Mats Lewan's Test Report

  • @Mats Lewan now says that for him the most convincing test of Rossi's device was one that he himself measured on October 6 2011. Well - Mats certainly has better test methodology than Penon - the ERV!


    The report:
    https://animpossibleinvention.…st-of-e-cat-october-6.pdf



    My comments:


    Mats – I’ve read your report. I’d like to apologise. Your test methodology is pretty good and far far better than that of Penon – the ERV – who famously did not record any details of the equipment he used, even when its results disagreed with the readings from a kWh meter.


    I could nevertheless pick a few minor holes in the electrical measurements – but I’m not going to. A 100% error here is not significant.


    Here is how I think it was done. The subtle indicators from the report are as below:


    (1)


    "Water temperature at the input and the output of the secondary loop was measured with thermocouples attached on the metal connections at the heat exchanger where the hoses were attached (see video) and the difference in temperature was used together with the value of the waterflow to calculate the output power."


    The problem is that the TC temperature is not the water temperature. There is no reason to think the water heats up to equal the heat exchanger temperature – in fact at high flow rates in particular this is unlikely. The deltaT is small which means it is difficult for you to check in any other way what is the real output temperature of the water, although I guess a thermometer in the collected 1 litre buckets when the water is flowing fast would help. (You need several to ensure you are really measuring water that was flowing fast through the heat exchanger – I don’t assume speed stays constant).


    To get a COP of 10 (and maybe even believe it himself) all Rossi needs is a heat exchanger where the casing is hotter than the water inside.


    (2)
    "At start:Total AC current 142 mA. Over all AC voltage 229 V => 32W
    During SSM: Overall current 544 mA. Voltage 230 V. => 130W"


    We don’t know how the power changes during SSM of course, because you are measuring it at start and end with maybe one more measurement. But if we suppose these readings correct throughout there is at least 100W going into the device to keep it hot during SSM. (You might want to multiply that by 2 due to electrical issues which generally I’m not going to address here, because I don’t think that is the main issue).


    (3)


    “Water from a tap was fed to the secondary loop of the heat exchanger. The heated output water was led through a long hose to a well outside the building.”


    OK, so can I ask, when you measured water flow rate, where were you? The main issue here is that water flow rate could be reduced during SSM. The e-cat has a high heat capacity so it is the average flow rate that matters. Could this not have changed? Having the water output in such a different place from the main equipment makes it very easy for the flow rate to be increased when you are outside measuring. Even if Rossi goes with you, he can increase the flow rate before he leaves the building and decrease it after he returns.


    (5)Comments about steam. That is irrelevant. If steam is generated and condenses there is no heat required so the whole steam thing is a distraction which makes it seem as though this device does something useful. In reality it heats up water by a few degrees (or much less than that) depending on the water flow rate in the heat exchanger and possibly other variables you do not control. (It could be for example that the water flows mostly through parts of the H-E that are much colder than the part measured with TC).


    Conclusion:


    I can’t tell what went wrong in this test. there are too many possibles:


    (1) At least 100W appears to be going into the device when it is off. That could be significantly higher due to electrical issues. also, that could change when not being checked. Why was the input power so much higher at the test end (when this would help keep it hot during SSM) than the test start?


    (2) the test is designed so that you can only check the flow rate when out of the building…. enough said!


    (3) The determination of water temperature coming out of the device is not safe because the heat exchanger temperature will not equal the water temperature.


    Overall I think the SSM comes from power in during SSM and flow rate change. The device has very high specific heat capacity so would naturally stay hot, especially with a little power top-up. The COP when active comes from water temperature different from TC temperature and possibly flow rate change and possibly electrical measurement issues (relatively small factor).


    You see there is quite a choice. You might have additional info that could narrow it down.

  • This one test is what Mats considers the best evidence for Rossi's devices working, in spite of its being an "informal" test. It did have multiple observers, including Mats, and was a very impressive demo.



    Rossi Oct 6 2011 test


    Meta-analysis


    There are four technical analyses of the data from this test as shown in Table 1.


    Author DeltaT evidence Other evidence
    Higgins yes no
    Rothwell no yes
    Roberson yes, then no yes
    Ascoli no no


    Table 1
    The deltaT evidence from thermocouples on input and output of the secondary water flow is what should be used to determine device COP. If it is correct then it shows clear high COP from this device. In this case deltaT cannot be used because the experiment setup makes this impossible.


