H. Zhang: First successful independent Mizuno R20 replication!

  • Or, you could do it. Read the report. Look at the calibrations, and do the arithmetic. You will see the calorimeter can measure 150 mW with confidence.


    Jed, let me spell this out so that there is no misunderstanding.


    (1) I'd expect this setup to be much more accurate than Mizuno's, obviously.

    (2) I'd expect the calorimeter accuracy to be much better than the transient detected (9W)

    (3) I don't know whether the accuracy is higher than the 1.9W continuous excess you note (2% accuracy).

    (4) I strongly question your accuracy of 150mW (0.15% accuracy). That is very high for any calorimetry.

    (5) More generally I don't have enough info from your link to do an error analysis.



    The link is a lot of pictures with a few numbers. The only error estimation I can find is 0.01C as the temperature measurement accuracy. You may be better off than me able to read the Japanese.


    For error estimation here we would also need:


    • Input power measurement error (?)
    • Sensitivity (temperature change / W) (can be worked out from data)
    • Second order effects (change in calibration with room and sink temperature)


    I think you are confusing power out resolution - which I agree is very high here, with accuracy. Anything that might change between calibration and active runs and affects the result degrades accuracy. That is power measurement drift and resolution, voltage measurement drift and resolution, room temperature, water inlet temperature, etc. This setup is relatively insensitive to room temperature change because of the water cooling and the differential temperature ,measurement, but even so that does not mean there is no sensitivity there.


    THH


    PS - i mention resolution because many of the presented figures have a 3 digit resolution as displayed, which worst case leads to +- 0.5% resolution error.

  • JedRothwell , do you know who H. Zhang is ? I have been doing some googling and there is plenty of H. Zhang researchers, so far the only one I have not found a picture of to know if matches the face in the paper you posted is a Haifeng Zhang who has done some Pd and phonon work and belongs to the Chinese Academy of Sciences, something that should impress some because currently the Chinese Academy of Sciences is the highest ranking research institution by the Scimago index. Is this Haifeng Zhang the same H. Zhang that made the analogue experiment we are discussing here?


    https://www.mendeley.com/authors/56957750000/

    I certainly Hope to see LENR helping humans to blossom, and I'm here to help it happen.

  • JedRothwell

    Quote

    Ah. I think you misunderstand. The numbers would be changed by insulation. Suppose it takes 100 W to heat the inside of the reactor to 250°C. * Now suppose you wrap the cell in insulation. It will take much less power to heat the inside, because the heat will not escape as easily. It is like the difference between heating the inside of a Dewar flask versus a regular test tube. Or like heating a house with poor insulation versus a well insulated house.

    You're right. I'll have more on this if I find time.

  • In other words, the heater is needed to reach a certain temperature, and it would reach that temperature with much lower power. Instead of 100 W and 110 W, you might see 30 W in and 40 W out. Or 5 W in and 15 W out. That is, if 5 W was enough to raise the temperature to the operating point, with something like a large Dewar.

    HOWEVER, having said all that, there is one possibility I did not cover. It may be that the heater is somehow triggering the reaction with something like IR radiation. In that case, if you were to reduce power to only 5 W in a well insulated container, the temperature might go up but the reaction might not turn on.


    In that case, you would use a lamp or something that produces more IR than the present heater does.


    I do not know if IR plays a role, but some people have speculated that it might.

  • Zhang tells me the cell is again producing excess heat. It recovered from argon quench. It has been running for about 14 hours, 6 W maximum, with a lot more fluctuation up and down to zero and back up than last time.


    However, there may be a problem. As shown in this graph, two of the fluctuations appear to be significantly below the zero line. They are endothermic. I do not think cold fusion ever produces an endothermic reaction, so I think this may be a chemical effect.



  • JedRothwell, do you know who H. Zhang is ? I have been doing some googling and there is plenty of H. Zhang researchers, so far the only one I have not found a picture of to know if matches the face in the paper you posted is a Haifeng Zhang

    I do not know his affiliation. His first name is Hang.


