Mizuno reports increased excess heat

  • Mizuno refers to "we" in the report, so that means he is not working alone on this.

    No, it is just him. Although another person in Japan replicated the experiment. He lives a good distance away.


    I wrote the paper. I used "we" throughout because that is the correct style for a paper with two or more authors. (Okay, I guess I am more bound to academic traditions than I let on, in my jeremiad against Nature.)


    Mizuno asked me to be the co-author. It is a great honor. My main contribution to the actual work has been to kibitz and ask dumb questions. Dumb questions can sometimes be helpful.

  • Having read both pdf in the first post of this thread I react as follows:


    Almost idiot-proof recipe.

    Even more detail about tooling and procedures would help.


    Example: Buy this here, buy that here.


    I admit the papers make very precise parts of the shopping list.


    Thank you very much for sharing this Jed. Please give my regards to Dr. Mizuno and thank him for his untiring efforts.


    Admirable persistence inspires us all.

  • Congratulations on your work Jed (and congratulations to Sir Mizuno), I hope it is replicated elsewhere and cracks the can open.


    My only concern is that this is once again an experimental set-up which relies on a different calorimetry method, which risks getting everything bogged down into endless measurement discussions. Say what you want about the Nature/Google article, they identified clearly that this is one of the reasons the field is so fractured, and that the community should first agree on how to measure results prior to worry about how to produce results. Without a consensus on measurements, it is all too easy for sceptics to dismiss results that do not mesh with their own prejudices as simple errors.



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

    Jed; if the first paper is new - then is it a rewrite of Mizuno's earlier paper - which I read? Or is it describing new results? My memory is pretty hazy - so I apologise if i'm misunderstanding this.


    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?


    It is not a matter of belief. It is a matter of clarity, as you should well know. Teasing out what experimental results mean, and what effects could cause something unusual, requires calm, lack of personalisation or blame, and precision. Verbal or anecdotal statements do not help with any of these things. If you have a record of the entire methodology, in detail, then an exact deacription of the experiments, timeline, calibration tests, etc, would be immensely helpful.


  • Jed, I assume this is related to rubbing with Palladium. It might be good to emphasise that.
    Rubbing for cleaning wil decrease the weight, so this might be confusing.



  • Jed; you are quite right in asking me for this.


    1. Let me point out that if these extraordinary and if correct world-changing results (I say that with some hope that they are real, although not as yet much expectation) are correct you might want to help the world by drawing the widest possible attention to them. One part of that is writing them up as precisely as possible. In the process of doing that (I have found personally) often the person doing the work becomes clearer about it.
    2. I'm not demanding that this is done. It is entirely Mizuno's decision. I'm just surprised here that my helpfully laying out what would make things more precise, some of which maybe could be done, here seems to be disliked. I have the best of motives, and also no wish to impose my ideas on others. If you wish me to stop doing this please say immediately and I'll happily lay off this particular report.
    3. One merit of this report (not as a scientific write-up, but as a technical resource) is that it does specify so precisely the apparatus and materials - making replication much easier. Maybe the best route to publicising this work, rather than write it up more precisely, is for somone to replicate it. In that case what I'm suggesting here is still useful for whomever does the replication.
    4. Precision can always be improved, and is always helpful. there is a trade-off. In this case, due to the extraordinary nature of the results, a lot of precision is expected and helpful. If you come at this from the standpoint that these results are not extraordinary, then much less is needed. That may be what LENR researchers think appropriate, but it will not be so for most scientists who will find the proposition that way above chemical energies can be measured in this apparatus surprising. So decide who is your audience, but accept that if you don't treat the results as extraordinary you should not expect non-LENR converts to pay much attention.


    I did summarise what I think would help strengthen this in my first post. I don't think this has changed much, but I'll comment with a bit more context and detail.


    For this calorimeter:


    • Raw data from which the results are derived: temp in, out, air speed, air cross-section, air speed profile, exact geometry over which these are measured
    • Details of blower power vs voltage and back pressure, blower spec, Blower voltage in experiment. (Useful as sanity check for airflow measurements)
    • Description of precisely how calibration is done, and of calibration methodology. Ideally replacing the reactor by an identical object (with same D2 connections) without nickel gauze heated electrically
    • Clear description of the previously done tests of air speed, temperature constancy, quantitative estimate of errors due to these.
    • Clear description (true in this case, I believe) that D2 source could not be heating the device by burning D2. Clear methodology pointing out why total amount of D2 is known/limited and much less than that which could result in energy generated.


    Let us go through this list:


    Raw data

    For the tabulated results it would be good to know under what conditions they were measured. This data must exist. Thus here: what is the air velocity (might change with different input power). Whether the tabulated values are raw or converted in some systematic way does not matter, as long as any such conversion is made precise. Given air velocity, input temperature,and the equation used, output power and output temperature are equivalent in the sense that one can be converted systematically to the other, but the exact conversion equation should be given. Exceptionally, given these results are so extraordinary, it might be helpful to give both output temperature and inferred power, as well as velocity, but I realise that would not normally be done.