    Specifically, the water flow rate is set unnaturally high. At this rate a small deltaT corresponds to a large apparent output power. There are obvious error mechanisms which can generate a small deltaT without water temperature change. Therefore the error here makes using the mechanism impossible unless the observed deltaT is high.


    In this case the observed deltaT is very small, and within the range of likely errors.


    This error could easily be removed. Reducing the water flow rate until the deltaT increases significantly would both quantify this error and make it much less significant. Measuring water temperature directly away from the heat exchanger would make the measurement safe. Unfortunately that was not done.


    So the claim that this test provides evidence for the Rossi device working must come from other evidence. Before analysing this I want to say that it is easy to get this sort of indirect evidence wrong, which is why scientists tend not to use it. There is no evidence supported by more than one of the commentors above, even though each claim that the cited evidence proves beyond doubt that Rossi's device works. In each case I'm going to look at the arguments presented for the evidence.


    Rothwell evidence


    All of Rothwell's interpretations here of indirect evidence implicitly assume there is no "hot core". The model of a "hot core" with a loosely coupled external water circuit stabilised at 120C by a pressure vent explains everything without LENR. Rothwell does not reject this model - he just does not consider it. Once it is on the table it fits all the facts and is even supported by the known constructional details. I will look at each of Rothwell's points in turn.


    Sharp increases in T2 temperature during the "no power in" part of the test.


    Jed did not the possibility of a physical source of energy from hot cast iron "hot core". That provides the energy source for a sharp increase, and variation is expected due to turbulent flow, steam build-up, etc in the primary circuit together with water addition that alters the state of the system.


    The reactor surface remained hot


    This is expected by the hot core model.


    Room temp tap water was pumped into the reactor, enough water to fill the vessel twice.


    Using typical estimated values for the heat stored by a hot core the total amount of primary circuit water heated up and released by a vent is between 4 litres (dry steam) and 20 litres (wet steam). That seems within the range of what would be "filling the vessel twice".


    Without energy generation, with tap water being pumped in, the temperature would have fallen to room temperature within an hour. This was clearly show by the rapid temperature decline when the anomalous power stopped.


    The argument here supposes that there is only one internal temperature in the system. If there is a massive "hot core" loosely coupled to the primary circuit it is just wrong.


    Yet the vessel remained so hot that when a person accidentally touched the exposed metal 3 hours after the power was turned off, the person jumped back in pain from the heat.


    This shows that there is a large thermal heat content in the system which cools down slowly. It is predicted by the hot core model.



    Roberson Evidence


    Roberson looks in detail at the T2 temperature graph and infers from this that the only explanation is an LENR core with high variable power output. He put a lot of work into several versions of the analysis. Initially he accepted the deltaT evidence and used this, in the final version he agrees that the deltaT evidence is unsafe and cannot be used.


    I want to briefly discuss the second time constant and its implications. I propose that the electric heater is attached to the heat sink and somewhat insulated from the core modules. This conclusion can be drawn by analyzing the bump in the T2 curve that is maximized at around 18000 time stamp. This response stood out to me as strange when I was attempting to calculate the COP of the ECAT from the data set. This bump is obviously a result of the filtering of the final long power input pulse that occurs just prior to entering self-sustaining mode. You should notice that it has entirely been dissipated within a short period of time compared to the long time constant associated with the core insulation. It is very clear that power inputted to the heating resistor is subjected to the heat sink cooling. Heat energy within the heat sink is able to rapidly conduct to the water within the ECAT enclosure. This offers proof for the skeptics that Jed Rothwell and I are correct in our assertions that the fact that heat continues to be produced for hours at a high level is proof of LENR activity.


    Roberson's argument here is not clear so I'll elucidate it. The issue is that if the heater was attached to a hot core and separated from the primary circuit by insulation its switching could not give a "dissipated with a short time" short time constant. On the other hand, if the heater is attached to the primary circuit heatsink as assumed by Roberson then it cannot boost a "hot core" to a temperature well above 120C as is necessary for the "hot core" model to work.


    Roberson does not consider the very likely alternative. The heater is thermally isolated from both the hot core and the primary circuit heatsink. The effect of the heater switching on or off is therefore much faster than the hot core temperature change, but equally the heater can get up to high temperatures and so slowly heat up the hot core.


    There is further evidence to support this supposition. The final curve beginning at 30000 time stamp proves this quite well. Note that the temperature of T2 falls like a proverbial rock beginning shortly after the hydrogen is released from the core region.

    This fall is not visible on Roberson's own graph. But if it were, it does not prove anything.


