In conclusion, the only plausible explanation for this strange behavior is that, in the case of the active reactor, an internal source of electric heating was mistakenly connected to the power unit, rather than the normal external heater.
That is not plausible, because multiple meters are used to ensure there can be no mistake of that nature.
That is not plausible, because multiple meters are used to ensure there can be no mistake of that nature.
Data from spreadsheets are much more meaningful than unsupported sentences. The red and blue curves in the previous jpeg (*) come directly from the data in the spreadsheets that you have uploaded in the internet.
What is your explanation for the their strange behavior?
What is your explanation for the their strange behavior?
LENR is a combination of endothermic and exothermic processes.
If the endothermic reactions are in excess the reactor will cool
If the exothermic reactions are in excess the reactor will heat.
In LENR the reactions are nuclear... not chemical
but the exo/endo heat/cool geenral idea applies.
In chemical reactions this dynamic balance between the two modes is observed to
cause temperature oscillations in some cases.
eg the Belousov-Zhabotinsky Reaction
The frequency of the thermal oscillations will of course
be moderated by the thermal inertia of the reaction system.
Display MoreMizuno’s tests – Expected DTair trend for the 120 W run of the May 2016 experiment
The following jpeg illustrates the main reason which suggests that the an inner heater was powered during the active runs of the May 2016 experimental campaign, instead of the regular ouside heater.
In
Figure 5 of the 2019 JCNMS article (1) shows that the active and the control reactors are apparently identical. The same identity is shown in Figure 11 of the 2017 JCMNS article (2), in which, at point 2.6, we can also read: "The same type of reactor is used in the calibration, and is installed as a control for calibration of the heat balance in the enclosure described below. The design, size, weight, and shape of this calibration reactor are exactly the same as the reactor used for testing."
Therefore, everyone expects that, within the limit of the error guaranteed by the manufacturers, all the components of the two reactors (shown on the schematic of Figure 1 (1)) are also identical, including the external reactor heater, briefly described in (2) by providing these univocal parameters: "A 2-m-long heater of stainless sheath was wrapped around the reactor body. Its purpose was to heat the nickel mesh in the reactor. The heater capacity is 100 V, 600 W, with a maximum temperature of 500°C."
However, the power output curves of the active runs at 80, 120, and 248 W clearly show a strange trend compared to the correspondent control curves. This difference was also noted by the experimenters, as they reported in (2):
"The calibration test by the control reactor were performed at three input power levels. Figure 27 shows the output power of the control reactor for the input of 80 W, 120 W and 248 W. The output power is confirmed to be the same as input power.
The results of same inputs to the test reactor are shown in Fig. 28. These tests behave differently from the calibration test. …"
The weirdness of this different behavior is much more evident when drawing the active and control curves in the same graph. The above jpeg shows the two curves for the 120 W active and control tests, whose spreadsheets was release in summer 2017 (3-4). The represented quantity is the DTair, which is assumed to be proportional to the power output. Moreover, the two curves have been slightly shifted over time in order to make them both start at the power onset.
We can see that, during the heating phase, the active curve (in red) increases much more slowly than the control curve (in blue) and the two curves intersect only after a couple of hours. This is very strange, because - if the only difference between the two experimental setup was the activation of the internal mesh, which was claimed to generate excess heat in addition to the heat produced by the external heater - it was expected that the active curve would have remained above the control curve for the entire heating phase, as exemplified by the green dotted line reported on the graph.
In conclusion, the only plausible explanation for this strange behavior is that, in the case of the active reactor, an internal source of electric heating was mistakenly connected to the power unit, rather than the normal external heater.
(1) https://www.lenr-canr.org/acrobat/MizunoTexcessheata.pdf
(2) https://www.lenr-canr.org/acrobat/MizunoTpreprintob.pdf
Ascoli,
Consider those two curves. Both look (to 1st approximation) like single pole low pass filtered signals, such as you might get from a heat source on a thermal time constant. However the time constant of the active run is 7000s, the time constant of the calibration run is 500s (roughly - we could get a more accurate value from spreadsheets).
The air temperature in this system, for constant air speed, depends roughly linearly on the reactor surface temperature.
The active run time constant is equal on rising and falling edge, so this does not appear to be anything other than a thermal time constant effect.
Given that as the paper says the actual reactor masses are identical, how can we get these very different thermal time constant?
