Mizuno reports increased excess heat

  • From previous paper (and referenced here) pipe dia is 5cm. Measurements are made at "Up 1cm, Up 2cm, Up 3cm, R 1cm, R 2cm, R 3cm, Center".


    It is not clear what this means. The measurements cannot be referenced to the centre (only 2.5cm of movement possible). If referenced to edge they apparently show airflow uniform (very close) over the middle 3cm of the 5cm tube - assuming symmetry.


    It means the anemometer probe was traversed across the air flow outlet tube.



    Actually this ignores the size of the probe


    It is small. You could look it up.


    More to the point, if the calorimetry is not working correctly, how do you explain results such as the 50 W calibration? Quoting the paper "Over 24 hours, average input electric power was 50.6 W. An average of 46.6 W of heat was captured in the stream of air. After taking into account heat losses from the calorimeter walls, the average was 50.5 W." Do you think that is a coincidence? It just happened to be within 0.1 W? There are many other calibrations at other power levels. This is first-principle calorimetry. In other words, if the air flow is measured incorrectly, the results will be wrong. It does not depend on a calibration, although a calibration does confirm it.


    If calibrations do not prove the instrument is working correctly, what would? How can anyone prove anything, if we don't believe calibrations?


    To put it another way, when calibrations at different power levels done repeatedly over two years show that the instrument always measures output correctly to within 0.1 W of input power, how likely is it that 53 measurements of 30 to 100 W are wrong? How likely is it that several ~250 W measurements are wrong? The traverse test is important. It lays to rest one of the big concerns of air flow calorimetry. However, the calibrations alone should be enough to give confidence with excess power at such high levels. It was questionable back when excess heat was typically around 10 W, but it is not questionable at 30 W, 100 W, or 250 W.


  • Thanks for that, Jed. Beg to differ over professionalism of write-up. It may seem to you a game, but communication is often that, and matters.


    Ok, actually I did not note properly the R20 results. I still cannot say much about these. They are billed as sample results and only impressive because of the comparison with the control. Unfortunately we do not have enough information to evaluate this.


    If the control and the rest of the setup is identical I agree these results are impressive. That is likely not true.


    You can see that the active graph from Figure 5 takes some time to reach its final value, presumably due to stuff heating up. The control graph is flat. No heat-up. We need much more information about how this control is made.


    Re the previous paper - please see my note about the velocity profile measurements. If the probe is small then my point is exactly as stated - proof of at least 36% that assumed. of course it will be more, but no way to know how much more, and the amount will vary with speed of flow a lot. These measurements are useless - and proper peer review would pick that up.

  • More to the point, if the calorimetry is not working correctly, how do you explain results such as the 50 W calibration? Quoting the paper "Over 24 hours, average input electric power was 50.6 W. An average of 46.6 W of heat was captured in the stream of air. After taking into account heat losses from the calorimeter walls, the average was 50.5 W." Do you think that is a coincidence? It just happened to be within 0.1 W? There are many other calibrations at other power levels. This is first-principle calorimetry. In other words, if the air flow is measured incorrectly, the results will be wrong. It does not depend on a calibration, although a calibration does confirm it.


    If calibrations do not prove the instrument is working correctly, what would? How can anyone prove anything, if we don't believe calibrations?


    Jed - if you read my original post properly you will see that I agreed either calibration or direct calculation (first principle calorimetry) could be enough to make this safe. I'm OK with calibration only, though I'd rather both. But for calibration you need exactly the same conditions guaranteed between calibration and active tests. For first principle calorimetry you need safe measurements.


    Now in the original paper the results were too small for possible differences to be ignored.


    The new paper has much better results. But these are quantitative (R19) based on we do not know what, and the quantitative measurements referenced (e.g. airflow profile) prove nothing. Or they are referenced to a control (R20) but it is unclear what is the methodology or differences between the two cases. That would need tying down. perhaps wait till we have definite, rather than sample, results before analysing R20.


    All I want is one precise description of tests that show these replicable +50% results and how they were obtained. should not be difficult if they are real, but has not yet been done.


    The merit of this +50% is that with this calorimeter (and some care) both first principle results and calibrated results should be easily safe and prove large amounts of excess heat well beyond anything expected. Two different methods doing this provides better integrity. But, as shown in the second paper, neither method is proven.

  • Thanks for that, Jed. Beg to differ over professionalism of write-up. It may seem to you a game, but communication is often that, and matters.


    As I mentioned, I was trained in technical writing and translation at Cornell and at a major corporation, and I made a very good living programming and writing. I have written newspaper and magazine articles on various subjects in both English and Japanese, beginning when I was in college. I have edited roughly 200 papers for the JCMNS. So I guess that makes me a professional.



    If the control and the rest of the setup is identical I agree these results are impressive. That is likely not true.


