No need for a vertical scale.. Resistance is known by the report (Joule heating figures reported). Mains AC voltage is known. Time scale is known. Ohm's law is all is needed.
But you are right that the plot was intentionally published in an OL condition thinking it would not reveal the actual power. The problem is that there is enough information to derive it anyway.
I took a quick look at the other thread, yes it is more pertinent but last comment was 5 days ago so i'd rather continue on this fresher thread.
Actually the Lugano report shows in figure 5 that the instrument is overloaded (readings are "OL" everywhere).
In the last link in my post above I simulated the current waveforms expected for the dummy run, then for the experiment run with the load resistance compatibile with 900W (tardively alleged to have decreased to 0.36 ohm due to temperature and use of a specially doped SiC heater) , finally with the same 1.23 ohm loads resulting from the declared Joule heating in the dummy run (just in case the resistance drop isn't true, as expected if the wire were Inconel as initially stated). Well, the only case that matches the figure 5 waveforms is the last one, which yields 3 kW and would justify overloading the PCE830.
Rossi says that the plot referred to the setup phase. So we cannot state that both errors (Input power and emissivity) were present at the same time. However the plot says that at some point they were providing 3kW electrical power.
I speculate that they started off with input up to 3kW (fig. 5 of the Lugano report) and a thermocouple, but weren't getting excess heat, and blamed the thermocouple's poor contact to the ridged surface. So they removed the thermocouple and resorted to (uncalibrated) optical thermography. Given the good results in terms of apparent COP (thanks to the flawed infrared emissivity thus overestimated temperature) they did not need to go again as high with the input, and cherry-picked the favorable result.
And they didn't bother to correct the Joule heating used in the previous tables (discarded due to COP<1) that revealed the 3kW in the first place, and caused the tardive explanation of special SiC heating resistors.
This last paragraph is just speculation.. But the mistake on emissivity is a fact, and when accounting for plausible additional measurement errors it may well imply that COP was <1 all the time.
Yes Paradigmnoia - 3.0kW - post re-edited. Thanx
Alain Co,
Levi stated on Mats' blog that even with emissivity 1 the Lugano COP is well above 1. Let us discuss this.
What many of us criticized is not the total emissivity graph (fig. 6 of the report) but its use in the Optris camera setting which should be the average spectral emissivity over its detection range.
Indeed the reading of the Optris camera is proportional to infrared emissivity and to T^4, thus by using 0.4 vs. 1.0, the fourth power of T would have been estimated 2.5 times higher (i.e. the reading around 1400°C would rather be 1060°C).
This means that the radiated power, proportional to total emissivity and to T^4, would have been overestimated by a factor 2.5 to be compared to the claimed COP 3.1 to 3.6. Another contribution would come from convection, which would be lower by roughly 100W if 1400°C scaled to 1060°C, but this is a minor effect that would still leave a COP in the range 1.2 to 1.3.
But there are other sources of error.
The Lugano report says the error around the emissivity of figure 6 is 0.01 which is ridiculously low. When comparing several references the figures dance around by 20%.This alone casts doubts on the conclusions.
In addition there are serious reasons to not trust the input power measured, given the infamous figure 5 of the Lugano report, with the PCE830 overloaded by the current peaks, whose width matches a 3.0kW [*] electric power much better than the 900W declared.
Some refs of mine
Feb 2015: Applying the "Lugano report" method to compute COP of MFMP's dummy dogbone:
http://cobraf.com/forum/immagini/R_123579764_1.pdf
http://cobraf.com/forum/immagini/R_123579764_2.pdf
Oct-Nov 2014: Inconsistencies in current waveforms published in the Lugano Report
http://cobraf.com/forum/immagini/R_123571297_1.pdf
[*] thanks Paradigmnoia for spotting the typo
A question to Dr Parkhomov.
In the calorimetry setup, what is the material of the reactor vessel ? Is it enameled aluminium ?
Thank you in advance
Inconsistencies in current waveforms published in : Levi et al. “Observation of abundant heat production from a reactor device and of isotopic changes in the fuel” (referred to as TPR2), and related : Levi et al. “Indication of anomalous heat energy production in a reactor device “(referred to as TPR1).
Andrea S. , 14 October 2014, [reedited 23 Nov 2014]
I hate to say this, but the plot in TPR2 figure 5 and page 6, although not on a readable scale, may be revealing.
Please refer to
http://www.cobraf.com/forum/immagini/R_123566844_1.pdf
for a better reading.
The PCE-830, when displaying voltage and current waveforms ,normally synchronizes all waveforms to the upwards zero-crossing of the V1 waveform when present, or of the I1 waveform if V1 is not available, and sets the timescale to one period.
The displayed I3 waveform is determined by the triacs closure. Current flows when the triac on V3 is closed at the same time as V1 or V2, and it is a fair assumption that all three triacs are commanded with the same conduction angle (since a single feedback is used for all).