Therefore the active run would appear to have much less effective cooling than the control run.
That is possible, the two reactors are in a different position, but it is a surprisingly large effect (10X). It could be obtained by varying the air speed, with lower air speed in the active run.
This is a red flag. The very different time constants imply differences in the setup (large ones) just as the different heater resistances imply a difference in heater design, contrary to what is stated in the paper.
Ascoli may see deception or catastophe (you'd have to ask him) but I see mistake
The mesh sees heat. The mesh needs heat
The mesh will suck in heat in an endothermic phase
The mesh will give out heat in an exothermic phase.
The mesh is metonymy for the unknown number of
small active and less active NAEs on the Pd/Ni meshes .
Many of these NAEs may be running on different time
courses.. from the others... although some may be coordinated.
The net effect of these NAEs will
produce a hairier graph than shown at right.
Depending on how Mizuno has activated the mesh prior to
the experiment... ( it took hours and hours in 2017
the total time for both phases may be limited.
Without more experimental knowledge of the endothermic and exothermic reaction(s)
kinetics at NAEs there is only speculation.
The huge excesses in R20 cannot be explained by measurement error.
Mizuno’s May 2016 tests – Heat fluxes during the active and control runs at 120 W
Consider those two curves. [...] the time constant of the active run is 7000s, the time constant of the calibration run is 500s (roughly - we could get a more accurate value from spreadsheets).
[...]
The active run time constant is equal on rising and falling edge, so this does not appear to be anything other than a thermal time constant effect.
Given that as the paper says the actual reactor masses are identical, how can we get these very different thermal time constant?
These are appropriate considerations and your question highlights the crucial point: supposing that all the rest (reactors and calorimeter) of the two experimental setups (ie those of the active and control runs) is identical, the only parameter that could affect the time constant is the heated mass.
At a first glance, we observe that the reactor are identical, so we could conclude that the involved mass should be the same, but looking better at the paths of the heat fluxes we can see that when the powered heater is the external one, only a tiny part of the reactor body is crossed by the heat. On the contrary, if the heat source is placed inside the reactor, all the internal walls are exposed to the radiation heat coming from the inner source and the heat flux should necessarily cross the whole thickness of the body. Therefore, the different time constants can be easily explained by the very different extension of the metal mass involved in the heating process, as illustrated in the following jpeg.
QuoteThis is a red flag. The very different time constants imply differences in the setup (large ones) just as the different heater resistances imply a difference in heater design, contrary to what is stated in the paper.
Actually, all the photos show that the two reactor are identical, so the most plausible explanation for the different time constants is that, for the active reactor, an inner heater (maybe the ceramic heater shown in the schematic) was mistakenly connected to the PSU, instead of the regular external heater.
These are appropriate considerations
THH was quoted as saying "The active run time constant is equal on rising and falling edge,"
A closer look a the 120W active data reveals
no equal "time constant "
In addition the falling curve has a long slow tail.
The scatter of points and the shapes do
not conform simply with time constant exponential curves
In particular there is a huge scatter at the end of the rising curve
which tranforms to a much smaller scatter on the falling curve.
The scatter and shape can be result from a
sequence of net endothermic followed by net
exothermic processes which start to terminate at time 20,500 occurring in multiple small NAE (nuclear active environments) on the Ni/Pd
mesh in the 2017 120W excess heat production.
Display MoreMizuno’s May 2016 tests – Heat fluxes during the active and control runs at 120 W
These are appropriate considerations and your question highlights the crucial point: supposing that all the rest (reactors and calorimeter) of the two experimental setups (ie those of the active and control runs) is identical, the only parameter that could affect the time constant is the heated mass.
At a first glance, we observe that the reactor are identical, so we could conclude that the involved mass should be the same, but looking better at the paths of the heat fluxes we can see that when the powered heater is the external one, only a tiny part of the reactor body is crossed by the heat. On the contrary, if the heat source is placed inside the reactor, all the internal walls are exposed to the radiation heat coming from the inner source and the heat flux should necessarily cross the whole thickness of the body. Therefore, the different time constants can be easily explained by the very different extension of the metal mass involved in the heating process, as illustrated in the following jpeg.
Actually, all the photos show that the two reactor are identical, so the most plausible explanation for the different time constants is that, for the active reactor, an inner heater (maybe the ceramic heater shown in the schematic) was mistakenly connected to the PSU, instead of the regular external heater.
