FP's experiments discussion

  • Thus we need one paper describing one self-contained result with calibrations, controls as needed, with all checks complete, and the detailed analysis and description of each check perhaps referenced, but with a clear statement that it was done.

    Why do you need one paper? Several papers cover the details you demand. Why do you insist they be combined into one?

  • It seems that people don’t understand my issues with the ‘7% error is just experimental noise’ position that Szpak, Mosier-Boss, Miles, and Fleischmann (SMMF) (and McKubre by supporting it) take in their measurement of water exiting an F&P cell in their 2004 paper. I have two issues with it, one specifically scientific, the other ‘philosophical’.


    Scientifically, a 7% excess over that expected from 100% Faradaic efficiency indicates some other mechanism for water (possibly electrolyte) loss. Yet SMMF simply wave their hands and call it experimental error. But if there was recombination occurring in their open cell (which does happen) that number would actually be low. SMMF react and condense the electrolyzed gases that exit the cell. 100% Faradaic efficiency assumes 0% recombination. They cite 7.2 cc of water as 100%. So if they had 20% recombination ongoing that means that they would have expected 80% of 7.2 = 5.76 cc of water when they measured 7.7. That means their error would now be 35.8%, not 7%. This issue needs to be clarified by identifying exactly where the excess volume is coming from. Otherwise the potential error becomes huge.


    Philosophically, the hand waving excusing of the error is unacceptable under modern statistical practices. “It’s just experimental error…” is an unacceptable response to questions regarding it. For the record, modern statistical quality control (i.e. analytical chemistry) routinely achieves biases of less than 1% and relative standard deviations also less than 1%. A 7% error being ‘acceptable’ is very ‘old school’. The old school attitude when applied to the magnitude of the CCS error I pointed out in 2002 automatically (i.e., via knee-jerk) suggests the CCS error is ‘acceptable’ or more routinely ‘below the noise level’. But the ‘new school’ attitude says ‘check it out’, which I did in my last post (and in my 2002 publication), and where I found errors on the order of the signal, which is certainly not ‘unimportant’. And the problem that arises from hand-waving away the CCS is that the researchers end up ‘working in the noise’, the prime characteristic of pathological science.

  • But then they write:

    "Secondly, as has been noted in the main text, the current efficiency for D2 and O2 generation is close to unity, i.e. γ = 1." (which means they are claiming no recombination) and the gamma factor disappears from then on including in all subsequent publications AFAIK (including the ones we have been discussing). This assumption seems to be hard-wired into CFers.


    For what I have seen, there are many of such hard-wired assumptions about CF. It's impossible to deal with all of them. I prefer to give priority to the most influential, that IMO are a) the assumption that all the water left the F&P cells as dry steam, and b) the occurrence of HAD events in these cells. Both these anecdotes come from the paper that F&P presented at ICCF3, and which is still considered the Fleischmann "major paper".


    I think this discussion on the F&P experiments will be very productive if it will shed some light on these two fundamental aspects of the CF history.


    With this regard, McKubre made an important contribution affirming that "F&P were exquisitely aware of the calorimetric consequences of delivering unvaporized electrolyte out of the calorimeter". Of course this awareness conflicts with the fact that FWIK none of their public reports deals with this crucial problem and thoroughly illustrates the methods they used to ascertain the dryness of the vented gas stream.


    Quote

    In "Thermal behavior of polarized Pd/D electrodes prepared by co-deposition"; S. Szpak, P.A. Mosier-Boss, M.H. Miles, M. Fleischmann; Thermochimica Acta 410 (2004) 101–107, they write:


    Szpak? The first time I saw this name was in a mail to Vortex that remained impressed in my mind: http://www.mail-archive.com/vo…@eskimo.com/msg97821.html

  • It seems that people don’t understand my issues with the ‘7%error is just experimental noise’ position that Szpak, Mosier-Boss, Miles, andFleischmann (SMMF) (and McKubre by supporting it) take in their measurement ofwater exiting an F&P cell in their 2004 paper.