    Likely not true? Did you read the first paper? The control and active reactors are placed in the calorimeter chamber, equidistant from the walls and from one another. The calorimeter is sealed. It is tested and calibrated extensively with the traverse and heat recovery tests, stepped through about 100 power levels. Then the active reactor is run, followed by the control, back to the active, and so on until the end of the tests.


    Are you sure it is "likely not true" that conditions are identical? Why would they be different? Or, let me put it this way: Can you suggest a way that would make the conditions more identical than this? How would you do this experiment?

  • Jed, you are not precise in reading what I say and answering it.


    Are you sure it is "likely not true" that conditions are identical? Why would they be different? Or, let me put it this way: Can you suggest a way that would make the conditions more identical than this? How would you do this experiment?


    I don't think surity exists with this write-up? You are asking me am I sure about how much I'm not sure? LOL. I've given my reasons: exact conditions of control versus active not stated, and, from graphs, time scale does not match, because active shows heat-up, control appears to have already heated up.


    Did you read the first paper?


    Yes, I did, some time ago. But now we are discussing the second paper, and we cannot assume the methodology for R20 sample tests was identical to that previously used.


    As I mentioned, I was trained in technical writing and translation at Cornell and at a major corporation, and I made a very good living programming and writing. I have written newspaper and magazine articles on various subjects in both English and Japanese, beginning when I was in college. I have edited roughly 200 papers for the JCMNS. So I guess that makes me a professional.


    Jed, I respect you as a good and professional writer, not as a professional scientist experienced at writing stuff that can be published in high impact western journals. That is no criticism. It is tough.

  • I'm wondering if by adding an insulation layer to the reactor one could model the heat exchange coefficient such that it resambles air flow cooling although water

    is cooling at the outside.


    The basic principle of this calorimeter is good: as Jed has pointed out it provides ab initio (first principle) calculation of heat output with relatively small losses (maybe 20%). It can also quite easily be calibrated.


    However there is a bit of work needed to do either of these things safely.

  • I'm again :)

    I would like to know if Mizuno tested directly 3 Ni layers or one by one ?

    Why only 3 ?

    It should be nice to know XH correlation with number of layers rather than specific surface .


    Thanks again for this Lenr research contribution .

  • All I want is one precise description of tests that show these replicable +50% results and how they were obtained. should not be difficult if they are real, but has not yet been done.


    Yes, it has been done. The second paper describes the tests in great detail, and it describes how the results were obtained. I have never read a cold fusion paper with as much hands-on detail as this one. Only the Storms and Cravens papers come close.


    The description and photos are of the R20 reactor, but the R19 was very similar. The nickel mesh reactants and the procedures used to make them was exactly the same.


    This is partly a matter of luck. Some the nickel reactants work better than others. The method of making them is crude, so you have to expect large variations in performance.


    It is ironic. When you say this is a not "a precise description" and it "does not show" how these results were obtained, you are complaining about the one thing the paper does actually accomplish. It is the one aspect of this paper that is superior to most cold fusion literature, and far superior to the recent Google paper in Nature. It seems you will not take "yes" for an answer. If you were complaining that this paper lacks theory, you would have a point. I left 20 pages of theory on the cutting room floor.


    Regarding the methods and how these results were obtained, what more do you think we should have said? What is missing? What questions about the materials and methods do you have that are unanswered?

  • I'm wondering if by adding an insulation layer to the reactor one could model the heat exchange coefficient such that it resambles air flow cooling although water

    is cooling at the outside.


    Mizuno tried that, years ago. It did not work well. I wouldn't want to try it with these reactors because you must have a way to cool them off quickly.


    Looking at the data, I have a sense that heat flow through the reactor and temperature gradients are important. I cannot quite put my finger on why. (If I could have, I would have put it in the paper.) It is just a feeling.

  • Great news and remarkable simple method.


    Even with the current described nickel mesh preparation I would expect that with the first run with Deuterium gas at 100 - 300 degree C some remaining surface Nickel oxide will be reduced while forming small amount of D2O which preferably needs to be evacuated as well. Just a minor thing I guess, but maybe this is why a bigger reactor vessel performs better.

  • It means the anemometer probe was traversed across the air flow outlet tube.


    Right. So you agree with me it shows constant airflow at:


    1, 2, 2.5, 3cm relative to one wall or:


    -1.5, -0.5, 0, 0.5 relative the centre.


    Thus we have from -1.5cm to +1,5cm (the middle 3cm of the pipe, assuming symmetry) tested for constant flow. The remainder is not tested.


    Because area scales as r^2 this means that 3cm^2 / 5cm^2 has been tested or 9/25 = 36% of the cross-sectional area is known to have constant velocity.