Ascoli,
I'm not sure we will get a definitive answer on this, but in the interest of accuracy:
(1) increasing thermal mass is NOT the only way to get a lower time constant. Increasing cooling, so that the equilibrium reactor body temperature is lower, is just as good. To see why, note that the initial reactor body rate of temperature chnage depends on the power. If the final temperature is much higher |(due to less good air cooling) it will take longer to get there.
(2) True, if we could ignore the thermal conductance of the reactor body, a small heated part would heat up quicker. But I suggest that large stainless steel container will conduct heat well enough to heat up together.
(3) Another possibility - with an external heater - is that the thermal resistance between the heater coils and the reactor body is relatively high, so that the coils heat up much more than the body. That would provide a faster time constant because again the reactor body would heat up less. Interestingly, you could also perhaps distinguish this because there would be some (small) heat output within a few seconds of the power input, due to heat dissipated from the heater coils.
On balance, of the various mechanisms for the shorter time constant, I quite like (3).
We can now quite easily make progress because we can estimate the thermal capacity and therefore the time constant for given cooling.
It is important to note an inconsistency here.
The orginal paper describing the 100% excess results
A: (2017) https://lenr-canr.org/acrobat/MizunoTpreprintob.pdf Observation of excess heat by activated metal and deuterium gas
is quoted as reference [1] of the 1st (ICCF21) paper:
B: (2019) http://lenr-canr.org/acrobat/MizunoTexcessheata.pdf Excess Heat from Palladium Deposited on Nickel (preprint)
In paper A the 100% excess results are shown as coming from a 20kg reactor, fitted into the calorimeter with a similar control reactor.
The reactor body is a cruciform cylindrical shape as shown in Fig. 1. All parts are connected with metal seal flanges.
The upper part of the reactor comprises the heater power inlet, the high-voltage discharge electrode, and a thermocouple. Several platinum temperature measuring elements (Okazaki; Pt AA class ±(0.1+0.0017|t|) − 196 ∼ 450) were
installed in the reactor. As shown in Fig. 1, there is a Kovar glass window on the left-hand side of the reactor, and a
pressure gauge; and a mass spectrometer valve and the gas inlet valve is on the right-hand side. A vacuum evacuation
system and a quadrupole gas analysis system are connected to the body.
The reactor is made of SUS 316. Its volume is 2740 cm3
and its weight is 20.3 kg
The testing is described as using two identical reactors in the calorimeter:
The same type of reactor is used in the calibration, and is installed as a control for calibration of the heat balance in the
enclosure described below. The design, size, weight, and shape of this calibration reactor are exactly the same as the
reactor used for testing. The internal reactants are the same nickel, of the same weight, size, dimensions and position.
Both are washed and wound the same way. However, excess heat is not produced by the calibration electrode even
though deuterium gas is added to the cell, because the nickel material is not processed as described in Section 2.5.
Figure 11 shows the test and calibration reactors. The latter reactor is placed in the calorimeter to obtain test and
calibration data, and to demonstrate a heat balance of zero. Both reactors are shown in the photo.
In paper B the same 100% excess results are described as older method:
1.1. Old method
A reactor with a cruciform shape was first used in this project (Fig. 1). It weighs 50 kg. Later, 20-kg versions of this
reactor were used, as well as cylindrical reactors. All have palladium rods in the center. The rods are 250 mm long,
wound with palladium wire (Fig. 2). This is the positive electrode. The negative electrode is a nickel mesh which is
fitted against the inside wall of the reactor, and connected to ground (Fig. 3).
The older method is described in detail elsewhere [1]. It includes several rounds of cleaning the electrodes by
heating and evacuation to remove impurities, followed by weeks or months of glow discharge to erode the center
palladium electrode and sputter palladium onto the nickel mesh.
The highest power observed with the older method was 480 W output with 248 W input, or 232 W excess (Fig. 4).
Note that reference [1] here is in fact to paper A which does detail the results, but describes them as obtained using 20kg reactors.
The new method is clearly described as being tested differently, using the new-style reactors.
All of the tests with the new method used airflow calorimetry. Two different pairs of reactors were used. The first
pair were cruciform reactors similar to the one used with the older method, but smaller and lighter, weighing 20 kg
instead of 50 kg. They are shown in Fig. 5 placed side-by-side in the calorimeter. The second pair were horizontal
pipes (Fig. 6).