    They never said that. You made that up. The 7% is caused by evaporation (L - in Fleischmann's paper). It is not experimental noise.

  • It is reasonable to remind ourselves of the reaction Fleischmann had to my 2002 paper. In 2017 Mel Miles published an ‘unpublished’ draft of a paper by Fleischmann in Infinite Energy, which was discussing the same experiments found in the 2004 SMMF paper. In it we find the following (emphasis added) :



    “It is appropriate here to comment also on criticisms which have been made recently17 on the evaluation of the data sets for experiments on “cold fusion.” It is asserted that such evaluations are based on calibrations derived by the statistical methods of regression analysis; such evaluations are deemed to be unsatisfactory in the absence of a proof that changes in the global heat transfer coefficient do not take place. We observe that this criticism does not apply to the observation of the phenomenon of excess enthalpy generation as we do not need to rely on evaluations using regression analysis when making semi-quantitative estimates when these rates are adequately high (as is the case for the Pd-D codeposition system investigated in the present paper). In this case we can simply use the maximum value of the “lower bound” heat transfer coefficient as an estimate of the “true” value and thereby evaluate minimum values of the rates of excess enthalpy generation (e.g., see Ref. 13). It should be noted also that this procedure is quite independent of any method of calibration.”


    Yes, Ed Storms and I both used regression. The fact that F liked a really complicated calibration methodology doesn’t matter at all. The problem arises from something like ATER causing a shift in the steady state heat distribution in a calorimeter that has a spatial sensitivity to heat detection.


    “We also make two further observations on these recent criticisms. In the first place, we have never observed changes in the global heat transfer coefficients in experiments on appropriate “blank” systems.2,11,18 Secondly, the assertion that excess enthalpy generation can be explained by such changes develops a scenario which cannot be investigated by any simple methodology (the heat transfer coefficients and rates of excess enthalpy generation are parallel in the parameter space of the instruments).”


    ‘Blanks’ by definition don’t show the FPHE and therefore wouldn’t show the CCS problem. The ‘simple methodology’ might start with putting the gamma factor back into the calibration equation (energy balance equation) and evaluating its impact…


    “We believe that the onus of proof of invoking changes in the global heat transfer coefficients as explanations of the observation of excess enthalpy generation rests on the authors making suggestions of this kind rather than advancing the hypothesis that the detection of excess enthalpy generation cannot be believed unless it can be proved that there are no changes in these global coefficients.”


    This is funny! What does F think I did in the 2002 paper? Slice bread? What I did was show that trivial changes in the cal constants zeroed out large excess heat signals. I ‘proved’ my point. He simply failed to get it. Personally, I doubt he really read my paper. Of relevance on that point is some of his comments in the recently released Miles-Fleischmann correspondence (Fleischmanlettersfroa.pdf (as revised) on lenr-canr.org), to whit (emphasis added):



    “This brings me to the paper by Kirk Shanahan. Has this been published, will it be published? Incidentally, he did send me a copy of this paper earlier this year but I threw it away in a fit of irritation.” p. 547.


    That’s the recommended way to deal with critics…NOT!


    (BTW for the record, I don’t believe I did that. I sent copies to Ed Storms, had lots of email discussions with him, and then there was the 2-year battle to get it published, which gave ample time for the paper to have been sent around the world multiple times…)


    “I think it would be a serious mistake to regard Shanahan’s paper as an attempt to further the understanding of calorimetry: it is really in the nature of a “spoiler”. He takes a week [sic] paper, sets up a scenario of errors in the calibration (which may or may not be true) and then extends his negative comments to the whole field by innuendo.” Ibid, p. 548


    Hmmm….pointing out a systematic error in calorimetry is not ‘an attempt to further the understanding of calorimetry”? Really?


    And, not by ‘innuendo’, by inductive reasoning.