    Velocity cross-section details (rest of 1st two bullets)

    The issue is that the first principles (ab initio) power calculation - an excellent characteristic of this setup - requires total airflow to be measured. Flow velocity in a pipe is Ok for this, when not turbulent, but you need to bound the error due to the flow slowing to zero near the pipe edges. You can do this with the given measurements. (1) note from flatness of middle portion of velocity profile that flow is laminar - turbulent flow in a pipe would not have a flat velocity profile across middle 3/5ths. (2) Use known equations for laminar flow to estimate error due friction with pipe walls and therefore slowdown near walls. (2a calculate Reynolds Number as sanity check that indeed flow is laminar). (3) Note how slowdown near walls (calculated theoretically from Reynolds number) scales with air speed and temperature. (4) From this get a secure proof, given the velocity measurements in the middle 3cm of the tube, that the overall flow is say 90% of that it would be if the velocity was constant across the tube (e.g. that would bound ab initio error in a 10% band). Check this against calibration to see what that means - calibration will mostly tell you heat loss of course, so can't be used to determine this, but it is still helpful. Check this against the stated airflow from the pump manufacturer (as a sanity test).


    Is this needed? No. If you view the calorimeter as needing calibration with no first principles analysis you can leave this out. But, in that case, why bother to measure the airflow speed at all?


    Is it helpful? Absolutely. When results are extraordinary every possible cross-check is helpful. Look at the FTL neutrino paper for an example of how anyone wishing to publish such results goes about it - ok they rushed to publish too soon, but they did at least try to cross-check the results internally as much as possible.


    Calibration methodology

    Calibration is absolutely a good way to ensure that these results are robust. But you need to be clear what measurements have been done under what conditions. For example, the procedure describe din the previous paper; is that carried out for each measurement? Or for a set of data points? If so, which? A lot of detail here will convince a skeptical reader that the calibration is valid and that conditions are the same as for the active test.


    Is this needed? not if the first principles calculation is fully done and robust

    Is it helpful? Yes, and indeed I think it would be weird if not done. Of course it was done here. So what is needed is making sure that it is fully documented that it is done well.


    D2 burning cross-check

    This is for nit-picking skeptics, who should be your preferred audience. Make clear how you know that the total amount of D2 (all of which might be assumed to burn) is limited to much less than that giving the observed energy out. I'm guessing from the setup that this is easy to do: e.g. a (low) pressure is established, after which there is no D2 connection. D2 loaded onto nickel can be bounded.


    Otherwise you have to argue that there is no way the D2 can burn because no oxygen present. That is probably true but a weaker argument, and should not be needed if the results are correct.


    Error bounds

    Basically, for every raw datum, and every inferred value from the raw data, state the error bounds and how these are established. It is always work doing this, but very helpful work, and good practice. It is not alas always done, but many would say it always should be done. Staker 2018 does quite a bit of this. In particular make sure to include error bounds with argument, for inference steps like deriving airflow from velocity measurement where the relationship cannot be precise. (A theoretical calculation for example will depend on the roughness of the tube walls).


    Precision

    In the R19 table of results, are the given heat outputs those measured from first principles, or those inferred based on calibration data that determines heat loss? Preferably give both, the results are strong enough that both would perhaps be interesting.

  • Congratulations on your work Jed (and congratulations to Sir Mizuno), I hope it is replicated elsewhere and cracks the can open.


    My only concern is that this is once again an experimental set-up which relies on a different calorimetry method, which risks getting everything bogged down into endless measurement discussions. Say what you want about the Nature/Google article, they identified clearly that this is one of the reasons the field is so fractured, and that the community should first agree on how to measure results prior to worry about how to produce results. Without a consensus on measurements, it is all too easy for sceptics to dismiss results that do not mesh with their own prejudices as simple errors.


    LCC - I actually disagree with this. Having robust positive results from different measurement setups strengthens the results because it is more difficult to imagine common systematic errors.


    In this case, as a variant of mass flow calorimetry, the method has some good characteristics but there are potential errors and those need very careful checking, as always.


    Skeptics will dismiss results if they can find possible more likely mechanisms for the data. Since skeptics view LENR as pretty unlikely, whichever method you choose you need cross-checking so there are no other feasible mechanisms.


    Repeating results with the same setup is only good if the repetition can be done with stronger instrumentation based on the defects of the original. That is very helpful.


  • That's reassuring to me if you feel that way! I hope it is the way it pans out. I see your point but I still think that designing experiments around calorimetry designed by "mainstream" scientists could be a way to avoid criticism on that specific point.