    张 Zhang (surname)


    航 Hang


    I believe Zhang is a common name in China.

  • I'd reckon that both exothermic and endothermic behaviour is unlikely. Could be some change from equilibrium conditions in the calorimeter or reactor that affects output. No idea what, but something that changed effective thermal resistance could drive both positive and negative excursions due to stored heat being released.

  • However, there may be a problem. As shown in this graph, two of the fluctuations appear to be significantly below the zero line. They are endothermic. I do not think cold fusion ever produces an endothermic reaction, so I think this may be a chemical effect.


    Cold fusion can produce fast & large endothermic effects. It is well known why, as we had many times measured the same. But I will not feed the trolls here as they anyway do no experiments and know nothing about LENR & real experiments.

  • About the

    Previously, I wrote that Zhang's tests tend to "peter out" after about 3 hours. I said they stop producing heat. That is incorrect. The spreadsheet shows the heat did not peter out completely. At the end of Test 1 the reactor was still producing 0.2 W. At the end of Test 2 it was producing 0.4 W, and at the end of the graph for Test 3, it was producing 1.8 W.

    Question to experts.


    Is the sequence of test a very good refutation of error in calorimetry? I'll tell the vision of a naive skeptic facing that chain of test.

    • before that a calibration, showing very sane behavior... no strange artifact observed.
    • First a positive results with clear power... but is it artifact? because something changed with mesh of deuterium?
    • then a test with noble gaz. no heat, a blank, a calibration? but is it because of noble gas different behavior? because of changes ?
    • then test with D2... nearly a blank... so blank is not because of noble gas... if it was only a D2 chemical of dissociation effect, why not now?
    • then new test, with more heat... what have changed if not LENR ? what other artifact ?
    • then new test and anomaly is increased... why does the artifact if it is have changed, increased again ... hard to defend the artifact theory.

    The observed quenching may be a very good tool to recalibrate the cell and prove it is not artifact ?


    is it good reasoning?

  • I'd reckon that both exothermic and endothermic behaviour is unlikely.

    What does that mean? If it is a chemical reaction, both behaviors might occur. With cold fusion, an exothermic reaction is likely, but an endothermic one is not, based on all other experiments on record.

    Could be some change from equilibrium conditions in the calorimeter or reactor that affects output.

    The calibrations rule that out, I think. If there are is no chemical or cold fusion reactions, then a test is nothing more than resistance heating, and it should work exactly the same way as a calibration. A control test and an active test would not be "close" or "similar" the way they are with electrolysis. They would be exactly the same in all respects. Indistinguishable. Calibrations with resistance heating never show any heat bursts or endothermic events. I don't see how the presence of a nickel mesh or deuterium gas can change amount of heat produced, unless they are reacting (either chemically or with a nuclear reaction). The deuterium cannot affect the sealed resistor. The resistor cannot "know" there is nickel and deuterium in the cell.


    No idea what, but something that changed effective thermal resistance could drive both positive and negative excursions due to stored heat being released.

    "No idea what" means you are waving your hands. Why would thermal resistance change in a sealed resistance heater? Why would deuterium change it, and why would argon change it back to what it does in a calibration? Neither gas reaches the heater. The choice of gas in the cell can have no effect on the heat once it emerges from the cell and is detected with the Seebeck devices.


    If you cannot think of a way your hypotheses might work, you cannot expect anyone else to think of one. I do not think it is valid. It is hand-waving, as I said.

  • You can bet is a common name. Many scientists carry the name.


    I kept googling with both names and the only Zhang, H. that came up related to LENR was an engineer working with Shongsheng Jiang at LENR China. Could this be the same Zhang ?

    I certainly Hope to see LENR helping humans to blossom, and I'm here to help it happen.

