MFMP: Next replication - GlowStick 5.4 aka GlowShell

  • The Martin Fleischmann Memorial Project announced a new replication approach by Alan Goldwater:


    *GlowStick* 5.4 AKA *GlowShell* undergoing initial calibration tests...


    Looks like it may be able to reach target temperatures with half the input power of previous *GlowStick*


    Alan Goldwater adds "The enclosures are made from whole kiln bricks,
    which were easy to cut and drill with ordinary hand tools. I measured
    around 60 °C on the tops with 400 watts input, so they insulate really
    well."



  • This is a great idea. The key thing is that it can help disambiguate excess power from calorimetry errors. Given that the reactor and its conditions are identical we should see the same levels of excess heat - if we see significantly lower that points to artifact. At the same levels, because input powers (with which artifacts scale) are lower, and with careful experiment design, that might be above likely artifact levels. Though you would if possible want at least a 4X increase in insulation for this happy situation given the current results are ~ 10%.

  • The simple truth is that if we get anywhere close to the "Rossi Effect" the excess heat will be huge and massive. It will be undeniable. Actually, they should be able to cut off the input power and see self sustain.


    This is tricky. We generally expect that LENR heat will incerase with temperature. So if a device is running at "self-sustain" power, it can easily run away. So Rossi, we understood, kept his devices below true self-sustain, to avoid runaway. This then created a limit on COP. How to move around that?


    Well, in a word, cooling. By controlling cooling, one should be able to run above the "self-sustain" temperature.


    Consider an approach with a destructive reactor, for demonstration purposes. The reaction is taken up to a temperature where it is showing XP. Then an insulating heat reflector is snapped around it, so that the assembly becomes a bomb calorimeter, and at the same time, the input power is shut off. Be careful, it actually could be a bomb! Almost all heat will be retained. If there is significant XP, the temperature should increase with no power input until the thing burns out. If the excess heat is artifact, it should slowly cool down based on whatever heat leakage remains. From the rate of temperature increase and reactor conditions, the XP could be estimated, i.e., increasing until meltdown.


    One could also run this with the "bomb" in place, simply slowly adding heat electrically, then shutting the heat down the moment that temperature increase shows up. There are details to consider, such as chemical reaction heat, so one might want to do the first part with significant cooling.

  • Is this only a test for checking out the new insulation or is it a full-blown experiment?


    The pictures are from a test run, actually before calibration. It was done merely to evaluate the effect of the K23 bricks on the input power to temperature. Alan will next run a true calibration run that will include measurement of the core temperature.


    Adding the insulation increases the sensitivity in temperature to a given excess heat. For example, if 100W of XP would have produced a delta-temperature of 20C without the insulation, it may take an XP of only 50W to produce the same delta-temperature with the insulating bricks. The bricks increase the sensitivity to XP ... but they also lower the threshold to thermal runaway. However, since only small XP has been seen before, thermal runaway is not expected.

  • Quote

    We generally expect that LENR heat will incerase with temperature.


    Just to be a D.A. -


    I agree we do expect this, on experimental grounds. Theoretically it is also what we would expect if LENR effects are artifactual, where the calorimetry artifacts scale roughly with deltaT.


    Otherwise there is no obvious reason for this scaling, although of course it is easy to speculate on things that would imply this. Without the test of strong quantitative prediction such speculation has a wide range.

  • (I wrote:)


    Just to be a D.A. -


    I agree we do expect this, on experimental grounds. Theoretically it is also what we would expect if LENR effects are artifactual, where the calorimetry artifacts scale roughly with deltaT.


    It's an experimental observation, THH. Yes. Some artifacts would scale this way, with delta T, specifically with heating of the electrolyte by electrolytic current. This is one reason why I'm interested in isoperibolic calorimetry where the "bath" or constant temperature reservoir is at an elevated temperature. This is simply raising the environmental temperature, it is not actually power input to the experiment (properly handled,of course). With high insulation, that environmental constant temperature bath or surround can also have substantially reduced power.