    There is a short period after the LENR activity has ceased and built in delays are satisfied. Within approximately 800 seconds, the decay begins at a rate similar to that seen due to the second time constant which establishes the conduction rate for heat stored within the heat sink. Review the falling edge of the pulse waveform around 19000 time stamp to see a similar decay rate.


    I agree these two decays have roughly the same slope. The final decay is clearly exponential and discontinuous from the preceding curve. The 19,000s decay is part of a "bump" and much more complex to analyse, it does not look like a simple exponential, so to assume they have the same cause is not safe, or even likely. The interaction between the heater and the primary circuit and its vent can create the 19000s waveform. From Roberson's graph the final 30,000s fall is 12C in 2000 seconds. I agree with roberson it looks exponential. The "half-life" time constant looks like 4.5 hours. I agree with Roberson's interpretation here, and it is explained exactly by the hot core model, which Roberson cannot consider because he assumes no hot core can exist. The half-life is what is needed for such a model to work.


    Roberson's analysis here is clearly not necessary because the hot core model explains these observations. It is also not good because to fit the data he must introduce unexplained arbitrary "built-in delays".


    I see absolute proof of LENR activity by pursuing this line of reasoning.


    Based on this analysis I must disagree. Roberson's key mistake here is dismissing the possibility of a hot core model due to his not considering a heating element loosely coupled to both the hot core and the primary system. Therefore he dismisses the possibility of a hot core, and has to invoke LENR to explain the rest of the data.


    Roberson's final argument is very indirect and very speculative. Even if you could not see how to explain the features in the T2 graph he notes it is a big ask to go from them to certainty of LENR. However his analysis is helpful in that the same features are what you'd expect from a hot core model.

  • @ Thomas Clarke,
    in another thread (*), you wrote:


    - "All Rossi's devices have a heater. In this case, to explain what was observed, the heating element needs to heat both a "hot core" with a large thermal mass and the primary water circuit. In order to explain accurately the rise and decay times of temperature in different parts of the experiment you need the heater not to be connected so tightly to either the primary circuit or the core that they have identical temperature. That "not identical temperature" is loose coupling."


    Perhaps we need to better explain the details of the processes of generation, accumulation and heat transfer within the fat-cat, as can be inferred by common sense from the information available for the test held on October 6, 2011.


    First, it is appropriate to define what is the "core". Referring to the diagram at the top left of the jpeg (1), the "core" is the set of elements placed inside the inner box (R). The "core" therefore includes the electric heater and the massive shielding (S), which accumulates the heat. About the path of the heat, it is generated in the resistor, which transfers it to the shielding, which in turn transfers it to the inner box casing, and hence to the cooling water. So all these elements are arranged in series, and heat goes through them all.


    The graph at the top right of the jpeg (1) shows the computed temperature trends for all the elements considered in the model. They don't include the electric heater, because it was assumed that the heat it generates was absorbed directly by the shielding (S). This is equivalent to consider a strong thermal coupling between the resistor and the shielding. We do not know actually how much these two elements of the "core" are thermally coupled each other, but we know that the most delicate element of the device is the filament of the electric resistor, whose temperature must not exceed a certain safety level, otherwise it breaks.


    It is also to be noted, at this point, that it is precisely the need of not exceeding the limit temperature of the resistor that determines the weird trend of the electrical power, shown in the graph on the upper left of the jpeg (2). The stepwise power increase at the beginning of the heating period and the on-off cycling toward the end can only be explained by this need. In particular, the most critical period goes from 14:00 to 15:00, when the shielding had already reached high values of T, but the inner box was only partially submerged by the coolant. Probably, the activation of the elecrìtric heating was controlled, automatically or manually, according to data provided by a thermocouple placed near the resistor, in the hottest part of the "core".


    With regard to the temperature difference between the casing (R) and the water, it is driven initially by convective mechanisms of heat exchange and subsequently, after the boiling point is approached, by evaporation mechanisms that determine a high delta T across the boundary layer on the surface of the casing.


    (*) Rossi: “Steam Was Superheated” in 1MW Plant Test

    (1) http://i.imgur.com/FTstZqe.jpg
    For an explanation in English of the images in (1) please see here: https://animpossibleinvention.…i-ih-affair/#comment-4465

    (2) http://i.imgur.com/nGaK4jz.jpg
    For an explanation in English of the images in (2) please see here:
    https://animpossibleinvention.…i-ih-affair/#comment-4883

  • Quote

    Roberson's final argument is very indirect and very speculative. Even if you could not see how to explain the features in the T2 graph he notes it is a big ask to go from them to certainty of LENR. However his analysis is helpful in that the same features are what you'd expect from a hot core model.