For a 50kg reactor (304 stainless steel) we have 500J/kgK => 25kJ/C
At 250W input we therefore have a temperature rise of 1/100 C/s assuming no cooling (true enough for the initial part of the curve).
For the active reactor we have (roughly) 5000s for an initial rise of 50% of the final (assuming the reactor temperature / air out temperature relationship is linear - approximately true.
So this corresponds to a 50C temperature rise. Or, for the whole thing, 100C rise.
That sounds about right. the calculation is quite approximate: if it become important we could fit a more accurate model of the rise, incorporating cooling is well.
Now consider the 10X faster temperature rise from the control reactor.
In that case we have 50% of final temperature in approx 500s.
That corresponds to a temperature rise (total) of only 10C for the reactor body.
For the newer (20kg) reactors we have 2.5X higher temperatures from the same calculation: 250C rise and 25C rise.
This contradicts the claimed calorimeter output. For a calorimeter body at 10K above ambient the expected output temperature rise would be much lower.
Conclusions
The 2017 100% excess results, discussed here, are described in the 2017 paper A, and also in the 2019 paper B. However it is not clear how they were obtained. Paper B says these (old method) came from (old method) calorimetry using 50kg reactor. Paper A 2017 says these same results came from calorimetry which is described as identical to the new method calorimetry in paper B: dual 20kg reactors side by side in the calorimeter.
Therefore we cannot proceed further with the analysis of these results until this is resolved. Are they obtained as stated in paper B (2019), or paper A (2017)?
Jed please help.
NB I am NOT assuming that old method (mesh conditioning) is the same as old method (calorimetry). However it appears from Paper B that the 2017 100% excess results are both old method (mesh) and old method (calorimetry).
The very different time constants imply differences in the setup (large ones) just as the different heater resistances imply a difference in heater design, contrary to what is stated in the paper.
Not necessarily.
In contrast to the inactive reactor which has only thermal inertia
with a small time constant the active reactor does not just have
thermal inertia.
The active reactor has endothermic processes which soak up heat
plus exothermic processes which produce heat.
The balance of these changes as the reactor temperature changes.
The soaking up of heat will of course slow down the release of heat from the reactor
Certainly the reactor cannot be modelled as having one time constant as TTHnew has asserted.
The scatter in the temperature and the shape of the temperature profiles suggest otherwise.
The rising and falling different shapes do not simply suggest a reversal of processes.
The reactions arising from multiple NAEs in the mesh cannot be modelled simply.
Jed: using the spreadsheets of raw results from the R19 runs we could make a lot of progress, I believe, by looking at the rate of change of reactor temperature in response to a heater power change. It would tell us what is the overall power driving that change independently of the calorimetry result.
There would be another time constant: with two 20kg reactors in the enclosure the second reactor would need to heat up as well, but this would be decouples from the active reactor and perhaps we could identify the two different time constants.
In any case the above uncertainty about the 2017 data remains. Whether the data comes from 50kg or 20kg reactors there is still a problem caused by an inconsistent time constant. Basically the control reactor time constant is much faster than expected from the power and the mass of the reactor.
THH
RB:
In contrast to the inactive reactor which has only thermal inertia
with a small time constant the active reactor does not just have
thermal inertia.
The active reactor has endothermic processes which soak up heat
plus exothermic processes which produce heat.
The balance of these changes as the reactor temperature changes.
The soaking up of heat will of course slow down the release of heat from the reactor
Well, by positing unknown endothermic and exothermic processes you can match any temperature curve under the sun! So I cannot technically disagree, although Occam's razor does not favour such coincidental matching of a single expected thermal time constant by two unknown reactions.
How about we consider just the control reactor: and ask how that time constant can come about in a 50kg (20kg?) reactor heated by 250W?
THH
Display MoreTHH was quoted as saying "The active run time constant is equal on rising and falling edge,"
A closer look a the 120W active data reveals
no equal "time constant "
In addition the falling curve has a long slow tail.
The scatter of points and the shapes donot conform simply with time constant exponential curves
I said the THH considerations were "appropriate", I didn't say they were "exact". THH himself warned that he provided a roughly estimation of the time constants and, given the complexity of the reactor body, there is no wonder that these constants varies during the heating and cooling phases.
