MFMP’s Ryan Hunt Comments on Feb 3 Dogbone Test

There’s a comment on the Martin Fleischmann Memorial Project’s Evernote site by Ryan Hunt, who was one of the team members conducting the test on the ‘dummy’ Dog Bone test yesterday. He comments about the emissivity measured in the testing:

” The main revelation was that the emissivity required for the camera to correctly interpret the temperatures on the surface was very close to .95. When we plugged in the emissivities cited from literature in the Lugano report (0.8 to 0.4), the apparent temperature was 1200 to 1500C at 900W in. Is our cast alumina significantly different than other alumina materials? ”

One part of his report in particular caught my attention. Ryan writes:

Meanwhile, here is some other interesting data.  The uptick in temperatures at the end looks interesting.  We saw something similar in the previous calibration on 2014-12-31, only we have extended it one more data point to 900W input.  We have no clue what to make of it.  We had no nickel or Hydrogen anywhere near the hot dog bone.  Any suggestions?  Could this be a change in material property like a thermal conduction change or radiant heat transmittance effect that could be misinterpreted?


This uptick in temperature will have to be looked at carefully, especially in comparison to what is measured in a fueled reactor to see whether ‘false positives’ could be being generated.

  • Andreas Moraitis

    Emissivity is a critical factor. Since the values from Lugano and from the DogBone differ significantly, the power/temperature relations cannot be directly compared. The energy that is emitted by radiation at the same temperature depends linearly on the emissivity (Lugano report, p.8). The higher the latter, the higher the radiated energy, and vice versa. But emissivity is also used as a variable for the estimation of temperatures by the Optris camera. Here, the influence goes in the opposite direction: If the emissivity is set to a lower value, the apparent temperature rises, and vice versa again. It would be necessary to know the internal algorithm of the camera system, or at least the exact measurement readings of the same temperature under different supposed emissivity values, in order to estimate the possible effect that a wrong value would have. Maybe it would even make sense to run a complete calibration sequence with alternating settings.

    • Alain Samoun

      “make sense to run a complete calibration sequence with alternating settings.”

      Yes it would make sens to calibrate the measured emissivity against a calibrated thermocouple puts inside the reactor at the place where the camera is looking at.

      • Ged

        That’s what they did yesterday, exactly, with help from the Williamson pyrometer. But seems the camera was killed when the computer shorted :(.

        • Alain Samoun

          Ah! I did not understand that the pyrometer was used in conjunction with a thermocouple inside the reactor? or was it outside?…It seems to me that they should be more interested by a correlation with the temperature inside…

          • Ged

            There is a b-type thermal couple inside, as well as another one outside. The pyrometer has to be outside, as it would never survive the temps, and it works by viewing from a distance. There is also a k-type thermal couple outside as well. So the camera and pyrometer looked at the spots the external thermal couples were at, while the internal tc monitored the center of the core.

            So, they had anticipated your good ideas and put them to practice. Part if what makes these replication efforts so solid; more so than Lugano itself, I’d say.

    • I have tried to do that.
      the camera as documented is an array of bolometers, measuring heat received in the wavelength range.

      I dis some estimation (excel…) of the integral of a grey body radiation and the result is that the camera measure a signal proportional to (T-220C)/230C, with 100% at 450C here

      the result is that is the emissibity does not plunge from 0.7 to 0.4 from 450 to 1200C as expected, but stay stable, then temperature is 800C instead of 1250 and 900W instead of 900C…

      the energie radiated should increase of more than 40% while power in increased by 12%.
      convection shoul increase proportionally to temperatur difference to ambiant… which is 12% about…

      if few heat is radiated, COP=1 is possible.
      if radiation is the main way to evacuate heat, (it grow with temperature, and calibration at 450 estimated both path for dissipation), then COP>1

      but COP>3 seems undefendable.
      one idea I have, if the problem is confirmed, is that the reactor was kept too cold to have the expected COP, but that Rossi and IH were happy with the measured performance and temperature that matched their own observation…
      I cannot imagine they are stupid enough to expect this kind of error, and not to fear that one of the tester put a black dot or a black fluid or even a thermocouple on the reactor.

      one thing that make me doubting of even that failure theory, is that there was a thermocouple inside the dogbone, expecting a temperature set by Rossi…
      If it was so different from what was expected, Rossi would have moaned.

      problem is the doubt.