    What is more to the point as to possible artifact is the well-known correlation between excess heat and current density. SRI P13/P14, one of the most striking displays of excess heat, crystal clear, has that obvious "explanation." Why, of course, the current was increased, so some error in measuring XP would increase with current. There are two problems with that idea: P13 was a light water control, running in electrical series, so exposed to the same current. P14 was with heavy water. The hydrogen control showed no XP, only an increase in calorimetry noise at higher current, i.e., expected. The deuterium cell showed the clear phenomenon. And then, what was not normally stated (cold fusion research has often been very poorly presented), that displayed run (this is in the 2004 DoE review paper) was the third time with that current excursion, The cell had been operated for many hours, loading and deloading. The first two times, there was no significant excess heat in either cell. I've called this a display of the "chimera of cold fuison." The chimera walked into the lab, licked McKubre in the face, then sauntered out. Okay, maybe some connection went loose. Maybe something. Garwin: They must be making some mistake."


    That's what heat/helium is about. The heat-only arguments, even as supplemented with hydrogen controls, are circumstantial, with each example, it's possible to claim that there might have been some error. This gets much more difficult when helium is correlated with the heat.


    Quote

    Otherwise there is no obvious reason for this scaling, although of course it is easy to speculate on things that would imply this. Without the test of strong quantitative prediction such speculation has a wide range.

    Sure. Strong quantitative prediction is be highly desirable. There are, however, fairly obvious reasons, in various theories:


    Hagelstein Phonon Theory: vacancy rates increase with temperature.
    Storms crack theory: full supply rate to the cracks increases with temperature.
    Takahashi TSC theory: dual dueterium molecule occupancy of a cavity is required. There is a potential barrier against that, increased temperature increases the number of such occupations.


    Letts found temperature variation corresponding with temperature variation in vacancy rate. However, that's relatively weak. Maybe. Maybe it also correlates with crack formation.


    However, the most interesting recent finding, my opinion, is that Storms ran a cell with elevated temperature, using a heater in the electrolyte, separate from the electrolytic current. When XP developed, he shut down the electrolysis, but XP continued. If this can be confirmed -- it hasn't been, and assuming his calorimetry is good -- loading will decline when the current is shut off. As it did not decline, he may have shown that loading ratio is not a critical factor in maintaining the reaction, it only looked that way because loading stressed the palladium, generating cracks. The presence of NAE and temperature (and fuel, but not necessarily such highly loaded fuel) are. Yes, it's obvious to suspect that he didn't actually have XP in the first place, but this is the standard "there must be something wrong" objection. He's an expert. So .... I'm hoping that confirmation/disconfirmation is attempted.


    Notice that the common idea that PdD is punk for commercial applications is based on the idea that it would need to be electrochemistry. Not necessarily! Key would be forming stable NAE. Storms is claiming that his cathodes can be stored and start right up instead of needing to go through extensive conditioning again. (And then he backed off a little, saying it needed more confirmation). If Storms is right, then readily reproducible experiments could be designed, using already-conditioned and confirmed cathodes.


    As to precise predictions, here is one: in electrrochemical PdD experiments, FP class, for each atom of helium generated, 23.8 MeV of energy is released as heat. It's precise, and it is verifiable. And that's the work under way, to investigate this with increased precision. Cool, eh?


    By the way, the above was about PdD. Very little is known about NiH reactions, by comparison, but it is not difficult to think that they might be similar.

  • Well, in a word, cooling. By controlling cooling, one should be able to run above the "self-sustain" temperature.


    Yes - it is entirely possible to operate above the critically-stable self-sustaining XP point with the use of active cooling (air flow, etc). The only problem is that the active cooling ruins the calorimetry. However, having a high XP instability occur would be a happy circumstance and it would be accounted for in subsequent experiments. I would like to think this is what happened to Rossi in the early days - he had some good success with XP and it caused meltdowns, causing him to dial back the nominal COP to about 6 by changing the designed heat removal which allowed control with simple thermostat and heater.

  • @ Roger Barker


    Please define "proper calorimetry", keeping in mind that the cells we are testing typically run at 1000°C. The team at HUG built some very nice calorimeters for testing Celani-type cells, which run at lower temperature and power density. Both air- and liquid-flow isoperibolic systems were built, reaching accuracy better than 100 mW if I remember correctly.


    Following those experiments, Mathieu Valat built an elegant concentric-tube calorimeter for a Celani cell, and only needs a lab space to set it up. That will hopefully be in place soon, so he can begin testing the system. Bob Higgins also started construction of a water-bath calorimeter for a Glowstick, but he was not happy with the design and has put it on hold for now, pending redesign.


    Details of these projects were published on the MFMP web site QuantumHeat.org, and should still be available in the archives there.


    If you can propose a calorimeter design for a Glowstick-type reactor cell at reasonable cost, we'd be happy to consider building it. We are self-funded, with some support from crowd donations, so cost is a critical factor. Re-usability, accuracy, and ease of assembly and calibration are also important of course.

