The following guest post was submitted by Michael Lammert.
Requirements of H-Ni LENR Devices and Their Implications on the Lugano Test
by Michael Lammert (AKA Dr. Mike)
October 31, 2014
I’ve continued to think about the results of the Lugano test and Rossi’s E-Cat devices. Two questions that I’ve been considering are: What factors are required to produce an optimum H-Ni LENR device (ala Rossi’s E-Cat) and are the active “consumption” powers in the Lugano test reasonable and correct?
As I see it, key requirements for a useful E-Cat type device include:
- Ability to generate enough heat within the Ni powder to produce a useful output.
- Ability to control the low energy nuclear reactions sufficiently to provide stable operation without thermal runaway.
- Ability to deliver the heat to a load upon demand.
- Ability to dissipate the heat from the Ni to its surroundings fast enough to prevent the Ni from melting.
Although we don’t know yet what exact combination of Ni powder, hydrogen, catalysts, and electromagnetic pulses are needed to produce a working LENR device, there doesn’t appear to be any question that an E-cat device can produce sufficient heat to be useful. Not only is Rossi currently working hard to get his 1 MW plant installed, others have independently demonstrated useful heat from LENR devices.
The control of an LENR device for stable operation is going to be a difficult, but not impossible task, at least until the theory of operation is well understood. The ability to deliver heat to a load upon demand should just be a straight forward engineering problem, except for the control function. Therefore, both items #2 and #3 will benefit greatly when a solid theory of LENR is established.
I believe that getting the heat generated in the Ni dissipated to its surroundings will be the primary limiting factor for how much power can be generated from a given size device. Ni has a specific heat of 0.44J/gr/oK, which means that 0.44 Joules of energy will raise the temperature of 1 gram of thermally isolated Ni by 1 oK. It would only take about 520 Joules (1 Joule = 1W-sec) to raise that 1 gram of thermally isolated Ni from room temperature to its melting point (0.44*(1455-273) = 520). Luckily the Ni powder in the E-Cats is not thermally isolated. In the Hot-Cat, the Ni is in contact with both other Ni particles and the alumina reactor. Both Ni and alumina have reasonable thermal conductivities (91 and 30 W/m/oK, respectively). The Ni will also radiate heat proportionally to T4, and also have heat transferred via convection to the mostly hydrogen gas within the reactor. (Although there is no data on what the partial pressure of hydrogen is during operation, cooling through transfer of energy to the hydrogen might be made somewhat more effective by removing most of the air within the reactor prior to start-up since hydrogen has about a 7 times higher heat conductivity than air.)
How fast can heat be dissipated from the Ni powder through conduction, radiation and convection? For the answer to this question I would request a thermal engineer to make a calculation based on assumptions that are appropriate to the operating conditions of the Hot-Cat. The closest electrical analogy that I found to this thermal dissipation problem is data on the internet for temperature vs. current data for NiCr 60 wire (60%Ni). A 1 gram piece of 18 gauge (1 mm diameter) NiCr 60 wire would be about 15.5cm long and have a high temperature resistance of about 0.24 ohms. The NiCr would melt at a current of about 33 amps or at a power of about 260W (332*.024). (This is for a straight wire in air.) I would expect a gram of Ni powder sitting on alumina in a hydrogen environment to have a somewhat higher maximum dissipation power than the 260W for a piece of NiCr wire. However, if the Ni was clumped in a pile, it might be expected to melt at even a lower power level than the 260W. The solution to the dissipation problem is to add more Ni powder, distribute the powder in a uniform layer over the entire reactor internal surface, and run the reactor at a lower power generation rate per gram of Ni powder.
Now let’s look at the numbers from the Lugano test. It was claimed that 0.55 grams of Ni was used in the reactor with average output powers of about 1660W for the first part of the run and about 2320W for the second part. Assuming all of this excess power is generated uniformly in the 0.55 grams of Ni powder, dissipation in the Ni would have to be 3018W/gr in the first run and 4018W/gr in the second portion. Although I will wait to see a calculation of the maximum possible dissipation per gram of Ni powder from a thermal engineer before making a final conclusion, I believe these dissipation numbers are about an order of magnitude too high for the 0.55 gram of Ni powder not to melt.