    “I find it next to impossible to make any connection between Shanahan’s paper and the work we have done” p. 548


    Reading it might have helped…


    “The reason that I regard Shanahan’s paper as a “spoiler” is because it falls pretty well into the scenario of activities which have been used before notably by the Tobacco and Sugar industries. Britain’s first Professor of Nutrition Science was so effectively rubbished that he never got another research grant!” p. 548


    Oh oh! Conspiracy nut thinking… (there’s more of this silliness later too)


    “This brings me again to Kirk Shanahan’s paper. As I said in my FAX of 30/8/02 it would be a mistake to regard this as an attempt to further the understanding of calorimetry. You certainly cannot tell from the paper alone as to what his real intent may have been although the list of acknowledgments is highly suggestive.” P. 553


    Again….pointing out a systematic error in calorimetry is not ‘an attempt to further the understanding of calorimetry”? Really? Also, more conspiracy nut stuff…


    “You could tell him that I believe that Kirk Shanahan has simply applied a routine analysis to Ed Storms’ data.” p. 553


    Pretty much absolutely correct! And I found ‘CCS/ATER’!


    “Of course, if you introduce arbitrary shifts into the experiments (the errors referred to by Kirk Shanahan) then all is pretty well lost - you cannot devise calibration procedures which will detect such arbitrary shifts.” p. 553


    Pretty interesting comment! That’s what I’ve been getting at for years…


    “Incidentally, you may note that the fitting procedures we used in 1989 always included a calibration - we simply did not use a predetermined heat transfer coefficient as seems to be alleged by Kirk Shanahan.” p. 553


    What I actually ‘allege’ is that they determined a cal constant and used it in a later run where the steady state changed and thus made the prior (if only by seconds) calibration invalid. This is exactly what F describes here.


    “It seems to me therefore, that Kirk Shanahan’s paper is just the first step in a procedure which is likely to develop into a useless slanging match designed to justify the inaction of the agencies such as the D.O.E.” p. 554


    Again….pointing out a systematic error in calorimetry is not ‘an attempt to further the understanding of calorimetry”? And that is ‘a useless slanging match designed to justify the inaction of the agencies such as the D.O.E.’? How about it’s an attempt to determine how artificial excess heat signals can arise and fool well-meaning experimentalists?



    Bottom line: Fleischmann obviously didn’t understand my 2002 paper.


    (Note that based on the correspondence dates he probably saw the prior manuscript which can be found in Jed’s database. Technical details are all the same, but some text was changed in the published version.)

  • Evaporation. I have already cited many references.


    No. Yes, you have cited millions of references. Which specifically are you referring to here?



    They never said that. You made that up. The 7% is caused by evaporation (L - in Fleischmann's paper). It is not experimental noise.



    No. I quoted them directly before. Adding a preceding sentence I quote (again):


    The comment has often been made that excess enthalpy

    generation can be explained by recombination of the electrically

    evolved gases. In this experiment, the total consumption

    of D2O was 7.7 cm3 instead of 7.2 cm3, assuming 100%

    Faradaic efficiency, which is within experimental error.


    No other mention of where the water comes from.


    Edit: I forgot you don't know science. "Within experimental error" means in the noise.

  • Ah. I thought you were talking about another paper. My mistake.


    This is indeed a 7% experimental error, for one measurement. It was reduced in subsequent papers. This error cannot occur with Storms, McKubre and others who use closed cells.


    As McKubre and Fleischmann pointed out, the total lost to vapor (evaporation) is very close to the predicted amount for the measured temperatures.

  • This error cannot occur with Storms, McKubre and others who use closed cells


    The problem is what I described here: FP's experiments discussion


    i.e., people who think a 7% is 'just noise' . The CCS/ATER error is a 3% or less error that gives hundreds of milliwatt errors. As I pointed out, that 7% becomes 35% with only 20% recombination.


    So, yes, they don't measure water lost in closed cells. They do however ignore the CCS/ATER problem, which I originally detected in Storms' closed cell.

  • As McKubre and Fleischmann pointed out, the total lost to vapor (evaporation) is very close to the predicted amount for the measured temperatures.


    As McKubre - yes - and I have noted that he presents a possibility. Fleischmann? I still want the exact reference.


    If SMMF knew about evap and considered it important, and if it exactly accounts for the excess their paper should have read something like this...


    "In this experiment, the total consumption of D2O was 7.7 cm3 instead of 7.2 cm3, assuming 100% Faradaic efficiency. The remaining volume arises from water vapor carried out with the electrolysis gases. Using the D2O vapor pressure at xyz oC and integrating that for the abc seconds of the collection period, we obtained exactly 0.5 cm3 as the expected evaporated water volume, fully accounting for the remainder of the collected volume."


    This simply seconds Ascoli65's motion regarding sloppiness in this paper.

  • Why do you need one paper? Several papers cover the details you demand. Why do you insist they be combined into one?


    Jed, there can be multiple papers, I do not say only one paper. However one paper must make clear that the specific results described therein are checked in all relevant ways. It is not good enough for a paper to describe checks in general that convince the author a specific error will not be problematic. To be convincing, a paper must state explicitly that those checks are carried out in the case described therein, otherwise, a sporadic positive could be because of sporadic failure in one of the normally true but not explicitly checked assumptions.


    A correct positive result can depend on 2 different assumptions, each of which is usually true, where the positives come from cases where one or other assumption breaks down. Clearly in such a case it is necessary to be precise about what is tested when.


    This type of issue is not a problem when effects have replicable results. It is immediately detected. Where effects are supposed to be variable then they cannot easily be distinguished from assumption failure unless everything is tested together.

  • However one paper must make clear that the specific results described therein are checked in all relevant ways.

    Again, why do you demand that one paper do this? Why not look at several papers to make it clear that specific results are checked in all relevant ways? The same results are often discussed in different papers. If you want to know a lot more about Fleischmann's calorimetry, for example, see:


    http://lenr-canr.org/acrobat/MilesMisoperibol.pdf


    Miles is better at explaining these things then Fleischmann was. I believe your focus in recent discussions here was on the effects of vaporization. That is an important issue, and as McKubre wrote here, it is measured and included in the calorimetric equations. See especially Appendix I, and look for the term L. See, for example, p. 26:


    "The term involving the enthalpy of vaporization (L) contributes -4.9103 mW and -52.9857 mW, respectively, at these two cell temperatures, (Ref. 2)."


    [The two temperatures are 40 deg C and 80 deg C. The cell is usually at 40 deg C, and ~5 mW is much smaller than the excess heat in the weeks leading up to the boil-off event. Obviously, vaporization dominates during the boil-off.]



    There is another issue. What you consider "relevant" may not be what the authors think is relevant. For example, you think that water might condense in droplets on the wall and then be driven out by steam which is barely above 1 atm. Most people probably agree with me that this is impossible, or if it does happen, it could only cause an error in the milliwatt range (or probably micro-watts), whereas the excess was 133 W, so this could not possibly be a significant factor, and it is not relevant. You can come up with an endless list of phantom concerns such as condensed droplets being pushed out of cells. However, you cannot point to an experiment showing this is a real concern, or any textbook, and you cannot explain why it does not happen during calibrations. For that matter you cannot even explain why the cell boiling in the first place, given the low input power, or why the Kel-F plug melted. How could droplets cause all that to happen? So I do not think this objection has merit. As I said, I do not think you have found any significant error in this or any other mainstream experiment.


    To summarize, your droplets might produce artifactual heat of ~0.001 W, but that is 100,000 times less than what was observed. This is a drastic quantitative failure to explain anything. Other reasons it fails include: artifactual heat would also be seen with the Pt-H2O control. You cannot give a reason why the choice of metal and water would affect this mechanism. The artifact could not boil water with far less input power than normal; it could not boil water after the power is cut off (when it falls below the anode and cathode) and it does not melt plastic. Only real heat can do that. So your hypothesis cannot begin to explain the facts. You have made no serious attempt to quantify this hypothesis or defend it.

  • As McKubre - yes - and I have noted that he presents a possibility. Fleischmann? I still want the exact reference.

    And I still gave you an exact reference.


    http://lenr-canr.org/acrobat/Fleischmancalorimetra.pdf


    The deviation from the mean for the heat transfer coefficients are described on p. 4 and p. 5. I believe that is what most people would consider the error margin.