  • Abd,


    Quote

    Hagelstein Phonon Theory: vacancy rates increase with temperature.
    Storms crack theory: full supply rate to the cracks increases with temperature.
    Takahashi TSC theory: dual dueterium molecule occupancy of a cavity is required. There is a potential barrier against that, increased temperature increases the number of such occupations.


    Letts found temperature variation corresponding with temperature variation in vacancy rate. However, that's relatively weak. Maybe. Maybe it also correlates with crack formation.


    However, the most interesting recent finding, my opinion, is that Storms ran a cell with elevated temperature, using a heater in the electrolyte, separate from the electrolytic current. When XP developed, he shut down the electrolysis, but XP continued. If this can be confirmed -- it hasn't been, and assuming his calorimetry is good -- loading will decline when the current is shut off. As it did not decline, he may have shown that loading ratio is not a critical factor in maintaining the reaction, it only looked that way because loading stressed the palladium, generating cracks. The presence of NAE and temperature (and fuel, but not necessarily such highly loaded fuel) are. Yes, it's obvious to suspect that he didn't actually have XP in the first place, but this is the standard "there must be something wrong" objection. He's an expert. So .... I'm hoping that confirmation/disconfirmation is attempted.


    Non-LENR literature provides evidence from SEM images that extremely small bubbles can form in the lattice, and a common area for them to form are within microns from the surface of a bulk sample of metal or close to the surface below cracks/fissures produced by hydrogen embrittlement (the penetration of the lattice with atomic hydrogen). As always, I don't claim to be a scientist or an expert in this field, but my hunch is that once you achieve the optimum combination of pressure/temperature in these bubbles (which may or may not form exotic hydrogen species) LENR begins. Once it starts, as long as certain critical parameters are maintained the excess heat will continue regardless of any additional hydrogen loading. And it makes sense that if the fuel is quenched -- locking the hydrogen in these bubbles so it is unable to quickly diffuse out over ordinary time frames -- the excess heat should return upon heating and stimulating the same sample of fuel.

  • Abd,



    What's interesting is that both Rossi's earliest reactors before opening the JONP (the ones with copper and palladium hydrogen spillover additives) and those after opening the JONP (which he claims contained no precious metals) often utilized an innermost reactor chamber made of copper which melts at 1085C. These reactors were both alleged to self sustain and in at least a few tests, such as the 18 hour test observed by Dr. Levi, self sustained for 18 hours (if you don't count the power required for the control box). This would seem to imply that self sustain can take place below the temperature range at which stainless steel or alumina reactors would rapidly fail. I also think in Rossi's earliest Italian patent applications that he lists a preferred temperature range of up to 500C, but I'd have to go check on that.


    You are correct that boosting the output of a reactor to the point that self sustain can take place could increase the risk of a runaway reaction. This would especially be true if the excess heat starts being produced at a high temperature (1200C or so) which is already close to the upper long term tolerances of the materials used. And considering the corrosiveness of lithium vapor, these high temperatures degrade materials extraordinarily fast.


    I think that if certain parameters are optimized (maybe the quantity of hydrogen filled bubbles/cavities/defects in the lattice) or the level of pressure inside them, that excess could potentially be triggered at lower temperatures. Hypothetically, lets imagine a situation in which the contraction of the nickel lattice just below the curie temperature stimulated LENR during a triggering routine. A reactor producing robust excess heat at that temperature (capable of self sustaining with adequate thermal insulation) would have LOT more headroom than one that only starts producing excess heat at 1200C. What would be fortunate is if a runaway did start, the headroom would give replicators enough room to start removing some excess heat. One idea I like is devising a system in which the insulating bodies can be moved farther away from the reactor, allowing more heat to radiate away. Andrea Rossi even stated on the JONP once that he was testing the concept.


    There may be another possibility to control a runaway. In the past, Rossi said repeatedly that the need for a backup generator in case of a power failure was critical to safety. He implied that if a runaway took place, the "drive" could be used as a safety mechanism. A hunch of mine is that especially at high temperatures, the reactions taking place in the hydrogen bubbles (my mind visualizes them as spheres filled with glowing plasma) become more susceptible to magnetic fields. Certain frequencies could possibly accelerate the reactions (dirty high voltage AC has been talked about frequently) but perhaps others could have an inhibitory effect.