  • Is the Ambient temperature being measured? Hard to see the display

    • Ged

      They are. You can see them on

  • Freethinker

    Are we nor mixing up emissivity from the IR imaging with the full spectrum values used in the energy computations?

    • Sanjeev

      Good question Freethinker. Do you mean that the Lugano report used full spectrum emissivity?

      • Freethinker

        Yes, in the radiant power computations, but in the thermal imager there is another emissivity, to correctly compute the temperature. I think they are two different things.

        • Sanjeev

          So how much is the difference between the two at 1400C ?
          Actually I can’t find this mentioned in the report, so please let me know your source.

          • Freethinker

            For the emissivity diff in IR vs full spectrum check my link above. I’d say some 0.5 or there about.

            No, it is not mentioned in the report, but it makes no sense that an emissivity constant that is applicable to the full spectral range of the black body at a given temperature should be used in calibrating an IR camera.

            On the same note, it does make no sense in computing the radiant power with Stefan Boltzmann with the emissivity in IR and not taken for the full spectral range, for a given temperature.

            Apparently emissivity is often rather high in many materials in IR, where as transitivity windows in other spectral ranges will drop the emissivity considerably there as emissivity+transitivity+reflectivity = 1 . Note also that the emissivity in IR for alumina is rather insensitive to temperature, but may still cause one or a few percent diff between 400 and 1400 C. See figure (from link above).

          • Freethinker

            You can also check:

            New Methods for Measuring the Thermal Emissivity of Semi-transparent and Opaque Materials By D. Demange, M. Bejet, and B. Dufour

          • Sanjeev

            Thanks for the graphs. Learned something new !

          • Sanjeev

            From your both graphs, it seems they took the IR emissivity. I guess most of the radiation would be IR and there will be very little power in visible or above. (Else there is no point in using an IR only camera). The testers many have done that to keep it simple, it will be cumbersome to calculate power in all bands.
            I’m pretty sure that they used the same e values for both camera and calculations. See page 15 and 16. The calculation shows same value (0.69) used there.

          • Freethinker

            Yes, page 15 is kind of indicating that they set the camera to such values. It is lower than what would be the mean in the the range [7.5,13] micron. If they did use those values for the 1400C run then there are no real problems, possibly a risk for a systematic 10% error, but there is no where where I can see that they explicitly state any camera set emissivity values for the active reactor.

          • Sanjeev

            They should have explicitly mentioned the emissivity values for active run. They should have also mentioned the wavelengths used. Or even better should have calibrated the dummy at 1400C and used same emissivities for both the dummy and active runs.

            They did not do it, its a shame. We can’t do anything about it now.

          • Freethinker

            True. Anyhow. About the emissivity, they clearly did do measurements that seemed quite self consistent with the dummy, and the dummy did have a COP 1. There was used a emissivity of about 0.7.

            The cat was baked in something, like the Bone was, and it could be that the surface characteristics with regard to emissivity is not straight of alumina, and that also the cat and the bone differs in this respect.

            But the emissivity is a weak spot in the Lugano report. If you play with those numbers it does quickly amount to a lot of temperature and power. Would be good if MFMP could get it going again, or perhaps try out the russian style cat experiment instead.

          • you make a very good point isn that emissivity may be wrong, but stable.
            the problem is that Lugano testers assumed a reduction from 0.7 to 0.4, which is probably wrong.

            note that about camera signal I just estimated boltzman radiation in the wavelenght window of the optris, and the signal is globally linear (not T^4 T^1/4) with temperature… to be precise some correction have to be done, but in that range the energy grows more or less linearly with temp…

            the result is that error in emissivity, translate into a linear error in temperature.
            the pivot temperature for me is around 220C… this mean that 10% error in emissivity is 10% error in the temperature difference with 220C (apparent 320C is 310C if emissivity is 10% higher than expected)

            my computation are rough, but the orders are probably right (I changed some hypothesis, integration step, and nothing changed much)

            this is enough to guess the real size of the catastrophe.

          • Freethinker

            There is mentioned in the report (similar to text in ref [2] in the report, ):

            “Every thermal camera contains a detector where sensitive components generate an electric signal proportional to the IR radiation received. This signal is then amplified and processed by the device’s electronics, and converted into an output signal proportional to the temperature of the object. This proportionality is expressed by an algorithm dependent on several parameters, such as the internal temperature of the detector (read directly by the camera sensors), ambient temperature, and the emissivity of the radiant body.”