  • The calibrations rule that out, I think. If there are is no chemical or cold fusion reactions, then a test is nothing more than resistance heating, and it should work exactly the same way as a calibration. A control test and an active test would not be "close" or "similar" the way they are with electrolysis. They would be exactly the same in all respects. Indistinguishable. Calibrations with resistance heating never show any heat bursts or endothermic events. I don't see how the presence of a nickel mesh or deuterium gas can change amount of heat produced, unless they are reacting (either chemically or with a nuclear reaction). The deuterium cannot affect the sealed resistor. The resistor cannot "know" there is nickel and deuterium in the cell.


    We's need to know that the sequencing and length of the calibrations was comparable with the active runs. Redoing calibrations, so you spend as much time doing them as active runs, would rule that out. Not clear that has been done here. After all, some active runs are null.


    Whatever there is, is intermittent. You need some effort to be sure it is active system specific, rather than just sporadic.


    If we had temperature variations in the reactor then a reaction could be both exothermic and endothermic driven differently at different times I guess.

  • "No idea what" means you are waving your hands. Why would thermal resistance change in a sealed resistance heater? Why would deuterium change it, and why would argon change it back to what it does in a calibration? Neither gas reaches the heater. The choice of gas in the cell can have no effect on the heat once it emerges from the cell and is detected with the Seebeck devices.


    Main difference between you and me. I accept "I don't know" as possible state when things don't add up. That is not waving hands, and not knowing what something is does not prove some other known but speculative mechanism. It does not disprove it, either.

  • We'd need to know that the sequencing and length of the calibrations was comparable with the active runs.

    No, we wouldn't. All we have to know is that the output heat in calibrations is always equal to 99% of input. Output never exceeds input. Which you can see in the report. Whereas in the active runs, output does exceed input. The time scale is irrelevant. Whether it takes 10 seconds or a week for the excess heat to appear, and whether the heat burst lasts for 1 minute or 3 hours, the calorimeter always shows output exceeding input. It might be from a chemical reaction, or it might be from cold fusion, but output is definitely above input.


    I do not understand why you think the time scale might play a role. I suppose if you started a calibration and ended it a minute later, before the cell reached terminal temperature, it would show less than 99% recovered. However, it would never show more than 99% recovered. You can leave it for a day, a month, or a year, and that never happens. Whereas an exothermic reaction of any sort shows up within minutes in this calorimeter.



    Redoing calibrations, so you spend as much time doing them as active runs, would rule that out.

    No, that is completely wrong, and by saying it you reveal that you do not understand calorimetry. You need to repeat calibrations before and after tests. You need to test at various power levels. You need to continue the calibration long enough to reach a stable temperature and stay there for a while. But there is no need to make the calibration as long as the active run. You cannot point to any physical law or anything in the calorimetry textbooks saying that. You just made it up out of whole cloth.


    The calibrations might take a day and the excess heat from the active run might last 111 days, or 3 months (as some have), but there is no need to go back and do a calibration 111 days long to confirm that. For that matter, if the heat burst only lasts 10 minutes, there is no need to do another calibration lasting only 10 minutes. The comparative time scale of the two events has nothing to do with it.

  • The comparative time scale of the two events has nothing to do with it.


    Well, I only partly agree. The longer the calibration runs the more the influence of ambient conditions (like the weather) becomes visible and can be taken into account. Because of the daily temperature cycles in the lab anything less than 24 hours would be folly, and preferably a calibration run should take in several working days and a weekend. This should be done at both the beginning and end of a very long test run, to allow for seasonal changes such as a lower or higher sunshine level inside the lab.


    But there is no way it must be the same duration as the test. In a better world I would like to see a matched pair of reactors, test and control which have been exhaustively calibrated before testing begins and then run side-by side in their own calorimeters. But all this costs time, money and takes up space Mizuno might not have. Old London vernacular 'When if's and ands become pots and pans I can cook some dinner' Alternative vulgar version 'If my aunt had balls she would be my uncle'.


  • Sigh, Jed, it is frustrating talking to you. You are full of certainty: things are either certain, or impossible to know. And all standard well-conducted experiments seem, from your POV, to be in the certain category. My view differs from this, I've found any experiment, no matter how foolproof, can go wrong not because of new physics, but because of mistake, or unusual intermittent equipment failure, etc. Those deviations could, for all we know, be caused by an intermittent measurement device lead with capacitive smoothing after and a more complex frequency response than a simple RC time constant.


