MIZUNO REPLICATION AND MATERIALS ONLY

Quote from Curbina

I understand you are talking about the R22? Please correct me if not. If you are talking about the R22, then The situation is not directly aplicable, as that reactor was heated internally. (Edit to add, probably I meant R20, the Mizuno reactor of which the data was presented by Jed at ICCF 22, R22 is a refrigeration gas, LOL).

I see these late “in oven”experiments as completely different, the oven is designed to reach an maintain temperature within it. Regardless of the losses, anything inert placed in the oven should reach the temperature of the oven and stay there.

If you put a stick in there, it of course will burn and release chemical energy and reduce the power input for a while and when the chemical energy is spent, it will get back to the set temperature.

A control reactor will simply get heated and stay heated while the oven is working.

A working reactor will start getting heated by the oven, and when the reaction is activated by the temperature, it will start producing heat, and it should start to be sensed by the oven controls and it would throttle down the energy consumption to maintain the temperature within the set point. The excess heat can be deduced from the power consumption to achieve a temperature, not very precise but good enough for practical purposes.

Display More

I don’t see why the model of reactor would make much difference as long as the activation heat was sufficient inside. Outside heat will eventually get inside, and the inside heat eventually gets out.

Otherwise, I was indicating what you appear to be saying here. A sufficiently heated reactor should (may) fire up without the usual electrical heat stimulation because the internal temperature will be the same (equivalent to the exterior heating, or interior heating, at some point). Unless the reaction is particularly picky about how the heat gets to some location or other unique circumstance.

The furnace should back off from adding heat commensurate to, if not directly proportional to, the reaction heat, in order to maintain the set temperature.

If the furnace is being used simply as a thermal box, and the usual heater (part of the assembly) is rated in 1000’s of watts, but only a few 100 watts are being used for the experiment, then the thermal mass of the furnace is grossly disproportionate to the experiment conditions and is possibly worse than the bare acrylic box for calorimetry purposes. A chest type freezer, much like Albiston was using is probably better.

Perhaps we can get some clarification on what the actual experiment involved.

The "incubators" I saw in Japan were equipped with many internal fans. The air is well stirred, so the temperature is uniform everywhere in the box. When you do isoperibolic calorimetry based on the internal cell temperature, I expect keeping the air well stirred will keep the results consistent. I do not think it means the outside temperature of the reactor is uniform. I wouldn't expect that. But, when you set the power to 100 W on Monday, and again on Wednesday, I would expect the internal temperature to be about the same. I do not know if a vigorous stream of air would help make the response linear, but I think it would make it consistent. This makes the incubator a lot better than leaving the cell in ambient unstirred air. The point of calibrating is to reverse the process and determine the power based on temperature. If it always goes back to the same temperature at a given power level, you can do that.

Years ago I had an HP A/D converter that had various methods of relating voltage levels to the instrument reading. For example, with a thermocouple, you could set a simple linear relationship: X volts = Y degrees. Or you could set up a more complicated curve. Or, if all else failed, you put in 0.1 V = X degree, 0.2 = Y degrees, 0.3 = Z degrees. You made a lookup table with as many entries as you wanted, so even if the response was not linear, the A/D gadget would return the right temperature. Or what was right when you first calibrated, anyway. If the surface temperature of the reactor in this case is not linear with power, with strange blips, you can use this ad hoc method. Test a bunch of power levels. Go back to the same power level several times. Manually add in each power level and associate it with whatever internal temperature the cell settles on after a while. Why isn't it linear? Who knows?? Who cares?

Interesting discussion but I think most are missing the forest for the trees. First of all the COP is 2.4, well beyond any kind of purported error. Secondly this high COP was achieved at a very low temperature relative to previous data wrapping a resistance wire around the reactor and using air calorimetry. This is not a peer reviewed published result but these results are illustrative of what is possible when the reactors are placed in an environment with a uniform temperature applied to the entire reactor surface. Experimental methods can be improved and teams are working on this now. Some of our academic and corporate partners are adapting their experimental methods to see if we can replicate these high COPs at lower temperatures. The calibration data is not linear (not sure if it "should be") and I am not sure how statistically robust the calibration process was (it should have error bars if done properly) but the temperature difference between control (90C) and active reactors (158C) is massive. The experiment in question was done by students under Dr. Muto's supervision. Professional researchers are now attempting a replication. Confirmation of large amounts of excess heat using different instrumentation and different methods does make Mizuno's claims for excess heat more robust.

