it has not been done. Everyone wants to have stronger results? You have a good experiment. Why not strengthen the results?
THH
THH, I did what you required. Here is the figure showing the result, which is in the paper I cited. I compared the heat produced by the resistor in the cell to heat produced by an incandescent light bulb that took the place of the electrolytic cell. This would be the worst possible condition. The bulb was heated by different sources of DC current. Each time, the results showed excellent agreement.
Also - it is easy to do this. So any non-LENR community reviewer, if you did not do this, would be suspicious why it has not been done. Everyone wants to have stronger results? You have a good experiment. Why not strengthen the results?
Because the results are strong enough for anyone not blinded by belief of impossibility!?
for anyone not blinded by belief of impossibilit
perhaps these guys can give an unblinded opinion..eh
THH, I did what you required. Here is the figure showing the result, which is in the paper I cited. I compared the heat produced by the resistor in the cell to heat produced by an incandescent light bulb that took the place of the electrolytic cell. This would be the worst possible condition.
Dear Ed,
Could you possibly explain to me why that would be worst case? The bulb would indeed be different from the cell, but because its heat (radiant + convective) would come from the same place it would not test imbalance between the panels on the recombinator side of the box and the electrolytic cell side. (Although I guess how much it did this would depend on exactly where it was placed).
However I'd expect it to to have a slightly different heat distribution from the cell. So documenting that difference would be helpful. Thank you for the graph. It would be very helpful to work out the best fit coefficients for both sets and put them in the paper. The closer they are - the more convincing the match. I don't think I can eyeball a 50mW difference easily on that graph...
Because the results are strong enough for anyone not blinded by belief of impossibility!?
Curbina - I will not reply to your repeated and incorrect implications about my beliefs further - just checking that you have read and understood my previous answers?
The Bayesian bit is maybe a bit technical and contentious (here).
But the other part of my reply is common sense.
Everyone in the LENR community - convinced LENR exists - will not see the need now to prove it as something extraordinary (which - it is - exist or no).
However mainstream science does not have that conviction, and therefore requires very strong evidence to be moved beyond a lukewarm "well, there is something there, maybe, but it does not look nuclear and the results are all over the place and not clear".
That will not encourage the amount of young v clever talent everyone would like.
So it really does not matter whether a more skeptical view is "right" or "wrong". The sensible thing for those in the LENR community to do is to strengthen their evidence, and write it up carefully, without gaps, in a way that makes it as clear as possible.
If the results are real it is always possible to do experiments and write-ups which are even stronger than previous results.
Since this has not been done so much in the past, and can always be done more, heeding the more skeptical voices willing to look at your papers might be helpful.
THH
PS - for example, compare McKubre with Storms and Staker. I'd expect it to be easy now to find a cathode now as good as McKubre used (this being the main variability) and instrument the heck out of it. Ed's calorimeter looks pretty good to me. And it is cheap and easily replicable.
However mainstream science does not have that conviction,
Do you have mainstream science's phone number..
I'd like to run this by him or her..
I have not seen a write-up of the solid-gas reactions,
Do you have mainstream science's phone number..
I'd like to run this by him or her..
Perhaps you can use me as a surrogate for the elusive entity. It would be helpful if you could link the written paper - which I see is out now? Not enough info in that screenshot for google. But I'd like to look at it.
Why is it those in the LENR community seem to think that graphs extracted from written papers are more convincing than the papers?
That graph shows a calculated excess power - right?
The only comment I can make at the moment is that if that excess heat comes from nuclear reactions it is most oddly coincidental that the heat should scale linearly with temperature above ~ room temperature and presumably - since no-one has observed endothermic LENR - be constant 0 for less than ~ room temperature.
This is the effect that I pointed out to you a while back which I believe at that time you misunderstood. That graph of excess LENR heat forms two linear segments with a sudden change in gradient around room temperature.
That is unprecedented for any reaction - nuclear or other. However it does make sense if the linear excess heat comes from some mistake or artifact in the calorimetry.