    Thomas, all your arguments are also indirect and speculative. But it does not prevent you from repeating them ad nauseam.

  • Quote

    Thomas, all your arguments are also indirect and speculative. But it does not prevent you from repeating them ad nauseam.


    (1) My technical arguments, as here, are neither indirect nor speculative


    (2) As for other arguments you will find that my repetitions are nearly all replies to contrary speculative arguments that seem to me to be questionable. You may feel no posts here would be preferable - but if there are posts, you will get speculation.

  • @ Thomas Clarke,
    I see that the discussion on the technical aspects of the October 6 test is continuing on the "Rossi-“Steam-Was-Superheated”-in-1MW-Plant-Test" thread. I know that you have well understood the rationale of the "hot core model" and you are responding adequately to all the objections that are raised.


    However, let me to add, here on this thread, my opinion about some of these objections, hoping that it can be useful to anyone, who is looking for an explanation of what could have really happened.


    IH Fanboy wrote (*): "your standard "hot core" model does not explain the exponential temperature drops observed in response to certain actions taken (which were various and occurred at different times)."
    and then he added (**): "not sure how it explains the temperature response when the hydrogen was let out of the cell."


    Probably he refers to this comment of Rothwell on Vortex (1): "At 15:50 the power is cut off. If there had been no source of anomalous heat, the power would have fallen off rapidly and monotonically, at the same rate it did after 19:55. It would have approached the zero line by 17:25. Actually, it would have approached zero before that, based on Newton's law of cooling."


    Rothwell rightly cited the Newton's law of cooling, but he applied it to the wrong element: the cooling water, the only element of the fat-cat, of which the values of the measured temperature were disclosed. In reality, he should have applied this law to the inner "hot core", whose temperature, as shown above in the upper right diagram of the first jpeg, starts dropping immediately after the cutting off of the electric power, in accordance to the Newton's law.


    The temperature of the water remains around 120°C for 3.5 more hours, because it receives the heat lost by the "hot core". Its value remains about constant due to the relief valve, which keep the internal pressure at about 2 bar. So there is no need to assume a loosely thermal coupling between the core and the water to give reason of their different trends during the so called SSM period. As shown in the cited diagram, the curves of these two temperatures begin to drop in reciprocal accordance after the internal pressure drops below the valve setting point. It happens shortly after 19:08, when it is reported by Lewan that: "Hydrogen pressure was eliminated. Flow from peristaltic pump increased. All electric power switched off".


    From this moment onwards, the water temperature drops, as pointed out by Rothwell, in accordance to the Newton's law, just because it follows the exponential decreasing trend of all the other more internal components.


    In conclusion, we can be sure that after 15:50, when the electric heating was shut off for the last time, no other heating source was active in that device.


    (*) Rossi: “Steam Was Superheated” in 1MW Plant Test
    (**) Rossi: “Steam Was Superheated” in 1MW Plant Test
    (1) "http://www.mail-archive.com/vortex-[email protected]/msg52546.html"

  • @IH fanboy, Ascoli65,


    If I may, one may simply say that as long as the exchangers A and B are still hotter than the boiling temperature of water at 2 bar, this will remain at 120°C, regardless of the temperature slope of A and B. Afterwards, it will follow that slope.


    The mass budget and thermal capacities assumed by Ascoli65 are all very reasonable. Let me add that that they do not need to be exact. We have a closed box with some data: if there exists a good explanation with known physics for its behavior, one should believe this explanation as much more likely than a dream of unexplained nuclear reactions. The burden is on the one who claims the results to rule out the "normal physics" explanation.

  • @ Thomas Clarke,
    I see that the discussion on the technical aspects of the October 6 test is continuing on the &quot;<i>Rossi-“Steam-Was-Superheated”-in-1MW-Plant-Test</i>&quot; thread. I know that you have well understood the rationale of the &quot;hot core model&quot; and you…


    Thank you for the additional analysis. But you did not address the temperature response when hydrogen was let out of the cell. Now, I am having trouble finding that bit of information, but I do remember quite clearly that those who witnessed the test were intrigued at how the temperature fell quickly when the hydrogen was removed.


    http://www.nyteknik.se/energi/…ces-proof-of-heat-6419717


    That is a link to the report, but unfortunately, the links within that report to the detailed version of the report and data appear to be broken.


    Do you happen to know where that data might be, and do you remember the discussion regarding the temperature response based on the extraction of hydrogen from the cell? Does your model account for that?