Its getting to the point where it may become dangerous (COP >2) at relatively low temperatures. If temperature are raised and the previous known phenomenon of exponentially increased power output with increased temperature holds, we may be approaching the "run away" zone. Systems have to be modified for safety before we proceed.

Quote from Daniel_G

Interesting discussion but I think most are missing the forest for the trees. First of all the COP is 2.4, well beyond any kind of purported error.

That is a dangerous assumption. First, we have not discussed any errors, purported or otherwise. We don't have enough information to do that. Second, you can always have an error. People have made gigantic errors. Rossi managed to confuse 1 MW with 0 W. (I do not think he actually confused them -- I think he knew better, but many people looking at his results did.)

Quote from Daniel_G

The calibration data is not linear (not sure if it "should be")

Yes, it should be. Something is wrong if it is not. The problem may be unimportant, or it may be serious. Until you find out why the calibration is not linear you need to be careful. You cannot be fully confident of the results.

Quote from Daniel_G

Interesting discussion but I think most are missing the forest for the trees. First of all the COP is 2.4, well beyond any kind of purported error. Secondly this high COP was achieved at a very low temperature relative to previous data wrapping a resistance wire around the reactor and using air calorimetry. This is not a peer reviewed published result but these results are illustrative of what is possible when the reactors are placed in an environment with a uniform temperature applied to the entire reactor surface. Experimental methods can be improved and teams are working on this now. Some of our academic and corporate partners are adapting their experimental methods to see if we can replicate these high COPs at lower temperatures. The calibration data is not linear (not sure if it "should be") and I am not sure how statistically robust the calibration process was (it should have error bars if done properly) but the temperature difference between control (90C) and active reactors (158C) is massive. The experiment in question was done by students under Dr. Muto's supervision. Professional researchers are now attempting a replication. Confirmation of large amounts of excess heat using different instrumentation and different methods does make Mizuno's claims for excess heat more robust.

Thank you clarifying, but could you please write up a brief and simple description of the oven experiment goals and method?

I feel that we don’t quite have a good handle on what the experiment conditions were, and therefore are guessing many of the details, leading to possibly pointless conjecture.

Quote from Paradigmnoia

Thank you clarifying, but could you please write up a brief and simple description of the oven experiment goals and method?

Just post the rest of the slides. I will translate them.

(You can Google translate them if they are in text format, but you may confuse a resistance heater wire with a wire that resists heat. Sort of like confusing a house cat with a cat house, the classic translator's dilemma.)

Jed resistance heaters are made of heat resistant wire. The words are interchangeable in this case and most certainly non-mutually exclusive.

As for dangerous assumptions, please go back and reread what I wrote. I said these results stimulated us to redo these types of experiments with professional researchers from both academia and corporate partners.

I didn’t say the results were conclusive, but they are interesting enough to us for spending more resources in pursuing this.

You can’t just hand wave 240% error away. If any of you have some concrete suggestions of where to find 240% error please do let us know.

Not sure why you would expect a linear relationship from power input to temperature. There are air leak loses, conduction losses, radiative losses which are exponential and each playing differently weighted roles at different temperatures. While I concede the data looks a bit imprecise, how can you explain a 70C temperature difference with a maximum temperature reading uncertainty of less than 5C?

R^2 of over 0.98 and you guys are saying this isn't linear! Exactly what R^2 would you accept? In the statistics that I studied a 98% linearity is pretty damn good. The temperature specs from this very well known scientific equipment maker is +/-3C at 300C and +/- 5C at 500C. There is a 70C difference in temperature between dummy and test reactors. The experiment was very simple. They are able to control input power manually and measure temperature. Input power is the independent variable and temperature is the dependent variable. Calibration runs were done with a dummy reactor and test runs were done with an active reactor. The phenomenon is repeatable and it is a large effect.