    Here's another paper. Have fun:


    http://lenr-canr.org/acrobat/MilesMisoperibol.pdf

  • This type of issue is not a problem when effects have replicable results.

    These results have been replicated thousands of times at ~180 labs. Fleischmann and Pons alone ran 16 cells at a time, about 100 times as I recall, and it worked in most cases. So what are you talking about? Compared to many other experiments, this is highly replicable.


    Perhaps you want to see results that work more often than they fail. A high success rate, in other words. That is not the same as replicability. You can have low success but high replicability by doing the same test many times, for example with an array of 16 cells (Fleischmann) or 100 cells (Bockris). If you reject results with a low success rate, you will reject things like cloning, which used to work in approximately 1 per 1,000 attempts. They had do it thousands of times but they ended up with many animals; more than enough to prove it worked.


    If you don't like low success rates, and things which are very difficult to replicate, you will certainly reject the Top Quark!

  • Regarding the difficulty of measuring the exact amount of make-up water per day, Fleischmann wrote (and Shanahan quoted): "In this experiment, the total consumption of D2O was 7.7 cm3 instead of 7.2 cm3, assuming 100% Faradaic efficiency, which is within experimental error."


    That is indeed 7%, but the error goes down over time, as you make more measurements. If you consistently measured 0.5 cm3 less than the actual amount added to top off the cell every day, after a week the water would overflow from the cell. You would know that you are doing it wrong. If your mistake was in the other direction, the water level would drop. You could not fail to see it. Even with the half-silvered Dewars they could see this using a small dentist's mirror.


    In most cases that I know of, make-up water is added to an electrochemical cell with a syringe. You can see for yourself what the error margin is for a syringe. Get a syringe, measure out 7.2 ml of water, and squirt it into a test tube that is mounted on milligram weight scale. Repeat 50 times. You will see that sometimes you measure out ~7.7 ml, and sometimes ~6.7 ml, but over time it comes out close to 7.2 ml. If you consistently add 0.5 ml more than you think you are adding, the water level in the cell will be much higher than it should be, and the weight will be 25 g too high.


    Also, they reduced the daily error after they wrote that.

  • And I still gave you an exact reference.


    http://lenr-canr.org/acrobat/Fleischmancalorimetra.pdf


    The deviation from the mean for the heat transfer coefficients are described on p. 4 and p. 5. I believe that is what most people would consider the error margin.



    Yah, not quite. What F&P are doing in that section on p4 and 5 is making a series of repetitive measurements on 1 experiment. Their replicability on the set of 19 is good. Now they need to do that '19' times on '19' experiments to get a true value for the variation in the k from experiment to experiment. Typically that variation will be at the 3% level or so.


    Here's another paper. Have fun:

    http://lenr-canr.org/acrobat/MilesMisoperibol.pdf


    Yah right. 34 pages to look for a comment on where they measure the mass of water that leaves the cell. I might get around to it...


    As I recall, Miles has published standard deviations on a few of his calibrations, and they are of the same order as I've mentioned already, i.e. 1-3%. That just means that he confirms what I saw in Ed's ICCF8 report. What this means to me is that there is a natural variation in calibration constant determination on the order of 1%. The CCS needed to zero out significant excess heat signals is also on the order of 1%. This means extreme care is necessary because everyone is working 'in the noise'. However, the usual claim is that 'the noise' is on the order of 50mW. That claim is incorrect based on Storms' and Miles' work, but because everybody is 'old school' on this, they don't realize this and think I am crazy for claiming their signals might be noise. However, I showed above with simple math it doesn't take much to wipe out apparently big signals. Bottom line: Everybody needs to check for a CCS.


    FYI - the equations in the F&P paper that have 'L' in them are for energy calculations. When I was writing my whitepaper, I derived the same terms independently to check them, and I found what they wrote to be fine as far as they went. Their equations do not contain lost energy due to entrained water or electrolyte. And while F&P may have included 'L' in their energy equations, they may not have measured water loss accurately enough to confirm no recombination during FPHEs. Published data to support their implicit claims is required. The unfortunate problem seems to be that F&P and most other CFers assume recombination is limited to 1-2%, but that is explicitly for electrochemical recombination. What I call ATER is not electrochemical, and it can go to 100%.