            If we assume a gray body, then fair enough, they can be same, but the very calibration curve, plot 1, they use indicate that emissivity and transmission is wavelength dependent, as an increased temperature will shift the peak of the black body curve blue-ward. Hence it is not a good idea to use the overall (gray-body) emissivity in the IR camera.

            Note that the estimated temperature from the thermal imager is roughly inversely proportional to the n:th root of the emissivity (see , page 8, Optris document and ref [2] from the Lugano report) )

            where n varies from 2-17 depending on wavelength in the measurement wavelength interval. ( I assume it is a sensitivity adjustment, as it should typically be n=4. )

            1/4th root of emissivit=0.4 -> 1.25
            1/4th root of emissivity=0.9 -> 1.03

            In the Optris document there are error graphs on the impact wrong emissivity as well. I’d say from looking into that graph that a 10% error in emissivity would translate into 8% for 1400C if the LT curve is extrapolated.

            With all that said, the discussion about the emissivity in the Lugano report is weak, unclear, and the emissivity is not clearly reported for the active reactor part. My take is that the extensive discussion is to describe how these value has gone into computing the radiant power of various segments, and that infact the emissivity in that camera has adjusted itself or been left at a static level, as to make it viable measurements of the temperature, and that the graph in “plot 1” has been used only to compute the radiant power.

            I was kind of hoping that MFMP could shed some light on that. Pity the camera fried. Hope they can fix it somehow…

            And about my thermal rant above…. I may well be that I am delusional. If any of you out there can blow any holes in what I write, then please do. I wish to learn more.

  • Sanjeev

    This graph is from the Lugano report. The emissivity values used in Lugano test at 900’C is about 0.45, half of what was determined yesterday.
    In the report the temperature at 900W was 1400C. If you take the 0.95 value instead of 0.45, the temperature would be 900C and radiative power would be 1274 W. COP is now 1274/900 = 1.4 instead of 2454/906 = 2.7. We are still overunity but roughly half COP.

    (only radiative power values taken from the table 7 row 16 on page 22, so its a worst case estimate. Please double check my calculations).

    • this looks possible, and if you accept the claim of parkhomov that few happen before 1000C, then a tiny COP is possible.

      the most important to look is
      1- the calibration was done at 450C (with a TC?) and you can already calibrate the convection whichj is roughly linear with temperature gap with ambiant.
      2-then you can observe the step from 800 to 900W and check that the COP cannot be 1

      to bad calibration was not done at 1400C

      • Ivanidso

        I know that recently it would appear that 1000C is the magic number, but as I recall Rossi’s initial devices operated under 500C and maybe even under 200C???

    • Ged

      The other interesting difference is the MFMP saw 450 C at 200 W input, while Lugano saw 450 C with 300 W input at the same spot with the same camera.

      Your calculations look good to me, Sanjeev. A great observation!

      • Sanjeev

        I guess the materials are different then. And the comparison will not be one to one, like Ryan noted. Thanks for checking it.

    • Freethinker

      True, that curve is in the report and it is used to compute the emissivity for different section of the rod that is later used in the Stefan-Boltzmann formula to compute the radiant power for a given temperature.

      I submit to you that the emissivity needed by the thermal imager to measure the temperature correctly, is something different, as it operates in the 7.5-13 micron range (or so I surmise if it is to be similar to the Lugano instrument). See for the emissivity in that spectral range. It should be that emissivity that is important for the temperature measurement, and the plot you show, is for computing the radiant power, where the average emissivity for the black body for a given temperature and spectral range is needed.

      My 25 cent worth …

    • luigi

      with the correct emissivity mfmp saw an average temperature of a little more than 900 c at 900 w. i’m sure we agree that mfmp cannot have overunity in a dummy test, so the calculation indicating a cop > 1 with 900 c from the lugano report must be wrong somewhere. the dogbone is more or less the same dimensions as the hot cat, so same temperatures for both must mean same energy out unless convection is totally different and that seems unlikely. also you used .45 but the lugano chart says .39 for 1400 c.