If it is assumed that it is not probable that the measured excess power seen in the Lugano test could have come from heat generation in the small amount of Ni, where did the heat generation come from? In my previous post I pointed out the problem with the “Joule heating” calculations, reprinted here:
- Problem with the “Joule Heating” Calculation
The “Joule heating” calculation for the Cu wire for the dummy run on pages 13-14 seems to be fairly straight forward. The “Joule heating” is simply the resistance of the wire times the current squared flowing through that wire. Sum the Joule heating in the 3 Cu wires from the controller and the 6 Cu wires to the device and you have the power that comes out of controller, but doesn’t participate in heating the Inconel coils. This is such a simple calculation, that it seems unlikely that an error would be made in other calculations of Joule heating. However, the “Joule heating” in the Cu wires for the active run has been calculated in Table 7, page 22 as about 37W for the input power at 800W and about 42W for the operation at 920W. These “Joule heating” calculations imply that the current in the Cu wires was 2.35 times as high in the 800W active run as it was in the dummy run (SQRT(37/6.7) = 2.35). The only way for this to be possible is for the Inconel resistors to have a very large negative temperature coefficient of resistance. Although the report did not specify what type of Inconel was used in the coils, the data sheets for various Inconels show well less than 10% variation in resistivity over a wide temperature range. For example, Inconel 625 has a resistivity of 135.9 micro-ohm-cm at 427 oC and 133.9 micro-ohm-cm at 1093 oC. Other Inconels have a slightly increasing resistivity as the temperature increases. Also it should be pointed out that if the Inconel used in the coils in this experiment had a large negative TCR, then the Joule heating as calculated in Table 7 would have been much higher than 42W for the 900W portion of the test. The calculated “Joule heating” powers are directly proportional to the “consumption” powers, indicating no change in resistivity of the Inconel coils as the temperature increases from about 1260 oC to 1400 oC in the two portions of the active runs. Questions for the authors: 1. What is the source of the error in the “Joule heating” calculation for the active run? 2. What type of Inconel was used in the resistor coils? 3. What was the current flowing through the resistors for each of the active power levels?
In re-reading the Lugano report I found that the authors actually answered my third question on the current levels on the active power: On page 3 they say that the potentiometer on the TRIAC power regulator was used to set the operating point, “normally 40-50 Amps”. Using the data in Table 7, page 22 for the Joule heating numbers, the average current for File #2 can be calculated as I2 = 19.5* SQRT(36.98/6.7) = 45.81A (where 19.5 is the current in the dummy run and 6.7 is the Joule heating in the dummy run), and the average current for File #7 can be calculated as I7 = 19.5* SQRT(42.18/6.7) = 48.93A. Indeed both of these currents are within the 40-50A range quoted by the authors on page 3. Since the resistance of the Inconel coils is known to change by less than +/- 5% over the operating temperature range (look this up if you don’t believe it) the input power for the active runs will equal to the dummy power times the ratio of the Joule heating powers. For File #2 the actual power supplied by the TRIAC regulator (“consumption” power in Table 7) is P2= 486*36.98/6.7 = 2682W (+/-5%) (486 is the power for the dummy run), and for File #7 the actual power supplied by the TRIAC regulator is P7 = 486*42.18/6.7 =3060W (+/-5%). (The “Joule heating” powers must be subtracted from these values to get the actual power delivered to the Inconel wire resistors.) Using these calculated “consumption” powers results in File #2 having a net power production of -217W with a COP of .918 and File #7 having a net power production of 180W with a COP of 1.06. If these numbers are correct, one can see that the power output due to the LENR effect is in the noise of the measurement technique. However, the analysis of the “fuel” and the “ash” clearly show that nuclear reactions took place within the reactor during the active runs, but perhaps at a power level in the low 100’s of Watts.
My speculation as to a possible source of an error in the set-up is that the connection of the power source supplying the “specific electromagnetic pulses” mentioned on page 1 of the report is somehow interfering with getting an accurate power measurement from the PCE 830 meter on the output of the TRIAC power regulator. There is no electrical diagram showing how this power source is wired to the reactor and the TRIAC power regulator, and the authors did not state whether this power supply was turned on during the dummy run. I would further speculate that the authors measured the higher current and relatively low power during the active runs and assumed that Inconel resistor coils just had a large negative temperature coefficient of resistance. (This is what I assumed until I looked up the TCR for the various types of Inconel.)
It would certainly be disappointing if the results of the Lugano test are clouded by an “instrumentation” error; however, it wouldn’t be the first time and won’t be the last time that the measurement technique caused an error in reported scientific results. In 2011 the scientists working on the OPERA project at Italy’s INFN Gran Sasso Laboratory reported that they had measured neutrinos that were traveling slightly faster than the speed of light. It took about 6 months with help from scientists outside their laboratory to finally confirm that they had an error in their measurement technique.
My recommendations to the Lugano authors and others working on E-Cat experiments are as follows:
- Have a thermal engineer estimate the maximum power dissipation rate for 1 gram of Ni in the Hot-Cat reactor based on assumptions that match the operating conditions for the Lugano test.
- Have the electrical set-up of the Lugano test reviewed by an electrical engineer to determine if the power source supplying the “specific electromagnetic pulses” could have interfered with making an accurate power measurement during the active runs.
- Re-examine the radial uniformity of the reactor temperature data. If the Ni powder was mostly at the bottom of the reactor, I would expect to see at least a 100 oC temperature difference between the bottom of the reactor and the top of the reactor (needs verified by a thermal engineer’s calculation). If the radial temperature uniformity is about the same in the active runs as the dummy run, then there is a good chance that most of the observed output power is coming from input power to the Inconel coils, not from LENR in the Ni powder.
- For future Hot-Cat experiments it might be beneficial to supply the “specific electromagnetic pulses” via a fourth coil wound on the reactor.
- The MFMP group would have a much higher chance of success if they tried to duplicate the original E-Cat, rather than the Hot-Cat. If they decide to go ahead and build a copy of the Hot-Cat, they should probably follow my recommendation #4 above and load the reactor with a minimum of 5-10 grams of “fuel”.