  • You are correct that boosting the output of a reactor to the point that self sustain can take place could increase the risk of a runaway reaction. This would especially be true if the excess heat starts being produced at a high temperature (1200C or so) which is already close to the upper long term tolerances of the materials used. And considering the corrosiveness of lithium vapor, these high temperatures degrade materials extraordinarily fast.


    In the original Rossi design, he was cooling with water, across a thermal barrier. If he was evaporating all the water, as claimed, he had no control from cooling, he could not increase it. His control, then, was entirely through heating. Many skeptics laughed at that, but it made sense. If he kept the temperature below self-sustain, removing the heating would shut the reaction down. However, if he did evaporate all the coolant, the reactor temperature would increase because cooling by water would then decrease. I concluded that he could not be evaporating all the water. He must have overflow. And then if he had overflow, he was producing less heat than claimed. Maybe a lot less.


    However, if he doesn't evaporate all the water, he could control cooling by regulating water flow. Still, that would be limited. It's important to realize that if the reaction rate increases with temperature, and if self-sustain is reached without an ability to increase cooling rate, it *will* run away, until the temperature reaches what will shut it down, one way or another. That design required not reaching self-sustain. It should be easy to reach self-sustain temperature, the problem is keeping it controlled, then. In furnaces, the fuel (mixed with oxygen) is above self-sustain (ignition) temperature, but the reaction is regulated by controlling fuel supply. Let's suppose that fuel supply cannot be regulated, it's fixed.


    To run self-sustain, the cooling must be designed to handle the heat and maintain temperature elevated above self-sustain temperature. I see two approaches: a normal cooling facility dependent upon water flow (with heat transfer across that thermal barrier, so that equilibrium is hot enough for self-sustain, and with reserve cooling available by increasing flow, however that is designed, plus an emergency cooling system with a plug that melts when the system gets too hot. (And that's replaceable, like a fuse.)


    All this engineering is premature when the basic reaction itself is not well-studied and characterized. Rossi kept everything important secret, and still does. And none of what he shows can be trusted. He can't even be trusted to be fake! So basing independent study on the "Rossi effect," imagining what that is, is not the most powerful approach. There are persistent low-level effects that might be the file drawer effect, and they might be real, and careful study will distinguish. I highly encourage this study. Done well, it's real science. Done poorly, let's say, it's a learning opportunity, if one persists and learns.


    Quote

    I think that if certain parameters are optimized (maybe the quantity of hydrogen filled bubbles/cavities/defects in the lattice) or the level of pressure inside them, that excess could potentially be triggered at lower temperatures.


    Yes. It should be possible. The effect would be expected to be smaller, though. It is important in this work to move away from the idea that Bigger is Better. Ultimately, for practical applications, yes, a few milliwatts would be meaningless. But if the calorimetry is precise, if one can reliably generate a few milliwatts, it could even be conclusive under some conditions. Context matters.


    I would rather be a small frog in a small pond, of clean water, than a big frog sitting on a desert island in the middle of an ocean, surrounded by undrinkable water.


    Quote

    Hypothetically, lets imagine a situation in which the contraction of the nickel lattice just below the curie temperature stimulated LENR during a triggering routine. A reactor producing robust excess heat at that temperature (capable of self sustaining with adequate thermal insulation) would have LOT more headroom than one that only starts producing excess heat at 1200C. What would be fortunate is if a runaway did start, the headroom would give replicators enough room to start removing some excess heat. One idea I like is devising a system in which the insulating bodies can be moved farther away from the reactor, allowing more heat to radiate away. Andrea Rossi even stated on the JONP once that he was testing the concept.


    A hundred degrees of headroom should be enough. It's an engineering problem, but for self-sustain, controlled cooling is necessary, unless there is some sort of fuel control or the like. Rossi's original control setup depended on staying below self-sustain. He could then control with heat. Remove the heating, the reaction would shut down.


    In fact, consider the recent Storms work. He maintained the electrolyte in his PdD experiment at an elevated temperature (obviously below boiling, this wasn't seriously pressurized). When he shut down the electrolytic current, which was maintaining deuterium loading in the palladium, the reaction heat did not decline. Only when he allowed the electrolyte to cool did the XP go down. This was a small amount of reactive material, and the reaction was not strong enough to boil away the electrolyte. So he could cool it (and shut it down) by turning off that electrolyte heating. That's how it appears at this point, unconfirmed.