  • Rothwell rightly cited the Newton's law of cooling, but he applied it to the wrong element: the cooling water, the only element of the fat-cat, of which the values of the measured temperature were disclosed. In reality, he should have applied this law to the inner "hot core", whose temperature, as shown above in the upper right diagram of the first jpeg, starts dropping immediately after the cutting off of the electric power, in accordance to the Newton's law.


    This paragraph is confusing. You state that the only element of the fat-cat of which values of the measured temperature were disclosed was the water. But yet, you graph the temperature of the core? Am I missing something?

  • @ IH Fanboy
    The links at the bottom of the Ny.teknik article still work.


    Concerning correlation to switch-off of H2, don't you think the demo was rehearsed several times before?


    The core temperature I suppose is computed given the input power, the assumed thermal capacities, thermal conductivities and convection. Lots of assumptions but if they are credible, the claims fall until the assumptions are ruled out by evidence.

  • @ IH FanBoy, you wrote:


    - "But you did not address the temperature response when hydrogen was let out of the cell."


    That the "hydrogen was let out of the cell" is written in the Lewan's report (1), where he writes: "19:08 Hydrogen pressure was eliminated. etc.". Presumably, this is what has been told him, but I personally have no reason to believe that there was any hydrogen inside the fat-cat.


    Anyway, you get the answer to your question just reversing the order of the causality of the two cited events: the drop (below a pre-established level) of the hot core temperature did cause the announcement of the end of the test, and hence of the hydrogen purging, if any.


    - "Do you happen to know where that data might be, and do you remember the discussion regarding the temperature response based on the extraction of hydrogen from the cell?"


    New links to the report and to the measured data of the October 6 test have been recently posted by Lewan on his blog (2).


    About the discussion on the hydrogen issues, I remember only the Rothwell's mail on Vortex, already cited on my previous comment.


    - "Does your model account for that?"


    My model doesn't account of the hydrogen, because it ignores any alleged heating from any source other than the electric resistor. But anyway, as explained above, the model accounts in some way of the contemporaneity between the announcement of the hydrogen removal and the decreasing of the coolant temperature. The contemporaneity is a simmetric property, it also holds in reverse.


    - "You state that the only element of the fat-cat of which values of the measured temperature were disclosed was the water. But yet, you graph the temperature of the core? Am I missing something?"


    The layout in the high left corner of the above first jpeg shows the only 3 temperatures (T1 - T2 - T3) which have been continuously monitored and recorded during the test. The corresponding curves, marked with (mis=measured), are shown, with different scales, in both the diagrams on the right side. All the other curves come from the numerical model. The red curve T2(mis) shows the only temperature measured inside the fat-cat. Its trend is well reproduced by the corresponding computed trend Tw, the green curve. This correspondence provides a good confidence that the numerical model reproduces adequately the real evolution of the other temperatures.


    (1) https://animpossibleinvention.…st-of-e-cat-october-6.pdf
    (2) https://animpossibleinvention.…s-on-the-rossi-ih-affair/

  • Thanks for the additional explanations. I am impressed that someone took it upon himself to do this analysis.


    The core temperature I suppose is computed given the input power, the assumed thermal capacities, thermal conductivities and convection. Lots of assumptions but if they are credible, the claims fall until the assumptions are ruled out by evidence.


    I'd like to ask an honest question and hope to get an honest reply. There are a number of assumptions that must be made for your model to fit, as you readily admit. Did you curve-fit the assumptions to the T2 curve? Or did you build the model, look at your calculated T2 curve, and essentially say, "by golly they match." ?

  • IH Fanboy


    It is Ascoli's model (and a clever one) not mine, but it is easy to answer.
    It is called reverse engineering. You define a model, set fixed and variable parameters. Graph results with a spreadsheet or math software e.g. matlab, and you tune or optimize the variables unril the graph fits the measurements. If all parameters are physically realizable and consistent with data: bingo, you get a water boiler design with the same performance as the fatcat.

  • It is Ascoli's model (and a clever one) not mine, but it is easy to answer.
    It is called reverse engineering. You define a model, set fixed and variable parameters. Graph results with a spreadsheet or math software e.g. matlab, and you tune or optimize the variables unril the graph fits the measurements. If all parameters are physically realizable and consistent with data: bingo, you get a water boiler design with the same performance as the fatcat.


    Ah yes, meant for Ascoli. Would like him to reply as well.


    The reverse-engineering approach (i.e., curve-fitting) is certainly one way to build a model in an attempt to explain a phenomena. It certainly isn't the only way.