    Suppose there is some intermittent (like, happens only on Mondays) change in the environmental conditions. Thus: lower water pressure means the cooling tubes develop air bubbles. Or, some metallic join connects or not. Or something. In the real world you can never be sure there are no such environment-determined unexpected errors. You try to control the environment (by using cooling water and monitoring its temperature) but you can never be sure that something has not changed. Which is why you do calibration before and after, as you have stated, and when experiments give noisy results you check all the apparatus to find the cause.


    In the case that behaviour is variable - some effect (in this case temperature excursions as measured by the seebeck devices) - is clearly unstable. You limit what that affect can be. To your credit you do not assume it is an extraordinary nuclear reaction - you allow it could be one or more chemical reactions. But you do not allow it to be something to do with the calorimetry or measurement.


    Well, I'll agree about half that, it could be chemical reactions. I'm pointing out that it could also be something changing in the apparatus. Something that trapped heat and then released it would naturally generate both positive and negative excursions from equilibrium. Some weird electrical intermittent fault could cause that. Also, given that the occurrence of this event is variable in the active runs, we cannot know it will happen in any specific calibration run. More time investigating calibration behaviour gives more chance to find such an artifact, if it happens during a cal run that rules out lots of things.


    It is however also possible that the presence of a Ni foil, or even Ni foil + Pd catalyst, is needed to generate this artifact. In that case it could be some chemical reaction as you suggest, although as I have said the fact that we have positive and negative excursions makes that a bit less likely, and i would not rule out some other (physical) effect correlated with the presence of the Ni foil.


    We live in an open world where in spite of best efforts strange things can always happen, no apparatus is perfect, especially when newly constructed. When things are weird you don't rule out possibilities.

  • But there is no way it must be the same duration as the test.


    Actually Alan, I'd partly disagree with that, at least when as here something anomalous is found. (And of course a positive LENR indication would be anomalous).


    It is quite common for equipment to fail only after some considerable time of operation. Short cal runs with much longer active runs would not detect such a failure which would then be categorised as LENR.


    In addition, suppose you have some error that only evidences one in 10 runs. According to specific uncontrolled environmental conditions. You then need enough time in cal to have this error happen there.


    While total cal time need not be as long as total test time, individual cal runs should be as long as individual test runs, and the amount of cal time you need depends on how intermittent the test run disturbances - which could be some unknown reaction, or some equipment error, are.


    When no intermittent behaviour is evidenced, sure, calibrate before, after, and occasionally in the middle if tests last a long time.


    THH

  • [The comparative time scale of the two events has nothing to do with it.]


    Well, I only partly agree. The longer the calibration runs the more the influence of ambient conditions (like the weather) becomes visible and can be taken into account.

    Yes. As I said: "You need to continue the calibration long enough to reach a stable temperature and stay there for a while." Ambient condition changes are the sort thing I had in mind. But I said the "comparative" time scales, meaning the duration of the calibration versus the active run. The recommended duration of the calibration depends on various factors, but the duration of the active test is not among them.

  • Suppose there is some intermittent (like, happens only on Mondays) change in the environmental conditions. Thus: lower water pressure means the cooling tubes develop air bubbles. Or, some metallic join connects or not. Or something. In the real world you can never be sure there are no such environment-determined unexpected errors.

    More bullshit, invented just now to cover up your previous bullshit. Yes, of course a calibration should continue long enough to detect problems. Yes, as I said, it should be conducted at different power levels, and it should be repeated before and after the test. But that is not what you said. You said the duration must be comparable to the duration of the test:


    "We'd need to know that the sequencing and length of the calibrations was comparable with the active runs."


    You made that up. It is an excuse to ignore the data. If you applied that standard, it would be an excuse to ignore every experiment on record, because no one ever calibrates for as long as an active experiment continues.