QuoteIn particular there is a huge scatter at the end of the rising curve
which tranforms to a much smaller scatter on the falling curve.
It's normal. Heating is different from cooling. During the heating phase the temperature in the system are less uniform, because the heating elements are hot. After the power off, temperatures tend to become more uniform, causing a lower scattering in the temperature of the turbulent air flow.
QuoteThe scatter and shape can be result from a
sequence of net endothermic followed by net
exothermic processes which start to terminate at time 20,500 occurring in multiple small NAE (nuclear active environments) on the Ni/Pd
mesh in the 2017 120W excess heat production.
Exhilarant!
Back on Earth. What we can say with absolute certainty is that the time constants of the heating and cooling phases of the active reactor are one order of magnitudes greater than those of the control reactor. This is evident to everybody and should have been enough for a long running researcher in the field to deduce that something went badly wrong in his experiment. It should have been sufficient also for a long running reporter in the field, like JedRothwell, which also wrote and signed the most recent papers about these tests. On the contrary, despite all these warning evidences and the several huge inconsistencies contained in the spreadsheets (*), these results have been proposed as extraordinary outcomes in a JCMNS article in 2017 (1), illustrated at the last ICCF21 in 2018 (2) and eventually confirmed as an effective way to produce several hundred watts of excess heat in a more recent JCMNS article (3).
IMO, the 2016 Mizuno's experiment is a paradigmatic representation of 30 years of CF/LENR history and worth all the attention at the next celebration of ICCF22 in Assisi.
(*) Mizuno Airflow Calorimetry
(1) https://www.lenr-canr.org/acrobat/MizunoTpreprintob.pdf
(2) https://www.lenr-canr.org/acrobat/MizunoTexcessheat.pdf
(3) https://www.lenr-canr.org/acrobat/MizunoTexcessheata.pdf
RB: the "long tail" is probably matched by a "long tail" on the rise time (not shown due to limited on time). It looks to be there from Fig 28. One reason to expect this, as I've said above, is the heating/cooling of the other reactor body in the enclosure.
One way to estimate the magnitude of complex second order dynamical effects here is to look at the differing power control traces here. No "exothermic and endothermic reactions" but see Fig 4 for differences in shape of rise time tail at different powers in what should be a purely passive system with fixed frequency response.
However ascoli is also correct:
It seems certain that there is not the exact same geometry between the control and active setups for this 2017 paper data, the apparent 10X difference in thermal inertia needs explanation. If the control data was with 20kg reactor, and the active data with 50kg reactor, we still have a 4X difference in time constant to explain.
In this complex system there is no reason to expect exact symmetry between power on and power off dynamics.
This is true.
There is no reason everto expect a simple capacitor like time constant.
as THH has proposed
"THH was quoted as saying "The active run time constant is equal on rising and falling edge,""
There are many different processes that contribute to
Deuterium -Deuterium fusion
the most important of these being the reactions which dispose of the 20 MeV fusion energy
primarily as energy in the subeV range.
over thousands of individual atoms in the many small nuclear active environments.
This will be discussed at Assisi.
This is true.
There is no reason ever to expect a simple capacitor like time constant.
as THH has proposed... This will be discussed at Assisi.
The best is the enemy of the good. As is often the case in real experimental work there are many second order effects that prevent exact results. However, in this case the approximate results which indicate time constants differing by a factor of 10 are enough to make a clear discrepancy. Any decent engineer (or decent experimental physicist) would agree that.
Indeed, Assisi will have the boring job (since it is nothing to do with LENR) of explaining the very fast time constant for the control data graph in those 2017 results. Perhaps it is not fair to expect an LENR conference to deal with this. Let us do so here using the expertise and intimate knowledge that Jed has relating to these results.
Given the very large reactor mass, and known heating power in thermal contact with the reactor, a single thermal time constant is expected - with second order corrections for the enclosure heating and other issues. Because of this, we obtain very valuable additional information from looking at the control, and active, time constants. (For smaller items inside the reactor body providing the heat the same thing applies since small mass items have a small time constant compared to the reactor body).
Given the very large reactor mass, and known heating power in thermal contact with the reactor, a single thermal time constant is expected
Given THH assumptions
there should be no scatter in the data.
The scatter is the individual temperature variations which are measured in a large mass reactor at intervals in the order of minutea
These are consistent with large fluctuations in endothermic/exothermic processes
in multiple small NAEs.