  • @ Thomas Clarke,
    I see you decided to leave this forum permanently. I'm sorry, because I appreciated the great expertise you have shown and the genuine passion you spent in the debate.


    I thank you for the attention you've reserved to the issue and to the model that we are discussing on this thread, which was opened by you.


    I hope there will be opportunities to still share some other technical discussion with you, somewhere else on the web.


    Ciao.

  • @ IH FanBoy, you wrote:


    - "I'd like to ask an honest question and hope to get an honest reply. There are a number of assumptions that must be made for your model to fit, as you readily admit. Did you curve-fit the assumptions to the T2 curve? Or did you build the model, look at your calculated T2 curve, and essentially say, "by golly they match." ?"


    The first way, of course!


    I did adjust part of the input parameters in order to fit most of the known experimental data, in particular the measured T2(mis) trend. The rationale of this way of doing have been already explained by andrea.s (grazie, Andrea). The aim is not to reproduce exactly the fat-cat tested on October 6, but to find out a configuration and a set of alleged characteristics which are reasonable and which comply with most of the experimental evidences.


    The purpose of the model is not to "demonstrate" that there has been no excess heat. For this purpose, the best demonstration is the one proposed recently by Lewan, provided that the value of 16 kg of vaporized water be updated to the correct value of 6 kg (1). The purpose of the model is just to give reason for most the single phenomena that have been reported, as for instance the water temperature drop after the alleged shutdown of the reactor.


    To this respect, it's worth noting that the most meaningful phase for the accordance between the measured and the computed curves of the cooling water it's not the intermediate one, during which they share the common constant temperature of about 120°C, because in the numerical model this value is easily obtained by imposing a relief value of 2 bar (abs) to the outlet valve. Instead, the most significant aspects of the accordance between the two curves are the heating phase, whose slope has been optimized by fitting properly some input parameters, and, especially, the starting time and slope of the cooling phase, which came out quite spontaneously from the optimized calculation.


    (1) Rossi: “Steam Was Superheated” in 1MW Plant Test


  • I thank you for your honest answer. If you are familiar with currency trading, you might be aware of the neural networks that attempt to curve-fit historical trading data in attempts to discover patterns in trading, so that they can possibly predict future trading behavior among a particular currency pair. The success rate is dismal. Curve-fitting in almost all cases, doesn't work. I have my reservations about your model because of the assumptions and the method used to build it.


    I also think Jed makes an insightful point:


    JedRothwell wrote:


    I believe I based that assertion on the specific heat of iron, which is
    one-tenth that of water. Even if the inside of the Rossi device were
    heated to incandescence, it could not hold much heat compared to the
    water surrounding it. When a blacksmith takes a heavy piece of iron,
    heats it to incandescence, and then quenches it in water, very little
    water boils away. The metal instantly cools.


    Can you comment as to Jed's insight?

  • I'll pop in a quick reply to that.
    The specific heat by volume of iron is about 1/3 of that of water. (We can figure out how heavy later).
    A small air gap, maybe 1 mm, between the iron poker and the water, (maybe a sleeve of metal, held in place by a very thin pins or small insulators) would allow the poker to remain hot for quite a while while submerged. (The heat being radiated to the sleeve from the iron ).
    And yes, this mostly affects the time constant, not the overall total heat delivery, since water has for (almost all practical purposes) the highest heat capacity of anything.
    A fairly significant lump of incandescent iron surrounded by an air-gapped sleeve would be required to keep boiling water for any significant period of time. And swapping in lead, tungsten or whatever would make very little difference compared to iron. The volume of the water in the heating area would be the most important variable.


    Edit: wrapping the heater wire tightly around the iron rod (with very thin electrical insulation) could be a way to get the rod much hotter than the sleeve when in powered mode.

  • A small air gap, maybe 1 mm, between the iron poker and the water, (maybe a sleeve of metal, held in place by a very thin pins or small insulators) would allow the poker to remain hot for quite a while while submerged. (The heat being radiated to the sleeve from the iron ).


    Can you show (using formulas or math) how the 1 mm air gap would affect the time constant of heat transfer from the 30cm x 30cm x 30cm wafer to the water? Would be interesting to run some numbers and see how long this could keep the water boiling. Could it maintain the boiling for the 3+ hours?


    Edit: Weight of solid steel wafer at 30cm^3 = 0.52 pounds. So, even with an air-gap sleeve, it would fall far short of being able to boil even a liter of water from stored heat alone.


