[UPDATED 1-5] On November 24, 2017, Andrea Rossi held a demonstration of his third generation E-Cat reactor, the E-Cat QX, in Stockholm, Sweden, for an invited group of about 80 people. Here are my reflections on the event, in my role as organizer and presenter. 

At the end of July, Rossi told me he had decided to hold a demo of the E-Cat QX in Stockholm at the end of October, and he asked me if I would take the role as the presenter at the event. I accepted on the condition that I would not be responsible for overseeing the measurements (which were instead overseen by Eng. William S. Hurley, with a background working in nuclear plants and at refineries). Later, the event was postponed to Nov 24 and as the event came closer, my role came to include also practical arrangements for the venue etc.

Although I would not oversee the measurements, I wanted to make sure that the test procedure was designed in a way that would give a minimum of relevant information. From my point of view, already from the start, it was clear that the demo would not be a transparent scientific experiment with all details provided, but precisely a demonstration by an inventor who decided what kind of details to disclose. However, to make it meaningful, a minimum of values and measurements had to be shown.

On one hand, I may think that it’s unfortunate that Rossi chooses to avoid some important measurements, fearing that they would reveal too much information to competitors. On the other hand, I may understand him, provided that he moves along quickly to get a product to market, which seems to be his intention at this point.

As for the technical details of the demo, I describe them briefly at the end of this post, along with the discussions I had with Rossi up to the event.

Eventually, the event was held at a conference center in central Stockholm and went off fairly smoothly. I was happy to see an interested audience follow the program, from my own introduction to the demo and eventually a theory in progress presented by Carl-Oscar Gullström, nuclear physicist, focusing on meson physics.

A video recording of the event, except for the theory presentation, can be found here:

Here’s also a pdf version of my slides, and a pdf from the theory presentation by Carl-Oscar Gullström.

And here’s the video recording of Gullström’s presentation. It was fairly challenging for large parts of the audience I would say, including myself, although I, in the end, tried to make a very high-level abstraction of the contents which I had discussed with Gullström earlier.

[Update 5]: Note—at the demo, as registered in the video above, Rossi several times states that the dimensions of the plasma inside the E-Cat QX are ‘0.08 x 0.6 mm.’ (49:20, 1:33:20, 1:57:23). However, Rossi later recognized that he was mistaken about the unit that should be cm. In other words, Rossi’s claim is that the plasma inside the reactor has the form of a cylinder with the diameter 0.8 mm and the length 6 mm [end update].

To sum up the demo, there were several details that were discussed, from the problematic electrical measurement to observations of Rossi touching something inside the control system just before an additional measurement was being made (see below). [Update 1]: It was also noted that the temperature of the incoming water was measured before the pump and that the pump could possibly add heat. However, the temperature did not raise at the beginning of the demo when only the pump was operating and not the reactor. Rossi also gave the pump to me after the demo so that I could dismantle it (which I have now done—see update 4 below)), together with a wooden block where a 1-ohm resistance was mounted, which he also advised me to cut through (also done—see below). [End update].

In the end, I found that there were reasonable explanations for everything that occurred, and the result indicated a clear thermal output with a very small electrical input from the control system.

However, if I were an investor considering to invest in this technology, I would require further private tests being made with accurate measurements made by third-party experts, specifically regarding the electrical input power, making such tests in a way that these experts would consider to be relevant. (See also UPDATE 3 on electrical power measurement below).

§§§

Technical details of the demo:

The E-Cat QX is a small reactor which like earlier E-Cat reactors presumably is a heat generator based on a Low Energy Nuclear Reaction—LENR—with an energy density similar to other nuclear reactions, releasing up to 1 million times more energy per gram fuel than chemical reactions, but without strong radiation, which is the common disadvantage with known nuclear reactions.

The claims of the E-Cat QX are:

– volume ≈ 1 cm³
– thermal output 10-30 W
– negligible input control power
– internal temperature > 2,600° C
– no radiation above background

– at the demo, a cluster of three reactors was tested.