    McKubre claimed people have measured water loss and even put pH indicators in the collected water to check for electrolyte content. I want to see that published, not just asserted with a "Trust me" appended. I have looked for that thing since I started examining this field in 1995 and I've never seen it in print. And as I told McKubre, for revolutionary physics, there's no room for "Trust me".

  • Yah right. 34 pages to look for a comment on where they measure the mass of water that leaves the cell. I might get around to it...

    I suggest you experiment with a syringe instead, as I described above. See for yourself. If they measured the make-up water incorrectly, the cell would overflow or the waterline would fall, so you really don't need to consider that possibility. If they only added make-up water once, you would need to consider that.


    The calorimetric equations describe the physics, which tells you where the water went, and how much went where for what reason. The measurements agree with the physics. If they did not, or if the measurements were made wrong, the cell would overflow, and it didn't.


    I realize that "see for yourself" is not what you do. You are allergic to trying things, and doing practical tests such as measuring water with a syringe to see what the margin of error is. That is why, for example, you think that a bucket of water will evaporate overnight, and you think there are many reasons why a commercially made piece of plastic designed to be used in a test tube might be distorted by electrochemical experiments. You don't actually know any reason, but you wave your hands and pretend there are such reasons and hey presto, you believe it!

  • McKubre claimed people have measured water loss and even put pH indicators in the collected water to check for electrolyte content. I want to see that published, not just asserted with a "Trust me" appended. I have looked for that thing since I started examining this field in 1995 and I've never seen it in print. And as I told McKubre, for revolutionary physics, there's no room for "Trust me"

    Kirk: Too many words above this quote, and too many errors and quirks (not just yours) to address. Thanks Jed for helping. I don't trust either. So I do my own experiments, attended the conferences and battled out personally some of these issues with the originators. Your are not to be faulted for "not having been there", and I agree that we as a (now CMNS) community could/should have done a better job of documentation in early work. I have complained to the community on this point. But, as Jed can (also) confirm - there was plenty of discussion. Some heated.


    THH: somewhere up there you mention something like "changes in calibration" of our calorimeters and Kirk also bring up this point. I just want to make sure that you recall that our mass flow method requires very little calorimeter dependent calibration. >99% of the thermal energy leaves with the mass flow. Presuming that the heat capacity of air-saturated water is well known (it is), all we need to calibrate are: the temperature difference, ∂T; the mass flow rate, ∂M in unit time ∂t. Of these only T is "difficult" and we used redundant pairs of platinum RTD's calibrated against a calibrated quartz crystal NIST secondary standard (often in the same bath). This method was spoken about often and early in nauseating detail, and is published in our first EPRI report (and elsewhere). The issues you raise are real and relevant. But I demur on the point that they have not been covered.


    I am headed off on travel shortly. Will log in periodically if possible. These are important concerns - to me not because there are unresolved errors, but because there are clear confusions that "those who were there" might help to clear up (if possible). Personally I would prefer to see our collective intellect (meaning the whole forum) directed to the present and the future. All this retrospective makes me feel old.


  • I'm mostly just following along this thread, but I have a quick question regarding the top-up:

    Does not one fill a syringe with (for example) 10 ml, then add water electrolyte or whatever as required to a specific level in the container, look at the syringe (which if 7 ml were added would now have 3 left in it), subtract 3 from the original 10 to arrive at the 7 ml added?

    In this example there would be two top-up measurement error sources: the syringe 10 ml starting amount measurement, and the remaining amount measurement, (ignoring simple subtraction errors), neither of which would cause the average container level to rise or fall beyond the mean level measurement error (another potential error source) that the container is being topped up to. The top-up-to-level error would average out over time, while the syringe measurements could have a cumulative error (that does not lead to an over-full or under-full container).


    Does that make sense?