    Here is a handy tool: http://www.aqua-calc.com/calculate/volume-to-weight


    Would there be any reason to include such an air-gap sleeve aside from outright trickery? In other words, if one believes this conjecture about an air-gap sleeve, then you would also have to believe that Rossi's demonstration was fraudulent (as opposed to a mistake in measurement or delusion).

  • IH Fnaboy I havent read what Ascoli65 have calculated, but you are in risk mixing two different things in your thinking. What comes to fitting stock price history on trying to predict future, does not work since opposite to early days of stock exchanges, number of variables has grown, sudden one time interference factors (9/11, wars sub prime crisis etc) simply cannot be predicted from history. Also robot traders (operating at ms speeds) are something that you cannot predict from history because they exaggerates even small interferences because of snowball effect caused by speed competition).


    Curve fitting in physical processes instead is more predictable and is daily routine in process industry. You may do black box testing on unknown process (cause change on input and measure response curve on output). There can be multiple different serial time constants inside black box, which makes exact calculations more difficult.


    But for this you don't need even that. It is enough to calculate how much mass at max. you could hide inside E-Cat and calculate its thermal capacity supposing it gets heated say 1000C. That is what Paradimonia above explains. Boiling water takes lot of energy compared to temp increase of same amount of metal mass. Air gap is just parameter to fit to make correct boiling speed, it has nothing to do with capacity.


    I checked my earlier estimations:
    - Heat capacity of steel is 0.466 J/g/K (=kJ/kg/K) half compared to aluminum but fits in smaller volume Source Wikipedia
    - Heat of waporization of water is 2257 kj/kg (see Wikipedia)


    So in other words, to boil each 1 l of water you need to add 4.8kg of steel losing 1000K of heat (so steel lump temp must be > 1100 to keep 1 l water boiling totally. End temp must be >100C )


    That is only a safe calculation, in practise it has to overcome temp leaks and heating first incoming water ( 4.1813 kJ/kg/K) depending on test arrangements (which I have not read).


    These calculations does not take account possible pressure tweaking (=suck vacuum and water boils even in room temp). But as I said in playground these old tests and sayings are not interesting because of so big other reasons in play.

  • But as I said in playground these old tests and sayings are not interesting because of so big other reasons in play.


    No doubt, bigger developments in play now. Thanks for your analysis. The Oct. 6, 2011 test always intrigued me though, and still does to this day. In many ways, this is the test that kicked off the Rossi mania. I think that is why it still gets attention.

  • @IH Fanboy,
    For a large amount of water, the amount of iron (or whatever) would have to be very substantial to keep water boiling for three hours.
    The geometry with the fins is complex. Not so much the fins themselves, but for how the heat gets sent to the base of the fins.
    It could be modeled, but with necessarily lots of guesses in the parameters.


    Aside from trickery, an air gap would insulate a reactor from being held close to the temperature of the water or steam, and possibly allow better control of the heater (and reactor) itself. The gap could allow precise heater temperatures to be achieved without the lower temperature water, water flow, and bubbles constantly and possibly chaotically changing the heater temperature, making control very complicated. By forcing radiant heat to jump a gap, the temperature of the core is limited to the capacity for radiant power, no matter what else is going on around the core assembly. (At least until the sleeve begins to approach the core temperature). This leaves the water flow to control the water-steam temperature, which is simpler.


    A hunch is that without any reaction, the air-gapped incandescent heater would be slower than direct conduction per L of water, so the size of it would have to be greatly increased compared heating the same volume of water with a simple immersed coil heater in the same period of time.
    Indeed, a good dummy for this test would have been an immersed coil heater powered the same way, and using the same amount of water.

  • @Paradigmnoia


    I agree on the good reasons to insulate the core, not for trickery but to reach a desired temperature.


    As for specific heat per volume:


    Steel density 7.8 kg/l
    Steel specific heat 0.47 kJ/kg °C
    Water density 1 kg/l
    Water specific heat 4.2 kJ/kg °C


    Heat capacity of 1 liter = 1 dm^3 of water : 4.2 kJ/°C
    Heat capacity of same volume of steel: 0.47×7.8 = 3.7 kJ/°C


    Not that far.



    8 dm^3 of steel are roughly 60kg (out of the 98kg). With an insulated sleeve 60kg of steel heat up say to 1000°C=1300K with 30MJ.
    Then there are infinite combinations of surface , emissivity, thermal conductivities from core to sleeve and exchanger that can delay this 30 MJ heat release to match the 3 hours self-sustaining.