  • But, in that case, why bother to measure the airflow speed at all?


    The basic equation used in 2017 for calculating thermal energy output .from the AirCalorimeter

    boiled down to

    AirC Energy out = flowratex area x heat capacity x density x tempdiff (air in -air out) x time

    or AirC Power out =flowratex area x heat capacity x density x tempdiff (air in -air out).

    see Figure at right= Figure 6 adjusted

    which shows how the tempdiff of 10C is converted to Power of ~180W


    The airflowrate was based on the fan power/velocity calibration given in Figure 9.

    The airflow speeds were initially measured to validate this calibration

    After this only the fan power only was used for routine measurements of velocity.

    It varied very little during the course of one test

    as can be seen from the 2017 data Blower voltage/current columns in the link below.

    https://docs.google.com/spread…EMipY/edit#gid=1856684673


    In 2019 the new heat recovery calibrations account for the convective/radiative loss

    from the AIrC.box which increases at higher temperaturesas seen in Figure 11.

    These losses vary with box/duct material and dimensions

    If replicators build an airC box/duct/fan setup they will need to make their

    own caibrations. for 1(the fan) 2(the box)

    .

    The effort and time involved in one replication might justify an initial visit to

    Mizuno's installation. or discussion with him.before doing it

    Of course one can put more detail into these two publications

    with supplementary files.. but they are going to become quite large

    There are at least 100 large excel files.


    Jed said on Japanese researcher already done a replication

    I presume he visited Mizuno??

  • RB


    Thanks for that. I was aware of all that you have posted, except the specifics of the flow calculations.


    The actual flow in this case is measured (as in Figure 9) in a 5cm diameter tube, area (pi/4)*d^2 cm^2 = 1.96/1000 m^2. The figure you provide (4.4/1000 m^2) is not this so I'm not sure where you or Mizuno gets it from.


    The question however is how much of this cross-sectional area has air at the speed shown. From Figure 9, as I've said twice before on this thread, the inner 3/5 of the tube must be so, from the tests. The outer 2/5 has not been tested. To take an obvious over-estimate of the error here, if the untested part all had low velocity, then the average velocity integrated over area would be only 36% of the measured velocity.


    Obviously calibration can help resolve this, but will be dependent on power and airflow, which is why I was keen to get specifics.


    This is why the 2019 results do not mean much to me without more information. I could speculate that the airflow is constant velocity across say 90% of the duct area, but I would have no evidence for that and I know you like comments here to be founded in fact.

  • This is why the 2019 results do not mean much to me without more information.


    They mean a lot to other people who are involved in active LENR research.


    Of course I know that you are burning

    to carry out you own replication

    but the information that is there now I feel is sufficient for


    legitmate researchers to carry out their own research.


    If I can sit at my armchair for two hours to figure out

    what' s going on... then they can to..

    and they will try to get at least 10C tempdiff

    or 180w thermaloutput.


    I so like that photo of the 3Kw heater in Sapporo in

    2018.


    I thought that Mizuno was stuck on COP 2 since 2017.


    COP 10 wow!


    その驚くべき!

  • A flat velocity profile signifies turbulent flow...


    (1) note from flatness of middle portion of velocity profile that flow is laminar - turbulent flow in a pipe would not have a flat velocity profile across middle 3/5ths.


    There some kind of academic arrogance on display here Huxley? As pointed out recently, this is entirely wrong... Completely ass-backward in fact: Turbulent flow in a pipe always has a flat velocity profile (apart from in a negligibly thin boundary layer).


    Don't believe me? Just check any fluids textbook ever written. Or Google 'turbulent flow velocity profile', and have a look at the pictures.


    Air has very limited viscosity - hence laminar flow breaks down easily.


    Does it surprise you that, even at the lowest air velocity Mizuno measures (2m/s in a 50mm pipe), flow would be turbulent? That's because you didn't bother to calculate the Reynolds number (~6800*), instead preferring to repeat several seemingly-authoritative, but entirely wrong, assumptions about a topic you appear to know little about.


    * i.e. well into the turbulent regime.

  • There some kind of academic arrogance on display

    Professorial wisdom "turbulent flow in a pipe would not have a flat velocity profile"


    Wow Zeus .. you actually read THH.. well done:)


    There was turbulent flow in the Mizuno aircalorimeter

    which made the velocity and flow measurement so much simpler.


  • Jed,


    I'm glad that this work has been done. or half-done. My point is that in the write-up you either do the work, or you don't. if you don't then why bother to give the speed measurements?


    Frankly I'd not expect a rough RN calculation here to take very long, and it would then allow the inference you draw to be made. I always forget this stuff (and was wrong previously about laminar vs turbulent). Looking more deeply:


    Re = D V rho/mu


    D = pipe diameter, V = air velocity (max).