    With various means, one could then imagine running at higher temperature under pressure, perhaps to the point where PdD could be used for supplemental heating. Palladium is still probably too expensive, but with gas-loading, viability does become possible; the palladium is not consumed. Almost certainly, the engineering would go toward gas-loading, which could run at very high temperatures, in theory.


    Quote

    There may be another possibility to control a runaway. In the past, Rossi said repeatedly that the need for a backup generator in case of a power failure was critical to safety. He implied that if a runaway took place, the "drive" could be used as a safety mechanism. A hunch of mine is that especially at high temperatures, the reactions taking place in the hydrogen bubbles (my mind visualizes them as spheres filled with glowing plasma) become more susceptible to magnetic fields. Certain frequencies could possibly accelerate the reactions (dirty high voltage AC has been talked about frequently) but perhaps others could have an inhibitory effect.

    This gets too far into speculation about reaction mechanism for my taste. A backup system that depends on power is unsafe. It's easy enough to design fail-safe systems.


    If the reactor is designed such that runaway simply melts the fuel, shutting down the reaction, and all that happens is that the fuel capsule must be recycled, that could be enough.

  • In the original Rossi design, he was cooling with water, across a thermal barrier. If he was evaporating all the water, as claimed, he had no control from cooling, he could not increase it.


    Well, you can control cooling by increasing the flow of water. That's how a steam engine works. But I think Rossi's experiments all employed a steady flow of water, making them strange flow calorimeters. I do not think this is how flow calorimeters are supposed to work. If I were doing this test, I guess I would use a tank of cooling water. I would make sure the steam is dry. You can do this varying the flow of water, or with a steam trap, which separates steam and water. You drain the water back into the tank. I would measure the entire mass of vaporized water, by weighing the tank.

  • Abd Ul-Rahman Lomax wrote:


    Well, you can control cooling by increasing the flow of water. That's how a steam engine works. But I think Rossi's experiments all employed a steady flow of water, making them strange flow calorimeters. I do not think this is how flow calorimeters are supposed to work. If I were doing this test, I guess I would use a tank of cooling water. I would make sure the steam is dry. You can do this varying the flow of water, or with a steam trap, which separates steam and water. You drain the water back into the tank. I would measure the entire mass of vaporized water, by weighing the tank.


    They were strange. This was all obvious by 2011. The best calorimetry done would not evaporate any of the water, it would involve measuring flow and temperature rise accurately. (There can still be errors in both of these, but those errors can be eliminated with caution and particularly with control experiments. This completely avoids the diffiiculties of measuring steam quality. Yes, if water is being evaporated, there are still ways of studying the heat flow, as mentioned. Using a water tank can make measurement of water usage not dependent on possibly flawed meters or pump setting. (Or could confirm these). If the true quantity of steam generated is measured, as described, this is first-principle calorimetry. Without that full measurement, it can be highly misleading.

  • @ Roger Barker


    Please define "proper calorimetry", keeping in mind that the cells we are testing typically run at 1000°C. The team at HUG built some very nice calorimeters for testing Celani-type cells, which run at lower temperature and power density.…


    Hi, I am no expert on calorimetry but Mary Yugo is. She knows all about calorimetry. We've used them often in our fact finding sessions.


    Mary, can you kindly share your thoughts on using calorimetry for MFMP.


  • I don't know that Mary actually recommended this calorimeter. Maybe. It's first-principle. It's said here "without even performing calibrations" except that what was done was a calibration! I.e., known input power. Yes, within certain boundaries, this calorimeter could be expected to have accuracy like that.


    However, it would probably not have been usable with actual test devices, because the heat transfer rate to the cooling fluid would probably be too high. It would need to be used with an insulating barrier to allow the "reactors" to reach operating temperature without requiring much higher input power that could then burn out the heating elements. The Lugano approach, with the IR camera, would have worked well enough if calibrated. (Which would have made all the complicated calculations unnecessary.)


    Notice that if the cooling fluid rises above the boiling point, some of it boiling, the calorimeter becomes far more complex to interpret. It is only if that fluid is kept below boiling that it's simple. If one can be assured that it all boils, then it could become simple again, but assuring that isn't simple, either. Jed has written about how to do it. Rossi never, ever did this but simply claimed full evaporation without actually measuring it, but assuming it from possibly defective data and assumptions, and kept doing so long after the problem was identified and widely discussed.

  • Mary Yugo has often recommended the inexpensive calorimeter built by the GSVIT group


    I suspect that Mary likes that calorimeter because it's so easy to find fault with it. With such small flow tubes and coolant volume, it would surely reach boiling temperature with 1 kW heater power. How much steam could pass through the thin plastic tubes before something failed? And that's just a start at the potential problems.