  • Steel density 7.8 kg/l
    Steel specific heat 0.47 kJ/kg °C
    Water density 1 kg/l
    Water specific heat 4.2 kJ/kg °C


    With you.



    Heat capacity of 1 liter = 1 dm^3 of water : 4.2 kJ/°C
    Heat capacity of same volume of steel: 0.47×7.8 = 3.7 kJ/°C


    Not that far.


    Yes, those two values are pretty close, and representative of 1 liter of water.



    8 dm^3 of steel are roughly 60kg (out of the 98kg). With an insulated sleeve 60kg of steel heat up say to 1000°C=1300K with 30MJ.
    Then there are infinite combinations of surface , emissivity, thermal conductivities from core to sleeve and exchanger that can delay this 30 MJ heat release to match the 3 hours self-sustaining.


    Wait a second. You base 8 dm^3 of steel from what?

  • I just looked quickly at molar specific heats, which for metals pretty much hang around 24-28 J/mol K, where water is 75.2 J/ mol K.
    From what Jed is quoted as saying, I'm guessing he looked at mass specific heat, where iron is ~0.45 J/ g K and water 4.2 J/ g K.


    I hadn't bothered to do the conversion to a L volume with iron, although clearly I should have. I am surprised they are so close.

  • I don't think the hot core model looks very likely even though it seems that the core (under certain assumptions) could be charged with a large amount of energy (30MJ). I would expect the heat transfer across the air gap to be highly dependent on the delta temperature between the water (100C) and the hot core, and the core must sureley decrease its temperature over time when it delivers heat. I would then expect to see the delta T of the secondary circuit to become significantly smaller at the end of the SSM, but that does not seem to be the case.

  • @ IH Fanboy, you wrote (*):

    - "I thank you for your honest answer."


    You welcome. Anyway, I have no reason to answer otherwise.


    - "If you are familiar with currency trading, you might be aware of the neural networks ... The success rate is dismal. Curve-fitting in almost all cases, doesn't work."


    I'm not familiar with, but I can imagine that it is not so easy to predict the future in the financial world. Curve-fitting the hot-cat behavior is much easier.


    - "I have my reservations about your model because of the assumptions and the method used to build it."


    As I already told you, the purpose of the model was just getting an idea of how the things could have gone. You can build thousands of different models, with different assumptions and methods. It depends from what you are looking for. If you want a quick answer on the real COP obtained in the October 6 test, the best way is the "corrected Lewan calculation" that I mentioned before: input 31 MJ, output 25.5 MJ, COP<1.


    If you want also to have an idea about where a part of the 31 MJ introduced in the first part of the test have been stored in order to keep the water at the boiling point for 3.5 more hours, you can give a look at the upper right diagram in the second jpeg shown on the initial page of this thread. As you can see, the inner iron mass, 40 kg out of a total weight of 98 kg, can store 11 MJ (gray line), when it reaches the maximum average temperature of 560 °C (see the first jpeg). This stored energy (more than 1/3 of the total input energy) is about equivalent to the energy accumulated in the water inside the pool (green line), and is sufficient to compensate the energy brought away by the exiting fluid (red line) up to the end of the so called SSM period, and even beyond.


    -The above diagram answers also to the first Rothwell's point: "I believe I based that assertion on the specific heat of iron, which is one-tenth that of water. Even if the inside of the Rossi device were heated to incandescence, it could not hold much heat compared to the water surrounding it."


    A iron mass of 40 kg (=2x the water mass) at T=560°C (=5x the water delta T) holds the same heat quantity of the surrounding water, contrary to what Rothwell claimed.


    - About the second Rothwell's point: "When a blacksmith takes a heavy piece of iron, heats it to incandescence, and then quenches it in water, very little water boils away. The metal instantly cools."


    I have no direct experience in this field, but anyway I could say:
    - first, I doubt that a blacksmith can easily handle a massive piece of iron of 40 kg;
    - second, this iron mass is hypothesized to be inside the inner box made by steel, and hence his surface is not directly in contact with the cooling water;
    - third, only a small part of the 26 kg of water pumped inside the fat-can could be evaporated, being the volume of the internal pool at least 20 liters.


    (*) Mats Lewan's Test Report


  • Wait a second. You base 8 dm^3 of steel from what?


    Just an example compatibile with 30MJ stored at a emperature in the range of 1000 deg, and 20×20×20 cm fits well in the 30×30×30 cm of the object described by Lewan as a "heat exchanger". But it is just a starting point for a plausibile model. Ascoli's numbers are more deeply thought since he fits both transients and steady state.