Again, we thought this data is exciting enough to send an early stage report to our friends and supporters in this forum and rest assured proper scientific technique is being followed and more robust calibrations with sufficient replicates to calculate error bars will be done in multiple locations and multiple teams with high quality equipment.

Finally Mizuno specifically asked me to express to Jed first of all and the rest of the forum that his personal situation does not allow him to participate directly in this forum at this time as his personal situation is rather difficult at the moment, so I am doing my best to remove that burden from him and communicate with our friends in this community. In Eastern culture it is not customary to be so open with personal problems but in the West this is expected and is seen as probably "rude" for not responding to emails, etc. Trust me Mizuno is not being arrogant or rude. He's dealing with a life and death situation of close family and doing what any educated and cultured Japanese would do (keeping his personal misfortune private and focusing on taking care of his family).

Mizuno also mirrored my opinion that he sees no a priori reason to expect linearity in this data. However we ran the numbers based on this assumption and <2% non-linearity is certainly quite low but again I am not sure what the meaning of this is. A more robust series of calibrations will be done or rather are being done now and something will be sent to peer review for publication so please be patient while this is done.

Quote from JedRothwell

That is a dangerous assumption. First, we have not discussed any errors, purported or otherwise. We don't have enough information to do that. Second, you can always have an error. People have made gigantic errors. Rossi managed to confuse 1 MW with 0 W. (I do not think he actually confused them -- I think he knew better, but many people looking at his results did.)

Yes, it should be. Something is wrong if it is not. The problem may be unimportant, or it may be serious. Until you find out why the calibration is not linear you need to be careful. You cannot be fully confident of the results.

Firstly both myself and Mizuno do not agree that there is any a priori reason to expect linearity as stated previously. If we go forward under your assumption that is should be linear, we can calculate the expected magnitude of error due to non-linearity. The result is +/-2%, not 200%.

To mention Rossi's 1MW and 0W in this context is not only unhelpful, but ludicrous on its face. Dr. Muto's scientific and personal credentials are unindictable. The equipment used is top quality scientific equipment. The experimental concept is beautiful in its simplicity and could very well represent a major leap forward in low temp/high COP LENR research.

As professional scientists we not only accept but actually encourage critical reviews of data and methods but please do us the favor of actually doing the calculations (the above took me <10 minutes) before making any claims. 2% non-linearity is insufficient to explain a 70C/214W difference from dummy to test reactor difference. To even suggest that it could is not a valid or logical scientific argument.

Yes we have a few reputable partners assisting.

Daniel_G is using "2% non-linearity" in the wrong way here. When you fit a straight line to some data and find that R^2 = 0.98, this does not mean either that the relation between the independent and dependent variables is actually (or even mostly) linear or that the 2% gap is an measure of nonlinearity. Instead, that 2% is intended to measure the stochastic variability around the fitted line that is not explained by variation in the independent variable. The only real way to measure the extent of nonlinearity is to fit a curved line to the data and see how much that reduces the unexplained scatter.

A thought experiment will clarify. Suppose that the true relationship really is linear and that at each level of input power you add the same amount of scatter in temperature readings. Depending on how much scatter you add, you can easily get R^2 = 0.98, or R^2 = 0.50 or even R^2 = 0.20 ... all from a perfectly linear relationship between input power and temperature.

Its a technical point. I'm not sure how important it is to the physics of things here. There appears to be only a single temperature measurement per input point in the data so far and more observations at each input would make it clearer whether this is a nonlinear relationship or a linear relationship with unexplained scatter.

I think that the relationship should be linear at all except the most extreme temperatures.

I already stated that more robust calibrations are being done with sufficient replication to be able to calculate error bars. I also stated that this was an early report.

Maybe my brain is numb from not being able to sleep in my new hotel room, but I can’t make that equation work

Daniel_G ,

Thanks for the additional information. We look forward to an update when it is available.

Best wishes and my heartfelt sympathy for the Mizuno family.