      • note that since the reactor are different, 450W in one may heat it at 900C, or 450C…
        this is the role of the calibration.

        for Lugano, 450W was at 450C…
        for MFMP 900W at 900C…
        both result are solid because confirmed by thermocouple.

        I imagine that at 450W MFMP should be at another temperature than Lugano…

        • andrea.s

          MFMP at 450W was >600°C versus 430°C in the Lugano report (dummy run). This is one good argument for Levi and co., which makes the emissivity discussion quite controversial.

          MFMP gets 600°C at 450W, 900°C at 900W, extrapolating (or hopefully by test by adding the inner hot core consumption) they could perhaps have 1400°C at 2.7 W.

          Then the question remains, where does the 1.8W excess power come from (I won’t insist).

          • yes, either the surface is not the same, or the emissivity is different than imagined.

            here absolute temperature is 25% higher at same power.
            at this temperature of 600C, the emission should be 2.4x greater than E-cat, per unit of surface, a same emissivity.
            One possibility is that the surface of MFMP dogbone is either 2.4 smaller.
            or that emissivity is 2.4 times lower for the MFMP dogbone, but if the dogbone emissivity is 0.95, this is impossible
            and Lugano dummy emissivity was 0.7 at 450C… this makes MFMP dogbone at 0.3 at 600C
            I don’t swallow that …
            best hypothesis is surface. are the dimensions the same?

            or there is a mistake somewhere

      • Sanjeev

        You have a very good point.
        At 900W input, we get 900C (that’s 1173K) for MFMP test. The SB law gives output power of 1274 W for e=0.95. Its a dummy so the output cannot be more than 900W. So the e value is surely wrong. (Assuming the area of the tube is 12.5×10^-3, same as the ECat area)

        If you assume e=0.65, you will get less than 900W, so for MFMP test the reasonable e values should be less than 0.65. If you take conduction and convection into account the radiative power will be even less, which means e should be lesser, say 0.5.

        But somehow the optris camera demands e=0.95 to display correct temperature. I do not know why. I guess MFMP is still doing their calculations, and my estimate can be wrong, but as you say, the 0.95 value does not make much sense.

        I just took e=0.95 and T=900C to compare with the COPs given in Lugano report. Of course we do not know what the T value will be if there is fuel inside, It can be higher, but I took the lowest, 900C. I did not use 0.45 anywhere, I just mention it because the graph from the report shows 0.45 for 900C rather than 0.95

        So what is the conclusion ? I have no conclusion….

        • my hypothesis was that the surface was différent, making the MFMP dogbone probably smaller…

          if it is the same, then yor argument is better.
          the fact that you go back to what the Lugano team assumed is another sign of credibility of your position.

          I hope so, but there is doubt and that is evil.

          • Sanjeev

            There is a lot of doubt. Firstly, this calculation assumes the whole reactor to be at 900C, which it is not, only the middle part is at this temperature. The temperature goes down towards the ends. So overall power radiated will be less than estimated here.
            I’m feeling too lazy to do calculations for all 10 parts, so we will wait for the next experiment with the optris camera, when its repaired.

  • Observer

    Is MFMP taking into account the added surface area due to the fins?

  • artefact

    From MFMP on Facebook:

    “”Move fast and break things. Unless you are breaking stuff, you are not moving fast enough.” — Mark Zukerberg

    AAARRRGGHH!!! We did not intend to take this inspirational quote so
    seriously. The Optris thermal camera is dead. It apparently died in
    the same event that took out the broadcast computer. Once we got
    another computer set up, we found the camera would not talk. Then we
    plugged it into two other computers we had tested it with previously.
    No go. A mouse and at least one web cam seem to be dead as well.
    Needless to say, we are disappointed and this will change our plans even
    more. Stay tuned.”

    • Alain Samoun

      No fuse to protect the camera?

  • Anon2012_2014

    Reading the discussion below, I can’t help but think burning out the camera was a good thing.

    The camera, with its emissivity input, fourth power power calculation (T^4), plus convection, has just too many degrees of freedom to lock down the COP. I just don’t like it.

    I prefer Parkhomov’s simple boiling water in a pot approach. Easy to calibrate. Adjust the reactor temperature by putting the device in a insulating blanket of alumina in the water. The energy to heat and then boil away the water is easy to calibrate.

    Find a way to seal the tube and then replicate Parkhomov.