    The equations for flow in pipes depend on Reynolds number but basically you expect flow to be much smaller close to the sides of the tube and constant high over the centre portion. So these measurements prove absolutely nothing of use here, except that it is not turbulent flow (that would not have the centre portion flat velocity profile).


    We could maybe bound (from above) the Reynolds number from this data and therefore bound (from below) the average air velocity relative to the centre velocity. This is the sort of thing that decent peer review would clearly pick up.


    You agree with this?

  • Did you read the first paper?


    Yes, I did, some time ago.


    I uploaded it at 11:00 this morning, so you could not have read it more than 5 hours ago.


    But now we are discussing the second paper, and we cannot assume the methodology for R20 sample tests was identical to that previously used.


    The paper says they were identical. It says that twice. I just told they were identical. So, yes, I think we can assume they were identical. Or are you saying you don't believe what Mizuno and I wrote?


    Maybe you don't believe him. Do you think he has been sending me fake labs notes and photos all this time? Why would he go to all this trouble pretending to use the same methodology when it was actually different?


    If he did use different methodology, I would have described the differences in meticulous detail. If he had so much as used a different brand of dish detergent, I would have listed it. As I said, I used to write manuals describing things in meticulous detail. We had to do that, because when things went wrong, Large Sums Of Money might vanish in a short time.

  • Even with the current described nickel mesh preparation I would expect that with the first run with Deuterium gas at 100 - 300 degree C some remaining surface Nickel oxide will be reduced while forming small amount of D2O which preferably needs to be evacuated as well


    Yes. As noted it is very important to purge oxygen before beginning the tests. See steps 1 - 7 where it says "After cleaning."

  • Did you read the first paper?


    Yes, I did, some time ago. But now we are discussing the second paper, and we cannot assume the methodology for R20 sample tests was identical to that previously used.


    The Ref (1) Mizuno refers to is the first (ICCF21) paper. Here is what he says in this newer paper:


    "The only major change to the experiment since our last report has been the design of the
    reactor. The reactants and methods have not changed, so we believe the reactor design is the

    cause of the improved performance

    The air-flow calorimeter used in this study was described in Ref. [1]. The reactor, nickel

    reactants and methods were also described."

    The air-flow calorimeter was used in this study is described in detail in Ref. [1]. It is briefly described here. The instrument has not been changed."


    In the ICCF21 paper, Mizuno describes in detail his "side by side' calibration set-up, and use of 2 types of calibration. So IMO it is important to read the first paper, when doing a critique of the second.

  • Mizuno tried that, years ago. It did not work well. I wouldn't want to try it with these reactors because you must have a way to cool them off quickly.


    Looking at the data, I have a sense that heat flow through the reactor and temperature gradients are important. I cannot quite put my finger on why. (If I could have, I would have put it in the paper.) It is just a feeling.

    Thanks! This gave me an idea. Assume that we can control the degree of insulation, from very insulated to basically conduction in an insulation layer around the

    cylinder and place it in a water bath. Then if temperature is only what trigger the reaction you could essentially just manage temperature without any heating at

    all and the only electricity would go to the controller. This sounds that i'm after something similar to control rods in nuclear reactors.

  • The equations for flow in pipes depend on Reynolds number but basically you expect flow to be much smaller close to the sides of the tube.


    Very, very close to the sides of the tubes. Only a tiny fraction of the flowing air is affected. Other tests confirmed this. The amount is not enough to affect the results more than 0.1 W, as you see in the calibrations. As I said, this is a first principle method. If the flow rate is wrong, the answer will be wrong. It is simply weight of air * heat capacity of air * temperature difference. If the flow rate is wrong, the weight of air will be wrong, and the answer will be wrong. There is no getting away from that. It is not possible that both the weight of air and the temperature are wrong, and they just happen to balance out and give the right answer at every power level.



    So these measurements prove absolutely nothing of use here.


    The reviewer disagreed with you. I do not know who he was, but he knew a great deal about this. More than me, and I suspect more than you. But again, you are ignoring the calibrations. They showed that the output was very close to input. Within 0.1 W at 50 W. That's one part in 500, which is incredible, by my standards. What is your take on that? Do you think it is a coincidence? Do you think it also a coincidence that calibrations at other power levels balance, and calibrations taken each time the calorimeter is sealed up also balance?


    If calibrations do not convince you the machine is working correctly, what would convince you? Can you suggest something better than a calibration to confirm that the flow rate is right? If there were any evidence the machine is malfunctioning, you might have a point, but I cannot see why you think the measurements "prove absolutely nothing" when the machine is working. A calibration is worth more than all of your speculation, or the speculation of a hundred people tied together. The instruments themselves tell us they are working. You cannot dispute them. That's how experimental science works. If you don't see an error in the results, there is no error.