Now the displayed waveform for I3, is not the one would expect, neither with clockwise nor with counter-clockwise three-wire three-phase connection
The plot in TPR2 figure 5, instead is well matched by a negative I3 plot in the CCW connection, such as would result from inverting the clamp ammeter. [Note: this statement has been challenged, as the I3 plot seems to trigger at the beginning of the timescale. I thank Franco (who writes on the Cobraf Forum) for his sound technical advice and S. Caggia (22passi blog) for his constructive criticism ].
Now, the effect of reverting the I3 probe would be to underestimate the actual input power by a factor 3, i.e. close to the declared COP.
[ A later analysis, admitting the downstream PCE-830 to be connected 4-wire by using the upstream neutral (although the neutral is not shown in the TPR2), does find a plausible reading that fits the peaks' orientation in the TPR2 figure 5. On the other hand, the same analysis shows that current pulsewidth in figure 5 is compatible with a 3KW consumption (implying COP around 1), whereas the 0.9kW consumption declared by the authors would yield much narrower pulses. The updated analysis can be found in
http://www.cobraf.com/forum/immagini/R_123571297_1.pdf .
Once more the authors should better describe their setup, clarify the instance when the plot was taken, clarify if the measurement was 4-wire, and justify the apparent 3KW consumption resulting. The reversed clamp hypothesis remains the one that best fits the three-fold power underestimation .]
It does appear absurd that such a gross mistake could have happened. What is most disturbing is that a similar discrepancy on the PCE-830 plots happened in the TPR1.
In the TPR1 the I1 waveform similarly appeared inverted with respect to the one expected, such as would result by inverting the current probe.
When simulating the circuit on a spreadsheet, with triac conduction angle set for low regimes (as in the experiment setup) the probe reversal results in overestimating the input power.
But when the triac conduction angle is set for high regimes (as in the experiment run), the probe reversal results in underestimating the input power by a factor 2.7, close to the claimed COP.
Conclusion: the discrepancies in the published current waveforms, unless otherwise explained, hint to a gross mistake in misplacing a clamp ammeter as a plausible explanation for the apparent energy gain in both TPR1 and TPR2. The authors, for their own sake and that of the scientific community, are invited to provide sufficient evidence (additional photos, raw data, non-ambiguous explanations) that such an occurrence can be excluded.
(Some of these questions may roughly coincide with the ones asked by "Raman", only partly leaked on the Cobraf forum, and hastily answered to by Rossi on JONP, however I believe they deserve a more careful reply).
The TPR2 on page 14 computes thermal dissipation of the copper wires by Joule effect during the dummy run as 6.7W when the power input is 486W.
This dissipation is computed starting from measured "average current" and from estimated resistance of the copper wires, as R*(I^2).
Q1.: "Average current" is understood to be a "Average root mean square current". Please acknowledge.
Then during the actual run, the copper wires dissipation is stated in table 7 in page 22 to be around 37W for total input power around 800W, and increasing to around 42W for total input power around 920 W.
Throughout the ten recording files the ratio of Joule heating of wires (in W) to total consumption (in W) is 4.6% and remains exactly the same in the two phases at 1250°C and 1400°C.
The ratio during the dummy run is instead 1.4%.
This is a very notable finding in that the Inconel resistors seem to undergo a huge shift in reactance from the dummy test to the active run.
However the reactance stays then exactly the same over 1250 to 1400°C and over 786W to 924W.
This seems an artefact of calculations rather than a real measurement, but the authors can probably better explain.
Q2.: Please confirm that total consumption is a direct reading of the PCE-830 instrument, or else explain how it is computed.
Q3.: Please confirm that Joule heating is computed from rms current measured directly by the PCE-830, or else explain how it is computed.
If both the above are confirmed, the shift in reactance is a phenomenon to be understood and explained.
Q4.:Do you expect sudden drop in Inconel ohmic resistance due to temperature close to melting point ?
Q5.:Do you hypotize magnetic coupling to elements in the reactor (though hard to imagine interaction with proton dynamics at 50Hz and few tens of Ampere) ?
On page 1 it is stated "In addition, the resistor coils are fed with some specific electromagnetic pulses".
Q6.: Is this an information given by Rossi, and/or have you detected such pulses, expectedly at much higher frequency than the PCE can measure?
Q7.: Do you believe the electromagnetic pulses fed to the resistor coils draw a substantial portion of the 360W consumed by the control box ?
Finally, on page 6 fig. 5 the PCE-830 reading is OL as in "overload".
Q8.: Please explain why the instrument was in this condition and why you chose to publish it as such, considering the obvious suspicions it would raise.
Q9.: Please publish a reading of the downstream and upstream PCE-830s in representative phases of the test, with readable scales, or if not available please justify the validity of recordings in overload condition.
Thank you,
Respectfully,
Andrea S
