The August 2012 Penon Hot-Cat report

Given that this report was the milestone that persuaded [lexicon]IH[/lexicon] to release $1.5M I thought I should look at it in detail. I have to say I have not done this before. These reports often go on giving with detailed study so this is just a preliminary comment.

The report was released by Rossi together with some raw data (the temperature and calculations, but not the input power) and a large set of corrections. I'm not sure who wrote the corrections, it looks like Rossi?

Anyway there are several issues in the report, not all of which I have time for here. The most obvious, corrected in Rossi's corrections, is that the radiated power is estimated to be the sum of that from outer and inner concentric cylinders. Of course that is weird. The outer cylinder surface radiates. All other radiation internally is reabsorbed, except for a small amount from the end of the device whether the inner cylinder is visible.

This is a very obvious mistake and something a competent (for this task) person could not have made. It overestimates power. However, it is corrected in the set of additional corrections published with the report. I guess this is why some people view Penon as unsuitable for the job of ERV.

There is one problem that for me beats all of the others, I'm not sure if it has been noted:

Quote from Penon

E-Cat power supply was effected through a control box panel provided with a kWh meter which did not allow separate evaluation of the voltage and current supplied to the module.For this reason, a voltmeter and a clamp ammeter were installed downstream from the control box, so as to monitor power data independently from the panel meter. Due to the fact that panel meter data were found to be quite discordant from those measured by the voltmeter and the ammeter, it was decided to ignore the former and use only the voltmeter and ammeter data recorded manually in the course of the test. Voltage was gradually increased step by step to higher values as E-Cat temperatures were shown to stabilize.

If this means what I think it means it is very bad news. There are a number of reasons why the voltage and current data (multiplied together) may be different from the actual delivered power and therefore the kWh integrates input power. But equally the kWh meter is uncalibrated and could be all wrong. The fact that the two readings are very different shows a problem. You'd expect only a small difference due to the control box dissipated power.

Here is why the actual power delivered will be possibly be larger than the power calculated by Penon from V and I (the A reading):

(1) If the V and A meters are average RMS (not true RMS, which requires internal processing) the pulse nature of the triac output from the control box, which has a high crest ratio (= is very spiky) will lead the delivered power to be larger than the average voltage X average current. With, for example, a square wave at 1/n duty cycle the average V X average I scales as 1/n^2 but the RMS power scales as 1/n. So you expect an underestimation of input power by a factor of n/2 (the half comes from the fact that average meters compensate to generate true power for sine waves). For triac control n can be anything up to 20 or even more so that is a power error of up to 10X.

(2) Even true RMS V & A meters can go wrong. If there is a high crest value in the waveform the V or A meters could saturate even when they are nominally in range, so again giving an underestimate of the true input power.

Penon has:

• Not specified the equipment used (so we can never evaluate whether (1) or (2) apply)
• Not specified how the equipment used was calibrated (I guess it was not)
• Not checked either (1) or (2)
• Not investigated the discrepancy he notes between power and V X I
• Taken the likely LESS accurate figure (V X I).

I have sympathy with 5. Using the kWh meter would be unsatisfactory because it was provided by Rossi and could have been very badly calibrated. However personally I think it more likely that it was OK, and either (1) or (2) caused a large underestimate of the input power delivered. (Although, for completeness, we don't actually know who provided the V and A meters. That could well have been Rossi).

So my take home is this:

We have a definite error mechanism for this test which would underestimate input power and fits the data quantitatively - it could easy generate the claimed data (average COP = 2.2) from an electric heater,

• Penon is grossly innappropriate to do this test (and therefore also the one year test, even at a most basic level). He is not aware of basic issues of test protocol, of electric theory, of radiative power calculation. The danger with such a person is that they will rely on Rossi's setup and ideas, and not be able to detect mistakes in them. We know from Lugano that mistakes can happen.
• [lexicon]IH[/lexicon] were either badly advised, or did not realise the importance of rigor, when evaluating this test. It is badly done at a most basic level and this can be shown from the report.

Perhaps all this stuff is known to everyone - I know the Penon report was not well received at the time.

Am I being unreasonably hard on Penon or [lexicon]IH[/lexicon]?

Obviously the plethora of ECW rationalisations for why [lexicon]IH[/lexicon] want to break the contract and be dishonest even though Rossi's stuff works can remain, I don't think they can ever be ruled out and if your assumption is that Rossi's stuff must work, then you are drawn to them. I guess though that if (as one poster on ECW claimed) Darden & Vaughn are both committed Christians that make sense of their "save the world with LENR" stance and also makes deceit from them a bit less likely.

(I have not dealt with the heat after death claims - these are very speculative - I'll do so below if anyone feels they prove extraordinary behaviour).

Do we have a report for the 24 hour test that released $10M?

@Paradigmnoia

That is a separate issue and not actually a problem. Whatever R is, true RMS VXI will give the correct power - if we assume the system is pure resistive which in this case is probably fair.

The problem, due to average rather than RMS error (1) or due to meter saturation at high crest ratio (2) is not knowable given the info we possess. Nor is it known whether the errors will be a constant multiple, or vary with power (it depends on the contral box algorithm).

Quote

Maybe Rossi had his wattmeter calibrated properly. Possibly it read the cleaner inlet power. It would have been nice to have known what it read compared to the downstream meter.

All we can be sure is that the VXI measurement was significantly different from the input power measurement.

The report is saying - hey, look! This result is inaccurate! But I'm not going to tell you how inaccurate!

Re versions: well this must be a minefield. The original report had many errors and Rossi took it under his wing and published a version, with corrections. I was reading the ECW (almost synonymous with Rossi) publiched for the August 2012 test (the second of two tests by Penon).

Re Errors. One error I did not mention is the classic thermography one. The surface measured is black painted steel using a special black paint Rossi obtained (Rossi says it is special).

Now black paint has opacity and emissivity (both are needed) from its pigments. It is a thin layer, so if it stops being opaque the spectral charcteristics are dominated by the low emissivity metal underneath.

What matters for our purposes is the difference between its response at 1.5-3u (the Planck peak at the relevant temperatures where most of the power is radiated) and its response at 7-13u - which is what the camera receives.

It could go either way and give a positive or negative error. We just don't know how much this is. It is quite possible because pigments are chosen to have specific properties in the visible range and what they do outside that is variable.

See for example: http://qirt.gel.ulaval.ca/arch…t2008/papers/15_08_12.pdf which is not very helpful but does show how variable the spectral emissivity of thin paint films over metal can be - because if the paint becomes transparent at any wavelength the metal low emissivity shines through - so to speak!

Re heat after death.

This is definitely knocked on the head. Rossi claims:

• Temperature drops from 1074 to 572 K in 34 minutes
• Still 299 C at end of measurments
• Energy released after death1460 Wh
• Needed Cp =1460*3.6/(4.4*502)=2.4 kJ/kgK
• Steel 450-700 J/kgK, Ceramics 700-900 J/kgK

However, obviously we have a X2.2 error to account for the average COP (if not you don't need heat after death to prove excess heat).

That gets us down to 1.1kJ/kgK. That is still a bit more heat storage than it appears (from Rossi's data) is possible.

Two factors make up the difference:
(1) heat capacity is not constant with temperature. Rossi has given the room temperature figures. At the temperatures here ceramics, in particular, have a significantly higher heat capacity
(2) Rossi is using external temperatures. However the internal temperature will be significantly higher (consider the conductive gradient) - also the resistors will not at the end of the test necessarily be in equilibrium with the outside temperature - you can see for most of the test this is not true indicating a higher resistor/surface deltaT than would be expected just from conductive losses.

So the "heat after death" argument is completely wrong - the data here is exactly compatible with the data for the rest of the test given no heat after death.

That Rossi should need to invoke this (bogus) heat after death argument shows I guess that he is aware that the actual test data with COP of X2.2 is not convincing.

Of course, the actual test data could have been convincing quite easily:

(1) Measure input power robustly with specified calibrated equipment capable of measuring true RMS and resolving very spiky waveforms (which many cheap digital meters are not). This would probably resolve the discrepancy that Penon notes, but for the test to be convincing this must be documented.

(2) Use a K type thermocouple to get an independent measurement of surface temperature. Errors here would tend towards a lower temperature and hence would not matter, so this is a pretty robust procedure.

So if you look at a true RMS good quality modern DVM:

You have to delve deep in the spec to find it, but we have:
– The maximum circuit limitation of peak signal is 300 % typically for voltage and current except 250 % and 150 % typically for range 600 V and 1000 V respectively.

And: the bandwidth is 30kHz.

A triode switched waveform will possibly have components significantly above 30kHz, and (more important I think) the crest factor limitation could lead to large under-reading.

With this type of circuit all of this needs to be checked carefully - which is probably why later experiments used a PCE-830 with better characteristics.

Quote from Paradigmnoia

I saw what looked like the same DVM (as in the Penon 2012 report) in an old photo of Mats Lewan pointing at an amp clamp, next to a blue box controller.

Can you identify which DVM it is?