    Both rho (density) and mu (dynamic viscosity) depend on temperature, so the value varies, but at 25C for a 5cm pipe Re ~ 10,000.


    Thus the mizumo case is typically turbulent flow, but only 4X larger than the boundary.


    At that value (from thermopedia):


    The velocity profile for turbulent flow is fuller than for the laminar flow (Figure 1), whereas a relationship between the average and axial velocities ū/u0 depends on the Re number, being about 0.8 at Re = 104 and increasing as Re rises.


    So we have a decent estimate of volumetric flow = v * A * 0.8 (80% derating of naive equal velocity value).


    Error bounds for this would need as always more work. At higher temperatures nu increases and rho decreases, so Re goes down and we get nearer the laminar/turbulent boundary with average velocity lower relative to centre of pipe velocity.


    So we can say: calibration will definitely depend on air speed and temperature. For given power dissipated, reducing speed will increase temperature and reduce Re because V, rho, nu all move in direction of lower Re (and therefore more derating of actual power output relative to that calculated).


    You can see why more detailed investigation of these issues is helpful? If we had the missing info from the data we could check this further with more precise quantitative values using calculators etc.


    THH


    PS - another quick google-mediated check:


    Harikrishna M Menon (HK), M. Tech Mechanical Engineering, Dr. B. R. Ambedkar National Institute of Technology, Jalandhar (2020)
    Answered Sep 13, 2016


    (Vmax/ Vavg) = 1+ 1.33 √f

    And f = 64/Re

    Re = (Rho x v x d) / μ


    so the derating factor is 1/(1 + 1.33*sqrt(64/Re))


    Substituting values of Re. The actual value here is between 12,000 and 5000 I think:

    10,000 0.9

    5,000 0.8


    Not quite consistent with the Thermopedia comment so more work is needed to determine which is correct. Unfortunately these derating factors are significant, so more work would be needed to do this precisely.


    Where Re (as above) =

    (Rho x v x d) / μ


    d = 0.05 (SI units)

    v = (typical, but varies) 3.5 (SI units)

    Rho, mu vary with temperature as can be found from tables.


  • Robert. Please stop personalising:


    (1) Stop calling me an academic. When have i claimed this? And what purpose does this serve? Nor have I claimed any specific expertise in any of these areas (except poss electrical stuff). I am just capable of reading stuff anfd have like you I'd guess enough maths to learn new things from resources. In fact I stay here because I so much enjoy learning new things and doing little random physical question calculations like this fluid flow stuff which I first got into when trying to analyse Rossi's claimed heat exchanger (it does not work, and his "expert" used an approximate coefficient value well outside its domain of application).


    (2) I posted (it took me a while because lots of attempting to find calculators etc) an update on the velocity profile issue. It is quite true I got laminar/turbulent the wrong way round. I knew nothing about Re before a while ago here when investigating Rossi heat exchangers. I'd always avoided flow calculations when young as messy, therefore knew nothing of them, in fact they are quite beautiful. Luckily I'm capable, when pushed, to spend time doing this because I'm curious and no-one else here bothers, and to look up more precise calcs and therefore correct errors.


    (3) I expected this would not be an issue (on calculation velocity profile v flat, as you say). Plugging in numbers (you can check, and should) it seems not. There is a significant derating and more importantly this derating factor varies with velocity, temperature. This is useful info because it means the extra precision about under what conditions calibration is done is indeed relevant. Although the derating is significant it is not likely that high, compared with +50%. But it could be 80% derating factor or maybe more - I have posted inconsistent info from different sources above - which means it needs to be considered properly.


    None of this is a problem for the validity of the results, I'd guess. I am only guessing, because without dotting is and crossing ts here I can't be sure. Neither could anyone reading the paper, which is why my set of "want more details" questions is relevant, and why it is always best to be explicit about assumptions made.


    Regards,

    A somewhat feeling I should not bother here THH

  • I made some minor changes to the paper and uploaded a new version. Biggest change, p. 12:


    Rubbing is done with a palladium rod, 100 mm long, diameter 5.0 mm, 99.95% purity. Before rubbing the mesh, weigh it with a precision scale. Then vigorously rub the entire surface, left and right and up and down. Turn the mesh over and rub the other side. Weigh the mesh again. Continue until the weight increases by ~50 mg.



    Some people say the images are blurry. I will check the Acrobat settings.

  • Why not immerse the cylinder in a bath of water and measure the electric input energy used to heat water between temperatures?


    The heat capacity of water is 4,18 KJ/kgK and would be very simple and accurate method, actually conservative:


    If the water bath is not insulated any excess energy calculated would be a Conservative Value, i.e. Actual excess will be higher.


    So this method can easily be used when we have large excess values as claimed here.