The test procedure contained two parts—thermal output power and electrical input power from the control system—essentially a black box with an unknown design, connected to the grid.

Measuring the thermal output power was fairly straightforward: Water was pumped from a vessel with cold water, flowing into a heat exchanger around the E-Cat QX reactor, being heated without boiling, and then flowing into a vessel where the total amount of water was weighed using a digital scale.

A second method for determining the output power was planned—measuring the radiated light spectrum from the reactor, using Wien’s Displacement Law to determine the temperature inside the reactor from the wavelength with the maximum intensity in the spectrum, and then, Stefan-Boltzmann Law for calculating the radiated power from the temperature.

These two results would be compared to each other at the demo, but unfortunately, the second method didn’t work well under the conditions at the demo, with too much light disturbing the measurement.

The method for measuring electrical input power was more problematic. The total consumption of the control system could not be used, since the system, according to Rossi, was using active cooling to reduce overheating inside, due to a complex electrical design. [Update 4]: One hypothesis for the overheating issue is that the reactor produces an electrical feedback that will be dissipated inside the control system and has to be cooled [end update]. At this point of R&D of the system, the total energy consumption of the system is therefore at the same order of magnitude as the released amount of energy from the reactor, and it, therefore, makes no sense to measure the consumption of the control system. Obviously, this must be solved, making a control system which is optimised, in order to achieve a commercially viable product.

Instead, the aim was to measure the power consumption of the reactor itself. Using Joule’s law (P=UI), electrical power is calculated multiplying voltage across some device with the current flowing through the device. However, Rossi didn’t want to measure the voltage across the reactor, claiming that it would reveal sensible information.

He would measure the current by putting a 1-ohm resistance in series with the reactor and measuring the voltage across the resistance with an oscilloscope, then calculate the current from Ohm’s law (U=RI), dividing the voltage by the resistance (being 1 ohm). Accepting to use an oscilloscope was good since this would expose the waveform, and also because strange waveforms and high frequencies would make measurements with an ordinary voltmeter not reliable.

But, as mentioned, knowing the current is not enough. Rossi’s claim was that when operating, the reactor had a plasma inside with a resistance similar to that of an ordinary conductor—close to zero. Electrically this means that the reactor would use a negligible amount of power, but it was just an assumption and I wanted to make it credible through other measurements.

My suggestion, which Rossi accepted, was to eliminate the reactor after the active run, replacing it first with a conductor, then with a resistance of about 800 ohms as a dummy, to see how the control system behaved. The conductor should provide a similar measurement value as with the reactor if the reactor behaved as a conductor. Using the 800-ohm resistance, on the other hand, should show whether the control system would possibly maintain the measured current, expected to be around 0.25A, with a higher resistance in the circuit. At 0.25A, a resistance of 800 ohms would consume about 50W, which would be dissipated as heat, and this could then explain the produced heat in the reactor without any reaction, just from electric heating.

[UPDATE 3]: I now think I understand why Rossi wouldn’t let us measure the voltage across the reactor. Rossi has described the E-Cat QX as two nickel electrodes with some distance between them, with the fuel inside, and that when the reactor is in operation, a plasma is formed between the electrodes. Most observers have concluded that a high voltage pulse of maybe 1kV is required to form the plasma. Once the plasma is formed the resistance should decrease to almost zero and the control voltage immediately has to be reduced to a low value. Normally, and as claimed by Rossi, the plasma would have a resistance as that of a conductor, and the voltage across the reactor will then be much lower than the voltage across the 1-ohm resistor (measured to about 0.3V—see below). Measuring the voltage across the reactor will, therefore, be difficult: The high voltage pulse risks destroying normal voltmeters and measuring the voltage with an oscilloscope will be challenging since you first have to capture the high voltage pulse at probably 1 kilovolt and then immediately after you would need to measure a voltage of maybe millivolts. [end update]

At the demo, 1,000 grams of water was heated 20 degrees Celsius in one hour, meaning that the total energy released was 1,000 x 20 x 4.18 = 83,600J and the thermal power 83,600/3600 ≈ 23W.