    Barker apparently didn't know what he was really asking, but I'll assume by "proper calorimetry" he meant something real, not just a provocative implied criticism. The answer to that is shown by the following links, which I hope he will follow, thus learning something about the subject:
    http://goo.gl/iPMM5I MFMP HUG Concentric Calorimeter
    http://goo.gl/tl9NHz MFMP Valat/Celani Mass Flow Calorimeter
    http://goo.gl/FuObGn MFMP HUG Dual Differential Concentric Calorimeter


    There were several other development projects, including a large air-flow calorimeter (HUG) and a complex dual vacuum-insulated mass flow system built by Nicolas Chauvin (LENR-cars). Bob Higgins large water-bath calorimeter is still under development and I will leave those details for him to post.

  • Flow calorimetry is inherently better than open heat emitters tested for temperature, because it is difficult to control or even know all the effects that change effective thermal resistance, AND the thermal resistance is not a priori known.


    You still use cal runs for the flow calorimetry to check everything is wrong - but you are one step ahead because you know what the cal result is supposed to be.


    It is hands down a better method if practicable.


    There are still errors: water flow mismeasurement, boiling, TC mis-siting. All easily avoided if you are not Rossi.

  • gameover wrote:


    Looks like about 0.1 °C / minute. So it would reach boiling in about 10 hours.
    More water or a radiator seems like a good idea. Let's keep discussing this - it may be workable.


    The measurement time was too short to notice that the temperature line would turn as equilibrium was reached. The increase in temperature showed that equiilibrium wasn't attained.


    for more precision, ideally, the temperature rise would be maximized, but kept short of boiling. Water flow would be increased only if it started to boil. The lack of equilibrium probably is due to the calorimeter itself heating up.


    A calorimeter like this, properly done, can show high heat definitively. All of this depends on the independence and probity of those doing the work. For maximum value of its work, MFMP will continue to strive for scientific excellence, including all the necessary and rigorous self-doubt.

  • Quote from Roger Barker: “Hi, I am no expert on calorimetry but Mary Yugo is. She knows all about calorimetry. We've used them often in our fact finding sessions.


    Mary, can you kindly share your thoughts on using calorimetry for MFMP.”


    As you may…


    Hi, yes! This is definitely one of the calorimeters Mary and the crew at ECN have familiarity with. They've described its use in some detail. You will have to trawl through ECN to find the relevant sessions but I believe Popeye (Joshua Cude) and Mary had numerous sessions with this calorimeter. I was an observer in most cases and did not participate.


    Mary, are you still in contact with Josh? If so can you kindly extend him an invite so we can get his input on this matter. Thanks.


  • Flow calorimetry is much more difficult than one might imagine. I don't recommend it for these types of experiments. Conduction calorimetry is a lot simpler and you don't have to deal with leaks, flow rates, and numerous problems that arise in flow calorimetry. MFC can be done well. Brian Albiston did it well. JPB did mass flow. Jeff Morriss did air flow. Guess what? No excess heat.

  • Recirculating water at a steadily increasing temperature might not be the best idea, since the heat losses from all parts of the calorimeter will increase if the insulation is as 'casual' as the photograph suggests it might be. With a calibrated dosimetric pump it might be possible (and preferable) to draw fresh water at a known and stable temperature and volume from a reservoir, pass it through the heat exchanger, measure its temperature again at the exit point and then dump it.


    What do you think?

  • Recirculating water at a steadily increasing temperature might not be the best idea, since the heat losses from all parts of the calorimeter will increase if the insulation is as 'casual' as the photograph suggests it might be.


    I am not sure what photo you refer to but anyway . . . I suggested phase change calorimetry, boiling off the water. I suggested a steam trap that would return hot water to the reservoir. To do this you need a well-insulated reservoir. Otherwise, as you say, the heat losses from the reservoir will increase significantly, making the calorimetry complicated.


    You have to calibrate this system carefully. You have to test it with a cold reservoir and a hot one. It is complicated. Phase change calorimetry is complicated and I do not recommend it. Sometimes it is the only practical method.


    The other phase change calorimeter is the ice calorimeter, where the mass of ice that melts indicates the heat. This is one of the first kinds of calorimeter ever devised, in the 1770s and 80s. I have seen some cold fusion results from ice calorimeters. Most were bad, in my opinion. I do not recommend this instrument.