The single thermal time constant repetitively postulated by THHnew
is plainly seen to be a poor fit to the experimental data.
Given THH assumptions
there should be no scatter in the data.
The scatter is the individual temperature variations which are measured in a large mass reactor at intervals in the order of minutea
These are consistent with large fluctuations in endothermic/exothermic processes
in multiple small NAEs.
RB: you are I am afraid not thinking clearly here. And putting words into my mouth that are not mine.
The noise on this trace is clearly interesting, but cannot be due to endothermic/exothermic reactions inside the reactor. Anything there would be smoothed out by the very large reactor body thermal mass. In fact the proposal that individual NAE variation could cause this is ridiculous. if you do not agree try some thermal modelling, noting the reactor body high thermal conductivity and heat capacity.
Ascoli's suggestion for this variation is possible: air turbulence results in the air stream removing variable amounts of reactor heat, or varying amounts impacting the RTD. That would explain why the noise appears at an amount proportional to the temperature difference from ambient. But I make no assumptions here when experimental results are noisy - it could something else we have not considered.
I'd just add; any engineer, or competent experimental physicist, would tell you that high frequency noise on a waveform (the scatter) in no way prevents accurate analysis of much lower frequency time constant,
In conclusion, the only plausible explanation for this strange behavior is that, in the case of the active reactor, an internal source of electric heating was mistakenly connected to the power unit, rather than the normal external heater.
That is incorrect, as I said. It is not plausible that several experienced scientists using multiple meters (including ones brought by me and other outsiders) would fail to notice this problem for many years.
I expect this difference is caused by latency. It is taking 83 minutes longer to heat up to the same temperature. The calibration was performed with the heater wrapped around the outside of the reactor, and the reactor did not get very hot inside. The excess heat run was heated inside, mainly with glow discharge. The entire reactor got much hotter before reaching a terminal temperature. The reactor is 50 kg of stainless steel, so the thermal mass is substantial. Once the reactor reaches the terminal temperature, the heat flow is the same regardless of how hot the reactor is inside, so the calorimeter will see the same level of heat if there is no excess, which it often did during failed runs. The reactor internal temperature was higher, and -- as I said -- it took longer to reach that temperature. It also took longer to cool.
I expect this difference is caused by latency. It is taking 83 minutes longer to heat up to the same temperature. The calibration was performed with the heater wrapped around the outside of the reactor, and the reactor did not get very hot inside. The excess heat run was heated inside, mainly with glow discharge. The entire reactor got much hotter before reaching a terminal temperature. The reactor is 50 kg of stainless steel, so the thermal mass is substantial. Once the reactor reaches the terminal temperature, the heat flow is the same regardless of how hot the reactor is inside, so the calorimeter will see the same level of heat if there is no excess, which it often did during failed runs. The reactor internal temperature was higher, and -- as I said -- it took longer to reach that temperature. It also took longer to cool.
Jed, it is helpful that we now know these two runs were not that same, with one heated inside and the other outside.
You say that the entire reactor got much hotter in the active test. From the output, you might expect it to get roughly 2X hotter. That would not explain the 10X difference in time constant. Are you saying that it got (maybe) 10X hotter? That is possible, but it would be interesting to understand why the active reactor cooling was so much less good. One option is that the external heater gets much hotter than the reactor it heats and is cooled directly by the airflow and so heats up the reactor much less than would be the case were it in excellent thermal contact with the reactor body and therefore the same temperature as it.
Note that in this system the 20kg or 50kg (which is it?) reactor has much larger thermal mass than any other part of the system.
I'm not ruling anything out here, but this is a very significant aspect of the recorded data, and it would be good to understand it.
We can now perhaps say that these published 2017 excess heat results are unsafe. If the reactor body is much cooler in the control than the active test, aspects of the airflow, possible waste heat transfer direct to the RTD, etc could determine these results. Or, differences in power input measurement (known to exist) combined with different heater resistance, could do this. I realise the new method calorimetry (two reactors at the same time in one enclosure) is different and might well be much more robust. I'm not therefore saying this makes the R19 results unsafe, but equally it makes extreme caution about all these results sensible.
Some comment from Jed about whether these 2017 results were old method or new method calorimetry would be helpful? (20kg or 50kg reactor, dual or single reactor in box, calorimeter with insulation or no, RTD in output tube after blower or somewhere else, etc, etc).