The voltage across the 1-ohm resistor was about 0.3V (pulsed DC voltage at about 100kHz frequency), thus the current 0.3A. The power consumed by the resistor was then about 0.09W and if the reactor behaved as a conductor its power consumption would be much less.

Using a conductor as a dummy, the voltage across the 1-ohm resistance was about 0.4V, thus similar as with the reactor in the circuit. With the 800-ohm resistance, the voltage across the 1-ohm resistance was about 0.02V and the current thus about 0.02A. The power consumption of the 800-ohm resistance was then 0.02 x 0.02 x 800 ≈ 0.3W, thus much lower than the thermal power released by the reactor.

These dummy measurements can be interpreted in a series of ways, giving a COP (output power/input power) ranging from about 40 to tens of thousands. Unfortunately, no precise answer can be given regarding the COP with this method, but even counting the lowest estimate, it’s very high, indicating a power source that produces useful thermal power with a very small input power for controlling the system.

At the demo, as seen in the video recording, Rossi was adjusting something inside the control system just before making the dummy measurements. Obviously, someone could wonder if he was changing the system in order to obtain a desired measured value.

His own answer was that he was opening an air intake after two hours of operation since the active cooling was not operating when the system was turned off.

[Update 2]: Someone also saw Rossi touch a second switch close to the main switch used for turning on and off the system. Rossi explained that there were actually two main switches—one for the main circuit and one for the active cooling system—and that there were also other controls that he couldn’t explain in detail. [end update].

Clearly this comes down to a question of trust, and personally, discussing this detail with Rossi for some time, I have come to the conclusion that his explanation is reasonable and trustworthy.

However, as I stated above, if I were an investor considering to invest in this technology, I would require further private tests being made with accurate measurements made by third-party experts, specifically regarding the electrical input power, making such tests in a way that these experts would consider to be relevant.

§§§

Here below is the test report by William S. Hurley, as I received it from Rossi:

DATA REPORT OF THE MEASUREMENTS MADE ON NOVEMBER 24TH 2017 ON THE E-CAT QUARKX TESTED AT THE IVA, GREV TUREGATAN 16, STOCKHOLM, SWEDEN.

Duration of the measurement period: 1 hour: the measurement has been made after the apparatus has reached a reasonably constant temperature
amount of water pumped through the reactor: 1 000 g
Water temperature at the input of the reactor: 21 C
Water temperature at the output of the reactor: 41 C
Delta T: 20 C
Energy produced:  20 x 1.14 = 22.8 Wh/h
Measurement of the energy consumed  ( during the hour for 30′ no energy has been supplied to the E-Cat) :
V: 0.3
OHM: 1
A: 0.3
Wh/h 0.09/2= 0.045
Ratio between Energy Produced and energy consumed: 22.8/0.045 = 506.66

Instrumentation used for the measurements:
Oscilloscope Tektronix TBS 1052B
K probes Omega supplied and calibrated by Prof. Bo Hoistad of the University of Uppsala
Water pump Prominent. The water pumped for 1 hour has been  poured in a plastic container seat on a scale to measure exactly the water passed through the E-Cat.
Temperature Data Logger: PICO Technology
The scale to weight the water passed through the E-Cat has been supplied by Eng. Mats Lewan of Stockolm

William S. Hurley
Senior Engineer- Endeavor
Los Angeles

§§§

Update 4:

I have now dismantled the pump and I found no hidden heaters or other modifications:

Pump nameplate:

I also cut through the plastic block on which the 1-ohm resistor, and later also the 800-ohm resistor, were mounted and found no hidden devices or energy sources inside: