New Experiment From me356 With New Reactor Design Using Nickel Wire, LAH (Test Restarted, Oct 16th))

Following the experiment the Rossi patent recipe, which did not produce excess heat, me356 has started a new experiment using a reactor of a different design which uses nickel wire as the nickel source, and LiAlH4 (lithium aluminum hydride) as the hydrogen source.

UPDATE: Thanks to Barty for posting that me356 has restarted his test today. Live data can be followed here: http://www.lenr-experiment.tk/

Me356 is making some comments on the LENR-Forum on this thread: http://www.lenr-forum.com/forum/index.php/Thread/2139-Tube-Reactor-design/?postID=8544#post8544

Here are some pictures from the MFMP Facebook page here:

me356
me356a

On the LENR-Forum here, me356 writes:

“I am doing experiment right now, it is running for few hours and it looks that there is excess heat for long time.

With increasing temperature/pressure excess heat is higher. This is second time that I have achieved it with very similar results.

Unfortunately COP at the moment is roughly 1,1 only so still we can account it to a various factors.
It is very stable and pressure is holding well.”

The experiment can be followed live at the link below.

http://www.lenr-experiment.tk/

  • Gerard McEk

    ME356, I assume that your remark about excess heat is based on a previous test without NiAlH4? Is the pressure equivalent with the pressure now?
    Maybe next test with additional lithium?

    • Axil Axil

      In the Lugano test, the addition lithium had its origin in the fuel preparation phase, not during the test run.

      • Gerard McEk

        That is indeed what I have assumed, Axil. The only question is where it was hided. I cannot imagen that it was in the fuel itself, because that was analysed. The other possibility is that it was evaporated at the inner surface of the reactor tube. But now we know that lithium ‘eats’ the Al2O3 tube, so where was it? Maybe the inner tube was first covered (sputtered) with an iron layer and after that with a lithium layer?

        • Axil Axil

          At high temperatures, lithium eats everything except diamond coating. Nickel will alloy with lithium.

  • Gerard McEk

    ME356, I assume that your remark about excess heat is based on a previous test without NiAlH4? Is the pressure equivalent with the pressure now?
    Maybe next test with additional lithium?

    • Axil Axil

      In the Lugano test, the addition lithium had its origin in the fuel preparation phase, not during the test run.

      • Gerard McEk

        That is indeed what I have assumed, Axil. The only question is where it was hided. I cannot imagen that it was in the fuel itself, because that was analysed. The other possibility is that it was evaporated at the inner surface of the reactor tube. But now we know that lithium ‘eats’ the Al2O3 tube, so where was it? Maybe the inner tube was first covered (sputtered) with an iron layer and after that with a lithium layer?

        • Axil Axil

          At high temperatures, lithium eats everything except diamond coating. Nickel will alloy with lithium. It completely covered the fuel and became part of the nickel fuel particles.

  • Bob Greenyer
  • Bob Greenyer
  • Jarea

    Wow!, congratulations! This is big news!
    Unfortunately, the COP is too tight, are we sure that the error tolerance is below that?

  • Jarea

    Wow!, congratulations! This is big news!
    Unfortunately, the COP is too near 1, are we sure that the error tolerance is below that 0.1 (10%)?

  • Bob Greenyer

    This type of reactor (with a range of fuels) needs to be run in a calorimeter or Parkhomov-style boiling water reactor.

    At the moment the calibration and live run data are taken from the hottest spot on the Optris PI-160

    At the end of the run – and average over the hot zone/reactor for calibration and live run will give a more accurate impression.

    The great thing about the design is that it can be assembled in 30 mins and runs hot with relatively low input power. Also, Like Celani’s cell, the powered wire is in the H2 and the outside is nice and smooth/regular allowing for simple Stefan-Boltzmann type determination of power.

    • Sanjeev

      Its a very good news that the COP is above 1, however we need to be careful.
      I think this becomes a good candidate for calorimetry.

      If, as you say, its easy to assemble or cheap, we can have 3-4 of these rods in the calorimeter to get a higher COP and better error margins.

      • Bob Greenyer

        I spoke to Mathieu Valat earlier today and suggested a similar experiment, which we have been planning with Celani wire + LiAlH4, be one of the first to be conducted in the MFMPs MFC. This desire has been accelerated by recent results that Jean-Paul Biberian has had with another group ten element in his MFC that should be reported at the Tolouse conference.

        • Sanjeev

          Great. I will look forward to the conf.
          However, I was thinking that Brian is all set with his state of the art MFC, so perhaps one such rod can be sent to him or he can also make one. Probably he can start such mashup experiment in a day or two, because all is ready there.

          • Bob Greenyer

            He needs the inner rod – but Brian should be able to do this quite fast.

          • Sanjeev

            Hopefully he will.

  • Bob Greenyer

    This type of reactor (with a range of fuels) needs to be run in a calorimeter or Parkhomov-style boiling water reactor.

    At the moment the calibration and live run data are taken from the hottest spot on the Optris PI-160

    At the end of the run – and average over the hot zone/reactor for calibration and live run will give a more accurate impression.

    The great thing about the design is that it can be assembled in 30 mins and runs hot with relatively low input power. Also, Like Celani’s cell, the powered wire is in the H2 and the outside is nice and smooth/regular allowing for simple Stefan-Boltzmann type determination of power.

    • Sanjeev

      Its a very good news that the COP is above 1, however we need to be careful.
      I think this becomes a good candidate for calorimetry.

      If, as you say, its easy to assemble or cheap, we can have 3-4 of these rods in the calorimeter to get a higher COP and better error margins.

      • Bob Greenyer

        I spoke to Mathieu Valat earlier today and suggested a similar experiment, which we have been planning with Celani wire + LiAlH4, be one of the first to be conducted in the MFMPs MFC.

        • Sanjeev

          Great. I will look forward to the conf.
          However, I was thinking that Brian is all set with his state of the art MFC, so perhaps one such rod can be sent to him or he can also make one. Probably he can start such mashup experiment in a day or two, because all is ready there.

          • Bob Greenyer

            He needs the inner rod – but Brian should be able to do this quite fast.

          • Sanjeev

            Hopefully he will.

    • Roger Barker

      Bob, MFMP has been trying to achieve a successful experiment for 4 years. Yet to date you have not achieved this. I applaud your efforts but have to seriously question your methods.

      What do you think is the single biggest reason you have failed thus far?

  • Ged

    Maybe one of the confounding factors for replicators all this time is the use of nickel that is Too pure. Per the NAVSEA info, and since Celani also has to treat (aka dope) his wires, perhaps all these 99% pure nickel mixtures are dampening the reaction strength and probability, leaving the reaction for when temps are in a narrow range were lithium can distort the lattice–and that would be limited in effect.

    • Sanjeev

      I agree Ged, nothing interesting in a perfect lattice of pure Ni. IF DeChiaro is right, we need defects in the crystals where action takes place. An impurity (dopant, as its called in semiconductor industry), is one way to generate such defects.

      • Ged

        It’ll take a bit of theory finding to figure out what element(s) to dope with and by how much (ratio), without access to NAVSEA’s model :(. We want a strained/deformed lattice yielding geometry inducted vacancies, but while keeping the crystal structure an FCC (face centered cubic) or BCC maybe (body centered cubic, based on reports of weak activity in Titanium). And material properties have to stay favorable towards hydrogen and compatible with our temperature range for lithium and hydrogen release.

        I wonder if semiconductor models exist we could snag access to. Also, bulk doping may yield highest results, but surface etching may still work to a high enough degree (ala Celani) by using acid or plasmolysis. This is Not my field, so I am just throwing around ideas.

        • Axil Axil

          Lithium will alloy with nickel at high temperatures and a lithium carbide will form that will release lithium Rydberg matter as well as hydrogen Rydberg matter. Carbon is important as a contaminate when combined with lithium.

          Parkhomov has a very high carbon content as a contaminate in his nickel powder. A poorly cleaned particles of carbon might be needed.

          A fuel preparation step might be needed to coat the fuel particle surface with carbon to contaminate it properly, for example, cover it a coat of graphite powder.

          • Ged

            Sounds like an easy enough thing to try.

            According to link.springer.com/article/10.1007%FBF02881419 carbon can be solid soluble in Ni up to 2.7% atomic ration while maintaining a FCC lattice for Ni. Can’t see the rest of the paper, but it may be of interest.

          • Bob Greenyer

            There is 3.8% CO2 in the air.

          • Ged

            0.038% actually (380 parts per million). 3.8% would be lethal I believe.

          • Bob Greenyer

            oops – dp issue – serve me right for being up nearly all night all weekend!

          • Ged

            All for a good cause!

          • Bob Greenyer

            for sure. Hey, thanks for the graphs again… helped us to see if we needed to go further with calibration or if a polynomial would fit and allow us to start the run faster.

          • Ged

            I’m glad I could help 🙂

            Some people like to stay up all night playing video games. Other people like to stay up all night trying to controlled light things on (nuclear) fire.

          • Bob Greenyer

            I can honestly say – I really enjoy it.

  • Ged

    Maybe one of the confounding factors for replicators all this time is the use of nickel that is Too pure. Per the NAVSEA info, and since Celani also has to treat (aka dope) his wires, perhaps all these 99% pure nickel mixtures are dampening the reaction strength and probability, leaving the reaction for when temps are in a narrow range were lithium can distort the lattice–and that would be limited in effect.

    • Sanjeev

      I agree Ged, nothing interesting in a perfect lattice of pure Ni. IF DeChiaro is right, we need defects in the crystals where action takes place. An impurity (dopant, as its called in semiconductor industry), is one way to generate such defects.

      • Ged

        It’ll take a bit of theory finding to figure out what element(s) to dope with and by how much (ratio), without access to NAVSEA’s model :(. We want a strained/deformed lattice yielding geometry induced vacancies, but while keeping the crystal structure an FCC (face centered cubic) or BCC maybe (body centered cubic, based on reports of weak activity in Titanium). And material properties have to stay favorable towards hydrogen, and compatible with our temperature range for lithium and hydrogen release.

        I wonder if semiconductor models exist we could snag access to. Also, bulk doping may yield highest results, but surface etching may still work to a high enough degree (ala Celani) by using acid or plasmolysis. This is Not my field, so I am just throwing around ideas.

        • Axil Axil

          Lithium will alloy with nickel at high temperatures and a lithium carbide will form that will release lithium Rydberg matter as well as hydrogen Rydberg matter. Carbon is important as a contaminate when combined with lithium.

          Parkhomov has a very high carbon content as a contaminate in his nickel powder. A poorly cleaned particles of carbon might be needed.

          A fuel preparation step might be needed to coat the very clean fuel particle surface with carbon to contaminate it properly, for example, cover it with a fine coat of graphite powder then heat it at high temperature to form surface imperfections. When the lithium is added, it will remove the carbon and leave a vacancy on the surface were the graphite particle had been melted into the nickel surface of the particle.

          Sputtering with graphite using vapor disposition might also be considered as a first step in the fuel preparation process,

          • Ged

            Sounds like an easy enough thing to try.

            According to http://link.springer.com/article/10.1007%2FBF02881419#page-1 carbon can be solid soluble in Ni up to a 2.7% atomic ratio while maintaining a FCC lattice for Ni. Can’t see the rest of the paper, but it may be of interest.

            Edit: Commercial “high carbon” alloys of nickel like the Alloy 800 series seem to only be a 0.2-0.5% range. Well below max. Dunno how much is enough, but in theory your idea of graphite mixing should make the surface layers take on their max 2.7% (interstitial alloy). The outer most layers may breakdown into Ni3C, but would flake from the crystal and expose the high carbon FCC nickel.

            Commercial alloy 800s may still work too if someone tries them. But those alloys also have iron and chromium (nickel-iron-chromium; somewhat related to stainless steel except nickel not iron based), and are known as Incoloy. So don’t know if they would work as both chromium and iron like to be BCC. Probably why stainless steel itself doesn’t work.

            Edit2: fixed link

          • Bob Greenyer

            There is 0.038% CO2 in the air.

          • Ged

            0.038% actually (380 parts per million). 3.8% exceeds OSHA’s maximum safe levels, and 10% is lethal in 30 minutes.

          • Bob Greenyer

            oops – dp issue – serve me right for being up nearly all night all weekend!

          • Ged

            All for a good cause!

          • Bob Greenyer

            for sure. Hey, thanks for the graphs again… helped us to see if we needed to go further with calibration or if a polynomial would fit and allow us to start the run faster.

          • Ged

            I’m glad I could help 🙂

            Some people like to stay up all night playing video games. Other people like to stay up all night trying to controlled light things on (nuclear) fire.

          • Bob Greenyer

            I can honestly say – I really enjoy it.

  • Bob Greenyer

    Here is calibration for current run (Ged already done chart on other experiment thread)

    https://goo.gl/oB6abV

    Here is data from the current run

    https://goo.gl/iyQO6p

  • Bob Greenyer

    Here is calibration for current run (Ged already done chart on other experiment thread)

    https://goo.gl/oB6abV

    Here is data from the current run

    https://goo.gl/iyQO6p

  • Axil Axil

    If this is a replication of the Lugano test, why was there no fuel preprocessing done where the nickel powder comes out covered with lithium and the 5 micron particles were sintered together at high temperatures into at least a few 100 micron particles?

    • Ged

      That may be an important (even critical) surface prep or activation method, but I still wonder what bulk differences Rossi had in his nickel.

      I doubt an e-cigarette wire is pure nickel (correct me if I am wrong), and it seems suggestive of weak activity here, so it may be at least in the right direction.

      • Axil Axil

        The fuel particle after preprocessing had a large amount of carbon on the surface of the 100 micro particle, but the ash had no carbon. Does that not mean that carbon was a consumable element in the fuel?

        Where is the carbon on the surface of the fuel for this test?

        • Bob Greenyer

          This is a Nickel wire in an environment of LiAlH4 – we have been wanting to do this for a while with a Celani wire. After calibration the LiAlH4 is added and the cell shook to allow dispersal of the LiAlH4 (180mg) – there is a lot of free volume in this cell.

  • Axil Axil

    If this is a replication of the Lugano test, why was there no fuel preprocessing done where the nickel powder comes out covered with lithium and the 5 micron particles were sintered together at high temperatures into at least a few 100 micron particles?

    • Ged

      That may be an important (even critical) surface prep or activation method, but I still wonder what bulk differences Rossi had in his nickel.

      I doubt an e-cigarette wire is pure nickel (correct me if I am wrong), and it seems suggestive of weak activity here, so it may be at least in the right direction.

      • Axil Axil

        The fuel particle after preprocessing had a large amount of carbon on the surface of the 100 micro particle, but the ash had no carbon. Does that not mean that carbon was a consumable element in the fuel?

        Where is the carbon on the surface of the fuel for this test?

        • Bob Greenyer

          This is a Nickel wire in an environment of LiAlH4 – we have been wanting to do this for a while with a Celani wire. After calibration the LiAlH4 is added and the cell shook to allow dispersal of the LiAlH4 (180mg) – there is a lot of free volume in this cell.

  • Great Work! I suggest putting R = Power / Current squared Reactor Live Data / Plots…

  • US_Citizen71

    Maybe the experiment should be tried with Inconel wire as the heater/lattice. Many of the trace elements in Inconel match up with elements detected in the Lugano Isotope analysis. It also creates cubic crystals at high temperatures and ‘cracking and microstructural segregation’ as well. – ” Inconel’s high temperature strength is developed by solid solution strengthening or precipitation strengthening, depending on the alloy. In age-hardening or precipitation-strengthening varieties, small amounts of niobium combine with nickel to form the intermetallic compound Ni3Nb or gamma prime (γ’). Gamma prime forms small cubic crystals that inhibit slip and creep effectively at elevated temperatures.[10] The formation of gamma-prime crystals increases over time, especially after three hours of a heat exposure of 850 °C, and continues to grow after 72 hours of exposure.[11]” … “Welding of some Inconel alloys (especially the gamma prime precipitation hardened family, e.g. Waspalloy and X-750) can be difficult due to cracking and microstructural segregation of alloying elements in the heat-affected zone.” https://en.wikipedia.org/wiki/Inconel

    It seems to be affordable enough for use. First listing I found: http://www.sisweb.com/ms/wire-inconel.htm#

    • Ged

      Incoloy is another that could be tried (see below). But don’t know if either are FCC. Inconel has been tried before with limited success I believe; but I don’t remember the circumstances or if we knew then what we know now.

      • US_Citizen71

        Appears to be FCC for both – “A superalloy, or high-performance alloy, is an alloy that exhibits several key characteristics: excellent mechanical strength, resistance to thermal creep deformation, good surface stability and resistance to corrosion or oxidation. The crystal structure is typically face-centered cubic austenitic. Examples of such alloys are Hastelloy, Inconel, Waspaloy, Rene alloys, Haynes alloys, Incoloy, MP98T, TMS alloys, and CMSX single crystal alloys.” – https://en.wikipedia.org/wiki/Superalloy

        • Ged

          Sadly “typically” doesn’t tell us specifically :(. Inconel I would expect is FCC (I’d like to know for sure though, but have yet to find the detail), but Incoloy is alloyed with chromium and iron which both want to be BCC (with carbon dispersed in the interstitium since nickel lattices are so nice and large spaced). Would depend on the relative abundances.

          • US_Citizen71

            If I am interpreting this correctly it appears that the FCC crystal structure is stable for iron inside our our magic temperature zone for the Hot Cat (Lugano Experiment) – “Allotrope of iron – From 912 to 1,394 °C (1,674 to 2,541 °F) alpha iron undergoes a phase transition from body-centred cubic (BCC) to the face-centred cubic (FCC) configuration of gamma iron, also called austenite. This is similarly soft and ductile but can dissolve considerably more carbon (as much as 2.04% by mass at 1,146 °C (2,095 °F)).” – https://en.wikipedia.org/wiki/Austenite

          • Ged

            Yeah, I saw that too, but didn’t make the connection to the “magic temperature zone” of the hot cat. Very good catch, and intriguing.

            Also found that Incoloy has been studied extensively for its use in hot fusion reactorsEdit3: interestingly, incoloy has been studied deeply in regards to its application in hot fusion reactors
            http://fti.neep.wisc.edu/pdf/fdm801.pdf&ved=0CCUQFjAFahUKEwjdo475pL3IAhWBdR4KHUAtDhY&usg=AFQjCNEgcTiVffscBijdjxkieKansbFbAQ&sig2=LlsP_sjTm0LOjxCA4SVN8w

          • US_Citizen71

            Your link is not working.

          • Ged

            Here we go: http://fti.neep.wisc.edu/pdf/fdm801.pdf

            Pasting with a phone from Google is quite a journey sometimes.

          • US_Citizen71

            Yep!!!

  • US_Citizen71

    Maybe the experiment should be tried with Inconel wire as the heater/lattice. Many of the trace elements in Inconel match up with elements detected in the Lugano Isotope analysis. It also creates cubic crystals at high temperatures and ‘cracking and microstructural segregation’ as well. – “Inconel’s high temperature strength is developed by solid solution strengthening or precipitation strengthening, depending on the alloy. In age-hardening or precipitation-strengthening varieties, small amounts of niobium combine with nickel to form the intermetallic compound Ni3Nb or gamma prime (γ’). Gamma prime forms small cubic crystals that inhibit slip and creep effectively at elevated temperatures.[10] The formation of gamma-prime crystals increases over time, especially after three hours of a heat exposure of 850 °C, and continues to grow after 72 hours of exposure.[11]” … “Welding of some Inconel alloys (especially the gamma prime precipitation hardened family, e.g. Waspalloy and X-750) can be difficult due to cracking and microstructural segregation of alloying elements in the heat-affected zone.” https://en.wikipedia.org/wiki/Inconel

    It seems to be affordable enough for use. First listing I found: http://www.sisweb.com/ms/wire-inconel.htm#

    • Ged

      Incoloy is another that could be tried (see below). But don’t know if either are FCC. Inconel has been tried before with limited success I believe; but I don’t remember the circumstances or if we knew then what we know now.

      • US_Citizen71

        Appears to be FCC for both – “A superalloy, or high-performance alloy, is an alloy that exhibits several key characteristics: excellent mechanical strength, resistance to thermal creep deformation, good surface stability and resistance to corrosion or oxidation. The crystal structure is typically face-centered cubic austenitic. Examples of such alloys are Hastelloy, Inconel, Waspaloy, Rene alloys, Haynes alloys, Incoloy, MP98T, TMS alloys, and CMSX single crystal alloys.” – https://en.wikipedia.org/wiki/Superalloy

        • Ged

          Sadly “typically” doesn’t tell us specifically :(. Inconel I would expect is FCC (I’d like to know for sure though, but have yet to find the detail), but Incoloy is alloyed with chromium and iron which both want to be BCC (with carbon dispersed in the interstitium since nickel lattices are so nice and large spaced). Would depend on the relative abundances.

          Edit: ok, found a paper saying Inconel 718 is definitely FCC http://www.emc2012.org.uk//documents/Abstracts/Abstracts/EMC2012_1002.pdf

          Edit2: ok, found a book saying Incoloy 825 is also FCC, so that solves that. http://tubingchina.com/Incoloy-825-Stainless-Steel.htm

          Edit3: more link fixing!

          • US_Citizen71

            If I am interpreting this correctly it appears that the FCC crystal structure is stable for iron inside our our magic temperature zone for the Hot Cat (Lugano Experiment) – “Allotrope of iron – From 912 to 1,394 °C (1,674 to 2,541 °F) alpha iron undergoes a phase transition from body-centred cubic (BCC) to the face-centred cubic (FCC) configuration of gamma iron, also called austenite. This is similarly soft and ductile but can dissolve considerably more carbon (as much as 2.04% by mass at 1,146 °C (2,095 °F)).” – https://en.wikipedia.org/wiki/Austenite

          • Ged

            Yeah, I saw that too, but didn’t make the connection to the “magic temperature zone” of the hot cat. Very good catch, and intriguing.

            Also found that Incoloy has been studied extensively for its use in hot fusion reactors

            http://fti.neep.wisc.edu/pdf/fdm801.pdf&ved=0CCUQFjAFahUKEwjdo475pL3IAhWBdR4KHUAtDhY&usg=AFQjCNEgcTiVffscBijdjxkieKansbFbAQ&sig2=LlsP_sjTm0LOjxCA4SVN8w

          • US_Citizen71

            Your link is not working.

          • Ged

            Here we go: http://fti.neep.wisc.edu/pdf/fdm801.pdf

            Pasting with a phone from Google is quite a journey sometimes.

          • US_Citizen71

            Yep!!!

  • Sanjeev

    So is the EM generator thing on at this time ?
    Is that a factor for COP of 1.1?

    • Bob Greenyer

      We did it at the start

  • Sanjeev

    So is the EM generator thing on at this time ?
    Is that a factor for COP of 1.1?

    • Bob Greenyer

      We did it at the start

      • Roger Barker

        When will MFMP be running another experiment?

        • Bob Greenyer

          Well, I was working with me356 all over the weekend.

          Mathieu is waiting for Celani wires and a pump for the MFC

          Alan is preparing his next experiment

          Bob Higgins is preparing his.

          • Roger Barker

            Ok thanks.

  • Roger Barker

    Have I been banned? Please don’t ban me! 🙁

  • Bob Greenyer

    Well, I was working with me356 all over the weekend.

    Mathieu is waiting for Celani wires and a pump for the MFC

    Alan is preparing his next experiment

    Bob is preparing his.

  • Ged

    Here’s the charts now that I’m able:

    http://s28.postimg.org/at2y0h71p/151012_mashup_Run1.png

    http://s29.postimg.org/bqo0u4vc7/151012_mashup_Run1_Cper_W.png

    The effect is definitely getting larger relative to calibration as the temperature gets higher.

    • nietsnie

      If I’m understanding the identity of ‘power’, at above 200 degrees it is taking considerably more electricity to produce the same temperature output as the calibration run. That is not hoped for – but it is interesting. Most especially the spike in output every 50 degrees or so. In as much as the calibration run had those (to a lesser extent…) too it can’t be due to something fuel related. What’s doing that? I don’t think I’ve seen anything like that in previous experiment runs.

      • Ged

        Other way around, actually. Take another look: the run produces more heat per watt.

        The spikes are steady points where the temperature and power were held. The looping inbetweens are points where the set point has been changed and the power is ramping up. Due to thermal mass, the device lags behind (due to the speed of the ramp) causing the “bowl” shape between the spikes. Only the spikes matter, or areas where the trace is slow enough to be steady state (like around 150-200 C). One can identify when its steady state because the line gets noisy as it bounces around the setpoint over a long enough time to get averaged, while fast motions are smooth.

        • nietsnie

          Thanks for setting me straight, Ged. And sorry for not responding for 11 hours – I didn’t check back immediately and then I went to bed.

          I haven’t looked at the raw data from either run. I might have it wrong, but, as I have imagined it, the difference between the two runs is just the addition of a very small amount of LiAlH4 that is just sort of poured into the tube and then shaken (not stirred…). Do you conclude that the relatively exaggerated troughs in the live run represent the additional energy needed to get the mass of the LiAlH4 up to temp? It’s hard to gauge exactly how much additional power is represented by the area in the troughs, but it looks like a lot – at some temperatures as much as 6 or 8 watts. How much LiAlH4 mass was in there? Is that reasonable?

          • Ged

            No worries! And no, you are completely correct about the difference between the two runs physically.

            Well, two things.

            1) The LiAlH4 is an insignificant mass compared to the reactor body. Really, the troughs represent transition points as the reactor jumps to the target temperature. During the calibration, there were smaller, longer steps, while during the run they rapidly brought it up to the temperature range of interest (400 C), so that caused a lot of “power moving faster than temperature so it looks like less C per W” offsetting. Once the temperature catches up, you get the “spike” where the majority of the data points are sitting (steady state hold temperature).

            2) There are a sparse few data points in the “bowl” areas, so it’s also partly from the way graphing in a scatter plot works. The majority of all the data, and thus energy, sits in the spikes, so they carry almost all the weight. The “bowly swoops” are ignorable, and can’t really be used for calculation, only steady states are reliable.

            So, in conclusion, the graph swoops actually aren’t “more energy” being needed, it’s just a faster power rate of change. The thermal lag is dependent on mass and isn’t going to change much between runs, but the speed of change is completely controllable. Meanwhile, that rapid change means there’s few data points in that area relative to the high amount of data points in the spikes. Oh the joys of graphing.

            Great questions.

          • nietsnie

            Ahhh… there *was* an additional variable besides the mass of the LiAlH4: the time between temp increase steps. That makes sense then. Thanks for your great explanation.

          • Ged

            Thinking on it more, it would have been a whole lot easier to explain if I had simply said “when the set point is changed, the PID shoots the power way up to fast drive temperature to that point, but then drops the power to maintain that point once reached”. It’s like flooring the accelerator to get to 60 mph, but then easing way off to cruise.

            All the previous tests we’ve seen used a slow ramp system where the PID or human controller only changed the power in small, exact steps to meet target temp, and wouldn’t make the next step till temperature reached the point, avoiding power overshoots of such magnitude (like applying the perfect amount of acceleration to coast into your cruise speed). This “ferrari” syndrome makes for some cool but unintuitive graphing. Sanjeev made easier to read graphs above using just the steady states and not everything in between like moi.

          • nietsnie

            But, had the changes in acceleration happened identically between the active and control runs – we would expect for there not to have been any significant difference between the results in the trough areas – or, anyway, we would not expect to see more power apparently used to ramp up during the active run. That additional variable is what caused them not to match.

          • Ged

            Precisely. The calibration was slower, ’cause they were doing a calibration sweep. For the run, they knew what temperature they wanted, and ran right for it.

    • Obvious

      These are totally cool-looking plots.
      Is there any indication of a resistance change from calibration?
      Just wondering about how the excess heat is manifesting.
      Improvement in conductivity due to an H-enriched atmosphere seems hard to believe, considering how little is likely to be in there.
      Ni-Li-Al alloying or coatings might show up as a change in resistance, although the low temperature of the start of the effect tends to undermine that idea.
      This is definitely something to be worked on in more detail and slight modifications to tease out the dynamics a bit more.

      • Ged

        Ask, and you shall receive!

        http://s23.postimg.org/cixe3ytkr/151012_mashup_Run1_Resistance.png

        Calculated via the R = P/I^2 relationship, as suggested by another commentor.

        Resistance is unchanged per temperature between the calibration and run. You can really see the set point, power stepping behavior in this.

        • Obvious

          I thought we started with 4 ohms.

  • Ged

    Here’s the charts now that I’m able:

    http://s28.postimg.org/at2y0h71p/151012_mashup_Run1.png

    http://s29.postimg.org/bqo0u4vc7/151012_mashup_Run1_Cper_W.png

    The effect is definitely getting larger relative to calibration as the temperature gets higher.

    • nietsnie

      If I’m understanding the identity of ‘power’, at above 200 degrees it is taking considerably more electricity to produce the same temperature output as the calibration run. That is not hoped for – but it is interesting. Most especially the spike in output every 50 degrees or so. In as much as the calibration run had those (to a lesser extent…) too it can’t be due to something fuel related. What’s doing that? I don’t think I’ve seen anything like that in previous experiment runs.

      • Ged

        Other way around, actually. Take another look: the run produces more heat per watt.

        The spikes are steady points where the temperature and power were held. The looping inbetweens are points where the set point has been changed and the power is ramping up. Due to thermal mass, the device lags behind (due to the speed of the ramp) causing the “bowl” shape between the spikes. Only the spikes matter, or areas where the trace is slow enough to be steady state (like around 150-200 C). One can identify when it’s steady state because the line gets noisy as it bounces around the setpoint over a long enough time to get averaged, while fast motions are smooth.

        Edit: I recommend taking a look at the raw data side by side if you want to get an idea of what I mean. A smaller averaging window than the 100 point average I used would probably also reduce the “bowl” effect.

        Edit2: Yeah, it’s just thermal lag. Changing to a 20 point moving average did nothing to the “bowls”. Just too fast a change between steady states to look good on a graph.

        • nietsnie

          Thanks for setting me straight, Ged. And sorry for not responding for 11 hours – I didn’t check back immediately and then I went to bed.

          I haven’t looked at the raw data from either run. I might have it wrong, but, as I have imagined it, the difference between the two runs is just the addition of a very small amount of LiAlH4 that is just sort of poured into the tube and then shaken (not stirred…). Do you conclude that the relatively exaggerated troughs in the live run represent the additional energy needed to get the mass of the LiAlH4 up to temp? It’s hard to gauge exactly how much additional power is represented by the area in the troughs, but it looks like a lot – at some temperatures as much as 6 or 8 watts. How much LiAlH4 mass was in there? Is that reasonable?

          • Ged

            No worries! And no, you are completely correct about the difference between the two runs physically.

            Well, two things.

            1) The LiAlH4 is an insignificant mass compared to the reactor body. Really, the troughs represent transition points as the reactor jumps to the target temperature. During the calibration, there were smaller, longer steps, while during the run they rapidly brought it up to the temperature range of interest (400 C), so that caused a lot of “power moving faster than temperature so it looks like less C per W” offsetting. Once the temperature catches up, you get the “spike” where the majority of the data points are sitting (steady state hold temperature).

            2) There are a sparse few data points in the “bowl” areas, so it’s also partly from the way graphing in a scatter plot works. The majority of all the data, and thus energy, sits in the spikes, so they carry almost all the weight. The “bowly swoops” are ignorable, and can’t really be used for calculation, only steady states are reliable.

            So, in conclusion, the graph swoops actually aren’t “more energy” being needed, it’s just a faster power rate of change. The thermal lag is dependent on mass and isn’t going to change much between runs, but the speed of change is completely controllable. Meanwhile, that rapid change means there’s few data points in that area relative to the high amount of data points in the spikes. Oh the joys of graphing.

            Great questions.

          • nietsnie

            Ahhh… there *was* an additional variable besides the mass of the LiAlH4: the time between temp increase steps. That makes sense then. Thanks for your great explanation.

          • Ged

            Thinking on it more, it would have been a whole lot easier to explain if I had simply said “when the set point is changed, the PID shoots the power way up to fast drive temperature to that point, but then drops the power to maintain that point once reached”. It’s like flooring the accelerator to get to 60 mph, but then easing way off to cruise.

            All the previous tests we’ve seen used a slow ramp system where the PID or human controller only changed the power in small, exact steps to meet target temp, and wouldn’t make the next step till temperature reached the point, avoiding power overshoots of such magnitude (like applying the perfect amount of acceleration to coast into your cruise speed). This “ferrari” syndrome makes for some cool but unintuitive graphing. Sanjeev made easier to read graphs above using just the steady states and not everything in between like moi.

          • nietsnie

            But, had the changes in acceleration happened identically between the active and control runs – we would expect for there not to have been any significant difference between the results in the trough areas – or, anyway, we would not expect to see more power apparently used to ramp up during the active run. That additional variable is what caused them not to match.

          • Ged

            Precisely. The calibration was slower, ’cause they were doing a calibration sweep. For the run, they knew what temperature they wanted, and ran right for it.

    • Obvious

      These are totally cool-looking plots.
      Is there any indication of a resistance change from calibration?
      Just wondering about how the excess heat is manifesting.
      Improvement in conductivity due to an H-enriched atmosphere seems hard to believe, considering how little is likely to be in there.
      Ni-Li-Al alloying or coatings might show up as a change in resistance, although the low temperature of the start of the effect tends to undermine that idea.
      This is definitely something to be worked on in more detail and slight modifications to tease out the dynamics a bit more.

      • Ged

        Ask, and you shall receive!

        http://s23.postimg.org/cixe3ytkr/151012_mashup_Run1_Resistance.png

        Calculated via the R = P/I^2 relationship, as suggested by another commentor.

        Resistance is unchanged per temperature between the calibration and run. You can really see the between set point, power stepping, temperature lag behavior in this.

        • Obvious

          I thought we started with 4 ohms.

  • Ged

    It seems like the average power is ever so slowly getting lower while temp holds. Emphasis on slowly.

    • Bob Greenyer

      That, and the pressure has risen from 3/4 bar to 1 bar over atm.

      • Sanjeev

        I suggest keeping the power steady and see if the temperature increases. A cross check.

        • Bob Greenyer

          difficult with the control set-up and the phase angle triggered triac. the power data comes from the PCE-830. Would need regulated DC for this purpose to be very precise.

        • Axil Axil

          Power needs to vary in a cycle.

    • Sanjeev

      Can be due to ambient temperature changes.

  • Ged

    It seems like the average power is ever so slowly getting lower while temp holds. Emphasis on slowly.

    Edit: Based on the highs and lows and where it seems to sit more of the time, seems about 1-3 W lower than previously. So about 10-11% now. The period on my graph is officially 8.5% lower power than calibration.

    • Bob Greenyer

      That, and the pressure has risen from 3/4 bar to 1 bar over atm.

      • Sanjeev

        I suggest keeping the power steady and see if the temperature increases. A cross check.

        • Bob Greenyer

          difficult with the control set-up and the phase angle triggered triac. the power data comes from the PCE-830. Would need regulated DC for this purpose to be very precise.

        • Axil Axil

          Power needs to vary in a cycle.

    • Sanjeev

      Can be due to ambient temperature changes.

  • Sanjeev

    This is a “simplified” plot. I took the steady temperature values at which the reactor was held (setpoints) and computed the average powers for calib and fueled runs.
    Plot attached.

    • Bob Greenyer

      Thanks Sanjeev!

      Well, there are two interesting points here

      1. the endothermic breakdown of LiAlH4 below 200ºC
      2. the appearance of apparent heat (based on calibration and ‘active’ run whole reactor hot spot from PI160) onset from 250ºC, as per IH patent application.

      Could you do a polynomial fit / extension of the calibration so that me356 can take it to higher power with something to compare too (input power to output temp)

      • Sanjeev

        Calibration extrapolated.

        • Bob Greenyer

          yes thanks, is there supposed to be an image attached?

          • Sanjeev

            Yes its attached. You may need to refresh the page. Its an Disqus issue.

          • Bob Greenyer

            Could you swap the axis, more intuitive to read that way

          • Sanjeev

            Sure. Its attached.
            With swapped axes, the poly fit is not so good at higher temperatures, so a linear fit was also added. It will be very approx.

          • Bob Greenyer

            try double entering the first point.

          • Sanjeev

            Sorry Bob, I was gone for a while. It didn’t work, but the log fits better. Here is the plot again with log fit. Probably Ged or Ecco can do a better job.
            Anyway I think a falling power will be a more reliable sign of excess than a computed calibration.

    • Ged

      Beautiful, Sanjeev, thanks for that!

    • Obvious

      This apparent smoothness does not seem to imply the onset of some sort of unexplained nuclear phenomena. The design of the device has, however, eliminated many potential sources of problems like binding coils, etc.
      Has the emissivity changed? Does Li and H vapour improve heat conduction? Significantly with this “fuel” load? What does 10% more and less fuel do? What does lesser Ni content in the wire do? Is there a Dirac sea loophole being exploited by certain atomic geometries? Do new electrical paths change the resistivity characteristics to that of a more efficient IR emitter (Nernst lamps, etc.)?
      This is a great device for the home experimenters to work with. Thanks me356.

      • Ged

        If I remember my stats right, the best way to analyze the “smoothness” or if there are “mode” changes in the graph is to do break point analysis. I might be able to do something like that using dT/dP ratios, or d(dT/dP) more specifically. Our brains naturally will smooth out a trace and miss slow behavior changes, so it’s hard to eyeball this sort of stuff.

        So much to glean from this, so much potential experimentation; all of those ideas you wrote would be great to explore.

        • Obvious

          Perhaps it’s just a result of my preconceptions, but I tend to think that an atomic effect, with generally thousands to millions of times the energy compared to chemical effects, should show up with a sharp spike in output. Maybe it doesn’t work that way, or the effect is being activated too ineffectively to cause a sudden spike in output. That would certainly make it easier to control. For the purposes of experimentation, a slow, gradual increase in output is definitely a bonus.

          • Ged

            I have an important question for you. If resistance does fall, how will that effect the heat production? You know a lot more about this, especially first hand, than me.

            If power goes down, because resistance goes down but amps hold, would heat also go down, since it is power that translates into heat (and back)? Or is it trickier than that? I tend to think power is power is power (since power is energy rate, and energy is heat), but I need to know for sure in this sort of setting.

          • Obvious

            If the amps hold steady, and the resistance drops, then power will go down in a “steady state” system (no adjustments to the other parameters). The altered wire or composite conductor simply becomes a better conductor than the original wire, which means less “loss”, which is heat. Heat is Current(squared) x Resistance for a partial conductor.
            P = (I^2)R

            In order to hold the amps steady, in case of improved conduction, the voltage would need to be adjusted lower to compensate, or the amps must increase. Duty cycle, PWM, and triac (etc.) changes can accomplish limiting of V and I, separately and/or simultaneously, depending on design parameters.

            In the beginning case, let P = 100W, I = 10 A, R=1 ohm, and V = 10A:

            Using P=I^2R, if R is reduced to 1/2, the power drops in half, (in the overly simple version of things).

            —P = (I^2)(R)(0.5)
            –2P = (2)100(1)
            —P = 50 W

            To accomplish the same power level drop as dropping the resistance in half requires a 30% drop in current with the original resistance.
            (I then = sqrt(50) = 7.07 A)

            ——P = (I^2)R ; I= sqrt(P/R)
            ———————I = (sqrt(50)[A])/(1[ohm])
            ———————I = 7.07 A

            Voltage will have to be dropped in order to reduce the current.
            (V =IR, and R=1), and P=V^2/R

            P = IV ; V = P/I
            V = P/I
            V = 50/7.07
            V = 7.07

            So if current is held constant, resistance halved, and V reduced appropriately to allow (control of) constant current:

            P = IV
            P= 10[A](7.07[V]) = 70.7 W

            or roughly a 30% drop in power (heat) with constant current, and 1/2 resistance.

            I hope I haven’t messed that up too badly.

            The squares, square roots and complimentary V/I changes make it a bit more complex than a straight function of resistance and power, when resistance drops, especially in a current and voltage controlled/limited system.

            In the case where voltage is held constant, current would double if resistance were cut in half: (If current is not additionally limited by some other means).

            —P = I^2(R)
            —P = I^2(0.5R)

            –2P = 2I^2(1)
            –2P = 20^2(1)
            –2P = 400(1)
            —P = 200 W
            or
            P = IV
            P = 20(10)
            P = 200 W

            Which means double the power with constant volts, one half resistance, and current allowed to flow unimpeded (except for the resistor described).

          • Ged

            Thank you greatly, and for such a detailed example. It is how I was figuring it, which is really good to know.

  • Sanjeev

    This is a “simplified” plot. I took the steady temperature values at which the reactor was held (setpoints) and computed the average powers for calib and fueled runs.
    Plot attached.

    • Bob Greenyer

      Thanks Sanjeev!

      Well, there are two interesting points here

      1. the endothermic breakdown of LiAlH4 below 200ºC
      2. the appearance of apparent heat (based on calibration and ‘active’ run whole reactor hot spot from PI160) onset from 250ºC, as per IH patent application.

      Could you do a polynomial fit / extension of the calibration so that me356 can take it to higher power with something to compare too (input power to output temp)

      • Sanjeev

        Calibration extrapolated.

        • Bob Greenyer

          That is great thanks

          • Sanjeev

            Yes its attached. You may need to refresh the page. Its an Disqus issue.

          • Bob Greenyer

            Could you swap the axis, more intuitive to read that way

          • Sanjeev

            Sure. Its attached.
            With swapped axes, the poly fit is not so good at higher temperatures, so a linear fit was also added. It will be very approx.

          • Bob Greenyer

            try double entering the first point.

          • Sanjeev

            Sorry Bob, I was gone for a while. It didn’t work, but the log fits better. Here is the plot again with log fit. There is a big difference compared to the poly fit.
            Probably Ged or Ecco can do a better job.
            Anyway I think a falling power will be a more reliable sign of excess than a comparison with computed calibration.

    • Ged

      Beautiful, Sanjeev, thanks for that!

    • Obvious

      This apparent smoothness does not seem to imply the onset of some sort of unexplained nuclear phenomena. The design of the device has, however, eliminated many potential sources of problems like binding coils, etc.
      Has the emissivity changed? Does Li and H vapour improve heat conduction? Significantly with this “fuel” load? What does 10% more and less fuel do? What does lesser Ni content in the wire do? Is there a Dirac sea loophole being exploited by certain atomic geometries? Do new electrical paths change the resistivity characteristics to that of a more efficient IR emitter (Nernst lamps, etc.)?

      This is a great device for the home experimenters to work with. Thanks me356.

      • Ged

        If I remember my stats right, the best way to analyze the “smoothness” or if there are “mode” changes in the graph is to do break point analysis. I might be able to do something like that using dT/dP ratios, or d(dT/dP) more specifically. Our brains naturally will smooth out a trace and miss slow behavior changes, so it’s hard to eyeball this sort of stuff.

        So much to glean from this, so much potential experimentation; all of those ideas you wrote would be great to explore.

        Edit: I already tried to do a Kolmogorov–Smirnov test yesterday to compare the development of the two traces statistically in time (to see if they belong to the same “behavior”), but my laptop is just not able to handle so much data crunching with that test so far. Hopefully I can get it working as it’s a good statistical measure to tell if there is indeed a significant difference in the behavior of the Run trace versus Calibration over time.

        • Obvious

          Perhaps it’s just a result of my preconceptions, but I tend to think that an atomic effect, with generally thousands to millions of times the energy compared to chemical effects, should show up with a sharp spike in output. Maybe it doesn’t work that way, or the effect is being activated too ineffectively to cause a sudden spike in output. That would certainly make it easier to control. For the purposes of experimentation, a slow, gradual increase in output is definitely a bonus.

          • Ged

            I have an important question for you. If resistance does fall, how will that effect the heat production? You know a lot more about this, especially first hand, than me.

            If power goes down, because resistance goes down but amps hold, would heat also go down, since it is power that translates into heat (and back)? Or is it trickier than that? I tend to think power is power is power (since power is energy rate, and energy is heat), but I need to know for sure in this sort of setting.

          • Obvious

            If the amps hold steady, and the resistance drops, then power will go down in a “steady state” system (no adjustments to the other parameters). The altered wire or composite conductor simply becomes a better conductor than the original wire, which means less “loss”, which is heat. Heat is Current(squared) x Resistance for a partial conductor.
            P = (I^2)R

            In order to hold the amps steady, in case of improved conduction, the voltage would need to be adjusted lower to compensate, or the amps must increase. Duty cycle, PWM, and triac (etc.) changes can accomplish limiting of V and I, separately and/or simultaneously, depending on design parameters.

            In the beginning case, let P = 100W, I = 10 A, R=1 ohm, and V = 10A:

            Using P=I^2R, if R is reduced to 1/2, the power drops in half, (in the overly simple version of things).

            —P = (I^2)(R)(0.5)
            –2P = (2)100(1)
            —P = 50 W

            To accomplish the same power level drop as dropping the resistance in half requires a 30% drop in current with the original resistance.
            (I then = sqrt(50) = 7.07 A)

            ——P = (I^2)R ; I= sqrt(P/R)
            ———————I = (sqrt(50)[A])/(1[ohm])
            ———————I = 7.07 A

            Voltage will have to be dropped in order to reduce the current.
            (V =IR, and R=1), and P=V^2/R

            P = IV ; V = P/I
            V = P/I
            V = 50/7.07
            V = 7.07

            So if current is held constant, resistance halved, and V reduced appropriately to allow (control of) constant current:

            P = IV
            P= 10[A](7.07[V]) = 70.7 W

            or roughly a 30% drop in power (heat) with constant current, and 1/2 resistance.

            I hope I haven’t messed that up too badly.

            The squares, square roots and complimentary V/I changes make it a bit more complex than a straight function of resistance and power, when resistance drops, especially in a current and voltage controlled/limited system.

            In the case where voltage is held constant, current would double if resistance were cut in half: (If current is not additionally limited by some other means).

            —P = I^2(R)
            —P = I^2(0.5R)

            –2P = 2I^2(1)
            –2P = 20^2(1)
            –2P = 400(1)
            —P = 200 W
            or
            P = IV
            P = 20(10)
            P = 200 W

            Which means double the power with constant volts, one half resistance, and current allowed to flow unimpeded (except for the resistor described).

          • Ged

            Thank you greatly, and for such a detailed example. It is how I was figuring it, which is really good to know.

  • Sanjeev

    Another plot for excess and COP. Very tiny effect, but encouraging.

    • Ged

      Also a pretty graph! Nevermind my earlier post, I misread something, hah.

      • Bob Greenyer

        The excess in W is LHS 6W and the COP is RHS which is just under 1.1, as I read it.

        • Ged

          I misread excess as a percentage, rather than watts. Whoops!

      • Bob Matulis

        How is it possible to have a COP less than 1? (Conservation of energy)

        • Bob Greenyer

          As determined against the calorimeter, input power to output temp

          How might this seem apparent? – for a start, at the lower temperatures below 200ºC, there is the endothermic breakdown of LiAlH4.

  • Sanjeev

    Another plot for excess and COP. Very tiny effect, but encouraging.

    • Ged

      Also a pretty graph! Nevermind my earlier post, I misread something, hah.

      • Bob Greenyer

        The excess in W is LHS 6W and the COP is RHS which is just under 1.1, as I read it.

        • Ged

          I misread excess as a percentage, rather than watts. Whoops!

      • Bob Matulis

        How is it possible to have a COP less than 1? (Conservation of energy)

        • Bob Greenyer

          As determined against the calorimeter, input power to output temp

          How might this seem apparent? – for a start, at the lower temperatures below 200ºC, there is the endothermic breakdown of LiAlH4.

  • Kristian

    Interesting! Where (how) was the temperature measured? (I’m sure that information is given somewhere, just hoping someone can give me a quick re-cap.)

    • Bob Greenyer

      Via Optris PI160 correctly calibrated with correct choice of emissivity, verified against K-Type thermocouple and Voltcraft IR2200.

      At the moment, both the calibration and live run data are taken from the hottest spot. After the experiment, the recording of the Optris data can be analysed further – with an average area temp in both cases for more robust results.

  • Bob Greenyer

    Via Optris PI160 correctly calibrated with correct choice of emissivity, verified against K-Type thermocouple and Voltcraft IR2200.

    At the moment, both the calibration and live run data are taken from the hottest spot. After the experiment, the recording of the Optris data can be analysed further – with an average area temp in both cases for more robust results.

  • Bob Greenyer

    Making of and EMI generator. This was used at the beginning of this run and at least once during it for a short period of time.

    https://youtu.be/gFHYx7qEofo

    • Ged

      Maybe it should be re-applied now that it’s at holding temp? Or intermittently?

  • Bob Greenyer

    Making of and EMI generator. This was used at the beginning of this run and at least once during it for a short period of time.

    https://youtu.be/gFHYx7qEofo

    • Ged

      Maybe it should be re-applied now that it’s at holding temp? Or intermittently?

  • Scott Jantz

    Nice experiment what is the schedule for temperatures above 400C?

  • Ged

    Power in still seems to be going down very, very slowly. The 80 W line has disappeared now in the autoscaled graph as the trace has dropped far enough away from it. Current seems unchanged though.

  • nietsnie

    Anyone know where COP is now?

    • Ged

      Not that much different. Probably around 1.11-1.12. Not all the meaningful, but encouraging.

      • nietsnie

        Still within margin of error then? How big is the margin of error?

        • Ged

          Honestly? I have no idea. We’d need a number of calibration runs to constrain the error bounds. As it is, I’d only start to feel comfortable with something above 20% (COP >1.2).

          • nietsnie

            Well – not too far then. And we’re still only at 400 C. Very promising.

  • nietsnie

    Anyone know where COP is now?

    • Ged

      Not that much different. Probably around 1.11-1.12. Not all that meaningful, but encouraging.

      • nietsnie

        Still within margin of error then? How big is the margin of error?

        • Ged

          Honestly? I have no idea. We’d need a number of calibration runs to constrain the error bounds. As it is, I’d only start to feel comfortable with something above 20% (COP >1.2).

          • nietsnie

            Well – not too far then. And we’re still only at 400 C. Very promising.

  • Winebuff67

    How long will this run go if we don’t have a problem?

    • Bob Greenyer

      A long time.

      Maybe after a few more power steps, me356 will open reactor and put more LiAlH4 in… then maybe later, some LiAlH4 and Ni.

      • Sanjeev

        I see its gone to 450C now.

        • Ged

          I’ll try to use your extrapolated equation to do a comparison at this new temp once I get home. My best guess at the C/W right now is 5.8, versus the 6.2 at 400 C, but we know it should shrink some as we push temperatures higher due to the whole P ~~ T^4 business.

          Note that 5.8 is roughly, if a little higher than, where the calibration’s C/W was when at 400 C–a good sign our diverging trend is both continuing and growing.

          • Sanjeev

            What I see is that the COP has gone down for step of 450C.
            That extrapolation is no good. Excel offers some “canned” curve fitting. We also do not know the zero crossing point. So may be the COP is increasing but we do not know for sure.
            It should be possible to do a calibration after the fueled run. But I suggest better to put this tube into a calorimeter asap.

          • Ged

            The C/W transform charts are more linear once above 200 C, or at least more regular, and should be easier to curve fit to. I’ll play with it and see if it can make sense. There should be some T^4 equation fit that’ll do; maybe even straight up S-B.

            Definitely second a calorimeter. Extrapolation can only give us a rough estimate.

          • Mats002

            The upside of C/W (T/P) measure us that the heat loss to the environment is the same in all runs, with or without fuel. The down side is that C is measured at one spot and definitely not at all spots at the same time.

            Using calorimetry and T/P compared to a reference run should be quite solid.

  • Winebuff67

    How long will this run go if we don’t have a problem?

    • Bob Greenyer

      A long time.

      Maybe after a few more power steps, me356 will open reactor and put more LiAlH4 in… then maybe later, some LiAlH4 and Ni.

      • Sanjeev

        I see its gone to 450C now.

        • Ged

          I’ll try to use your extrapolated equation to do a comparison at this new temp once I get home. My best guess at the C/W right now is 5.8, versus the 6.2 at 400 C, but we know it should shrink some as we push temperatures higher due to the whole P ~~ T^4 business.

          Note that 5.8 is roughly, if a little higher than, where the calibration’s C/W was when at 400 C–a good sign our diverging trend is both continuing and growing.

          • Kristian

            My understanding is that the temperature measurements so far refer to exterior measurements. If so, any idea of what the interior temperature would be?

          • Ged

            They are indeed external optical IR measurements. We’d need an internal thermal couple to know the ratio of outer to inner temps. Given it’s a pretty thin device, and all heat is coming from the center, I would imagine external and internal are pretty close.

          • Sanjeev

            What I see is that the COP has gone down for step of 450C.
            That extrapolation is no good. Excel offers some “canned” curve fitting. We also do not know the zero crossing point. So may be the COP is increasing but we do not know for sure.
            It should be possible to do a calibration after the fueled run. But I suggest better to put this tube into a calorimeter asap.

          • anonymous

            Estimated COP for 450C is between 1.15 – 1.25.

          • Ged

            The C/W transform charts are more linear once above 200 C, or at least more regular, and should be easier to curve fit to. I’ll play with it and see if it can make sense. There should be some T^4 equation fit that’ll do; maybe even straight up S-B.

            Definitely second a calorimeter. Extrapolation can only give us a rough estimate.

            Edit: C/W certainly seems the most reliable analysis method overall. For instance, since the C/W of the run at 450 C is about the same as the calibration at 400 C, that would imply a COP of 1.13, so only a little higher than before, but trending up at the same rate as we’ve already seen while holding at 400 C.

          • Mats002

            The upside of C/W (T/P) measure us that the heat loss to the environment is the same in all runs, with or without fuel. The down side is that C is measured at one spot and definitely not at all spots at the same time.

            Using calorimetry and T/P compared to a reference run should be quite solid.

  • Bernie Koppenhofer

    From Rossi, “harmonic co-operation between Licensee and Licensor”. Another clue for replicators?

  • Ged

    They are indeed external optical IR measurements. We’d need an internal thermal couple to know the ratio of outer to inner temps. Given it’s a pretty thin device, and all heat is coming from the center, I would imagine external and internal are pretty close.

  • Ged

    Ok, semi-official stats update (in the absence of raw data). The current set point seems to be averaging around 76 W, for a C/W of 5.92. The C/W of the calibration at 400 W was 5.67, meaning the current run is now above 1.13 COP. We can estimate this exactly by doing a trendline on the C/W (T/P) of the calibration averaged at the steady states of 200, 250, 300, 350, and 400 C.

    The relationship is incredibly linear over this temperature range, with an R^2 of 0.9997, for the equation y = -0.0108x + 10.012 (note that this is only valid for this temperature range), where x is the temperature. If we extrapolate using this equation, the expected C/W of the calibration at 450 C would be 5.15, which is reasonable. This gives us a COP of ~1.15.

    Getting closer to more confident levels for sure.

    Edit: Re-did the extrapolation with a polynomial, second order fit with the equation y = -2E-06*x^2 – 0.0097*x + 9.8488 and R^2 = 0.9998. This extrapolation yields 5.08 for the expected calibration temp per power relationship, which would put the COP up to 1.17. While a polynomial is going to be more accurate over the long run, due to the more conservative nature of the linear fit, I will stick with it for the future for extrapolation to prevent overestimation. This will fail at much higher C such as 600+, where we know for certain the relationship becomes non-linear.

    • nietsnie

      Thanks for all your good work, Ged.

  • Bob Greenyer
    • Ged

      Oh wow, congrats to Alan! Can’t wait to hear how it goes.

    • Stephen

      Wow Great News. Both Sveinn and Alan presenting should really be something.

  • Bob Greenyer
    • Ged

      Oh wow, congrats to Alan! Can’t wait to hear how it goes.

    • Stephen

      Wow Great News. Both Sveinn and Alan presenting should really be something.

  • Bob Greenyer

    Alan Goldwater has worked up a best guesstimate of an extended calibration curve.

    • Ged

      Hmm. That curve would put the calibration at 100 W at 450 C, for a COP of 1.26, due to the strong deflection of the trace.

    • Ged

      Using the known power and temps from the calibration, I could calculate the area under view by the Optris for the Stefan Boltzmann equation (~0.0486 m^2, using emissivity 1, or 0.0607 m^2 with emissivity 0.8, but that makes the estimated power much higher and far less conservative later, so going with the first measure). With that, it’s possible to calculate the power the calibration should be using at 450 C based on the S-B law. With an emissivity of 1, that would give 113 W; however the reference emissivity I’ve found for alumina is 0.8 (and then there’s this on the emissivity of alumina at 450 C, coincidentally enough http://www.shimadzu.com/an/ftir/support/tips/letter13/emissivity.html ), which makes the power 90.4 W (4.98 C/W, COP 1.197).

      Using my linear equation from the trendline, the estimated calibration wattage for 450 C is 88.6 W. So, it seems my trendline gives the lowest power estimate, and thus lowest COP estimate for the run for this temp point.

      So, the run is looking promising at the moment, at least.

  • Bob Greenyer

    Alan Goldwater has worked up a best guesstimate of an extended calibration curve.

    • Ged

      Hmm. That curve would put the calibration at 100 W at 450 C, for a COP of 1.26, due to the strong deflection of the trace.

    • Ged

      Using the known power and temps from the calibration, I could calculate the area under view by the Optris for the Stefan Boltzmann equation (~0.0486 m^2, using emissivity 1, or 0.0607 m^2 with emissivity 0.8, but that makes the estimated power much higher and far less conservative later, so going with the first measure). With that, it’s possible to calculate the power the calibration should be using at 450 C based on the S-B law. With an emissivity of 1, that would give 113 W; however the reference emissivity I’ve found for alumina is 0.8 (and then there’s this on the emissivity of alumina at 450 C, coincidentally enough http://www.shimadzu.com/an/ftir/support/tips/letter13/emissivity.html ), which makes the power 90.4 W (4.98 C/W, COP 1.197 for the run). Assuming I’m applying the S-B law correctly.

      Using my linear equation from the trendline, the estimated calibration wattage for 450 C is 87.3 W. So, it seems my trendline gives the lowest power estimate, and thus lowest COP estimate for the run for this temp point.

      So, the run is looking promising at the moment, at least.

  • Ged

    Looks like this device is shaping up to be the best candidate yet to move on to calorimetry, as Sanjeev wisely suggest. There’s too many unknowns with thermometry, and potential other sources of cumulative error; but the signal is very promising at this point, as far as can be determined. We’ll just have to see how it continues to develop.

    • Mats002

      The signal seams to be the mouse. The Cat is not awake yet, something more is needed.

      Me356 have ideas what to do as I understand this thread: http://www.lenr-forum.com/forum/index.php/Thread/2153-Rossi’s-Cat-and-Mouse-reactor-technology/?postID=8463#post8463

      • Ged

        Looks like others are estimating 1.22 and 1.27 COPs (last one based on a smarter use of the S-B law than I did). 1.2+ is definitely a point where I’d feel comfortable saying we’re starting to see something. Now there needs to be a very careful analysis to exclude all errors. But, from what I have read on those forum you linked, with Me356’s replies to other speculations, the setup is very sound and a lot of possible error is already excluded.

        Mouse indeed! But getting louder.

        Definitely need to move this design to calorimetry for precise measurements.

  • Stephen

    It seems to me that the pressure is very slightly higher this morning than last night perhaps by (5%) or 0.1 bar or so. I guess its is a good indicator of the variation in the amount of free Hydrogen Gas, I’m quite curious how it evolves with the Ideal gas law since i would presume the free volume is pretty much constant, and there are long periods of constant T. Or would other gasses still be present, outgassing and/or reacting?

    Is there any pressure and temperature data over the current run? Perhaps at 1 minute or so intervals? Or will this become available after the end of the test?

    Edit: I see the set point temperature has just increased to 460 deg C but I think I saw the slightly increased pressure mentioned above before the change in set point T.

    • Bob Greenyer

      Hi Stephen,

      me356 will publish the 10s screen grab sequence at the end of the run – pressure at this stage is manual.

      I think he will now dial down the temp – put in more LiAlH4 (to try to up the pressure to near 10 bar) and re-start.

  • Bob Greenyer
    • R101

      Oh wow, this is just fantastic! I believe this going to crack LENR wide open now. And then some other stuff.

      (IMHO, this may even explain Gravity 😉 )

    • Gerrit

      http://animpossibleinvention.com/2015/10/15/swedish-scientists-claim-lenr-explanation-break-through/

      “The scientists are now preparing for a well-planned experiment with all necessary safety measures, ideally with a transparent reactor body since the effect according to the scientists releases a lot of light.”

      “Our model describes quite a natural process. It’s probably one of the main sources for maintaining a high temperature inside Earth, since there’s high pressure, high temperature and good availability of neutron producing elements [through this process] with basically unlimited resources of deuterium,” Lundin said.

    • Sanjeev

      Great ! Yet another good theory, this time with experimental support.

      • Bob Greenyer

        The latter is the most interesting aspect. We hope that since they have tested this, with claimed success and have patent priority, they will share their triggering/stimulation approach.

        • nietsnie

          In the meantime, maybe you could try to experimentally zero-in on a triggering resonant frequency by, starting with a frequency-rich power source, temporarily splitting the input current into two parallel sources and applying tune-able frequency filters to them. Maybe one side is low pass and the other side is high pass. Then slowly rotate the filters in relation to each other, re-combine the two previously split (now filtered) sources, and watch for temperature/power anomalies. More or less the guitar phaser approach to LENR triggering.

          In case the effect is also temperature dependent, you could alter your software to accomplish a step-wise increase in temp, followed by, at each step, a slow step-wise dual-filter sweep. Keep a record of each filter position during each step. That way, if the result isn’t immediately obvious, you can notice anomalies programmatically later that you can use to fine-tune a later approach. If you can trigger the effect repeatably, then you can subsequently analyze the harmonics created by the phaser – which would likely be totally dependent upon the characteristics of the individual power source – later.

    • evleer

      “using the lowest possible amount of energy [through resonance] to shake loose the neutrons”

      This sounds very promising! If it would become possible to start and control LENR reactions using just a car battery as input power, all those off-the-grid applications that we hope for aren’t that far-off anymore.

  • Bob Greenyer

    Celani_Rossi_Mashup_Run1_Full_Data

    https://goo.gl/pIw5t9

    me356 says

    “Reactor was taken for the measurement (still under the pressure of hydrogen) and resistance is just perfectly same as it was during the calibration.

    Measurement was performed at the same ambient temperature.”

    Graph from Ecco

    • Sanjeev

      The active run (end) shows a tiny bit more excess.

      • Ged

        It’s really interesting that the active run (end) excess disappears at 250 to 200 C, the supposed magic temperature threshold.

        • Sanjeev

          This also shows that the gap between the fueled and calibration runs is not merely an offset. However it needs to be confirmed once again.

          • Ged

            Very true, and very good news. Also, completely agreed. Replication with calorimetry is the best next step I can think of.

    • Ged

      Any plans for this reactor now, since it was so, apparently, successful?

      • Bob Greenyer

        Take it apart, add more LiAlH4, Heat it up again.

  • Bob Greenyer

    Celani_Rossi_Mashup_Run1_Full_Data

    https://goo.gl/pIw5t9

    me356 says

    “Reactor was taken for the measurement (still under the pressure of hydrogen) and resistance is just perfectly same as it was during the calibration.

    Measurement was performed at the same ambient temperature.”

    Graph from Ecco

    • Sanjeev

      The active run (end) shows a tiny bit more excess.

      • Ged

        It’s really interesting that the active run (end) excess disappears at 250 to 200 C, the supposed magic temperature threshold.

        • Sanjeev

          This also shows that the gap between the fueled and calibration runs is not merely an offset. However it needs to be confirmed once again.

          • Ged

            Very true, and very good news. Also, completely agreed. Replication with calorimetry is the best next step I can think of.

    • Ged

      Any plans for this reactor now, since it was so, apparently, successful?

      • Bob Greenyer

        Take it apart, add more LiAlH4, Heat it up again.

  • Sanjeev

    Experiment restarted. Its now at 400C, 65W, roughly the same as previous run.

  • Sanjeev

    Experiment restarted. Its now at 400C, 65W, roughly the same as previous run.

  • Bob Greenyer

    From m356

    “Results are very similar altough it looks like there is a sealing issue. Interestingly until I have opened the tube, there was still pressure from the previous run.

    When the pressure was higher, COP was higher than previous time.

    I will leave it for tommorow.

    LiAlH4 that I am using right now has low amount of stored hydrogen”

    • Ged

      So the reactor may be “bleeding out” slowly. Will actually be interesting to see if it dies, but we know low pressure is part of the normal evolution of the phenomenon in other settings.

      Wonder what a higher quality LiAlH4 would do.

  • Bob Greenyer

    From m356

    “Results are very similar altough it looks like there is a sealing issue. Interestingly until I have opened the tube, there was still pressure from the previous run.

    When the pressure was higher, COP was higher than previous time.

    I will leave it for tommorow.

    LiAlH4 that I am using right now has low amount of stored hydrogen”

    • Ged

      So the reactor may be “bleeding out” slowly. Will actually be interesting to see if it dies, but we know low pressure is part of the normal evolution of the phenomenon in other settings.

      Wonder what a higher quality LiAlH4 would do.

  • artefact

    From me356:

    “The behavior is very interesting.
    We can exclude sealing issue, because today, the pressure is noticeably higher than yesterday without any change. The same thing happened to my previous experiment when the pressure decreased on its own. So definitively something is happening inside even that the temperature is pretty constant.”

    • Ged

      A lot of sub 60 lows now, so it seems power is once more doing that oh so painfully slow drop. Seems about 1 W or so less than last night at least. Fascinating how consistent the overall behavioral tends between the runs have been.

      • Ecco
        • Sanjeev

          Is ambient being monitored ?

          • Ecco

            I think it is, but it’s not in the publicly available data from Plot.ly

          • Sanjeev

            Any correlation with the ambient needs to be ruled out.

            It appears that the power is going even lower and the temperature is waving a lot more, sometimes +/- 3C.

          • Ged

            Me356 reported in the last run that during the slow decrease in power, ambient temps had fallen (night). So far there seems to be no correlation in any direction, according to the reports.

            Hopefully we can get some raw data on that to verify, but ambient temps usually don’t change over a wide enough scale to affect power this much given the temps we are at and the current ~6.25 C/W (that is, to change the power by 1 W, temps would have to change by over 6 C, and I doubt ambient is capable of doing that in a climate controlled setting; and definitely not without anyone noticing).

          • Sanjeev

            Probably you are right. If I recall correctly the ambient affected the old Celani setups of MFMP a lot more than expected, and that’s what makes me cautious. Even the popular Celani cell demoed in ICCF, which showed 20W of excess suffered from errors due to ambient.
            So I hope Me356 will record ambient as well in his next experiment, just to be sure.

          • Ged

            I sure hope so as well. Always better to verify than guess/assume!

          • Mats002

            This temp shift is over 24 hours, ambient is normally correlated with sun hours, so that is not a plausible explanation. The power shift is not linear either, hmm.

          • Ecco

            To be fair, chances that it might be coincidental still exist. I tried including additional samples since the last graph and now I’m not sure anymore that the trend is real.

            http://i.imgur.com/jHRKkWo.png

            Every data point represents the average of 1000 samples from plot.ly (after clicking “edit graph” on the bottom of the live data page, showing the data and importing it in a spreadsheet).

          • Mats002

            The datapoints should be more distributed in the time axis, why no data points for many hours between first and second point?

          • Ecco

            Because the full raw data isn’t available yet. Right now one can only access about 15 minutes chunks of live data a time; past data cannot be downloaded from the plot.ly interface with the stream set up by me356.

          • Ged

            Oh, the trend is very strongly there, in this graph as well. The R^2 may be lower than before, but probably not by much. This is a very good curve fit, and thus strong trend, so far as graphed.

          • Stephen

            Do we have a good idea of the free volume and do we know what the expected pressure would be at 400 deg C if all the Hydrogen in the last load of LiAlH4 was in the form of H2 Gas?

          • Bob Greenyer

            Actually it is not – but is planned for next test. In the future – there will also be a second geiger counter to help distinguish signal from noise

        • Ged

          Thank you, Ecco!

  • artefact

    From me356:

    “The behavior is very interesting.
    We can exclude sealing issue, because today, the pressure is noticeably higher than yesterday without any change. The same thing happened to my previous experiment when the pressure decreased on its own. So definitively something is happening inside even that the temperature is pretty constant.”

    • Ged

      A lot of sub 60 lows now, so it seems power is once more doing that oh so painfully slow drop. Seems about 1 W or so less than last night at least. Fascinating how consistent the overall behavioral tends between the runs have been.

      • Ecco
        • Sanjeev

          Is ambient being monitored ?

          • Ecco

            I think it is, but it’s not in the publicly available data from Plot.ly

          • Sanjeev

            Any correlation with the ambient needs to be ruled out.

            It appears that the power is going even lower and the temperature is waving a lot more, sometimes +/- 3C.

          • Ged

            Me356 reported in the last run that during the slow decrease in power, ambient temps had fallen (night). So far there seems to be no correlation in any direction, according to the reports.

            Hopefully we can get some raw data on that to verify, but ambient temps usually don’t change over a wide enough scale to affect power this much given the temps we are at and the current ~6.25 C/W (that is, to change the power by 1 W, temps would have to change by over 6 C, and I doubt ambient is capable of doing that in a climate controlled setting; and definitely not without anyone noticing).

            Edit: It’s also way more complicated than just that though. Ambient is still much much lower than the temperature of the reactor, so heat loss is still more or less the same. For ambient to affect the power going into the reactor, the rate of heat loss would have to change such that the reactor is -losing- 6 C less heat to allow a 1 W decrease, not that the ambient is 6 C higher than where it was. That would require a lot more than 6 C of change in ambient, but not sure how to calculate that exactly. It’s why we’ve never seen ambient do anything to these high temp runs–low temp runs closer to ambient would be more vulnerable.

          • Sanjeev

            Probably you are right. If I recall correctly the ambient affected the old Celani setups of MFMP a lot more than expected, and that’s what makes me cautious. Even the popular Celani cell demoed in ICCF, which showed 20W of excess suffered from errors due to ambient.
            So I hope Me356 will record ambient as well in his next experiment, just to be sure.

          • Ged

            I sure hope so as well. Always better to verify than guess/assume! The Celani experiments were cooler if I remember right, too.

          • Mats002

            This temp shift is over 24 hours, ambient is normally correlated with sun hours, so that is not a plausible explanation. The power shift is not linear either, hmm.

          • Ecco

            To be fair, chances that it might be coincidental still exist. I tried including additional samples since the last graph and now I’m not sure anymore that the trend is real.

            http://i.imgur.com/jHRKkWo.png

            Every data point represents the average of 1000 samples from plot.ly (after clicking “edit graph” on the bottom of the live data page, showing the data and importing it in a spreadsheet).

          • Mats002

            The datapoints should be more distributed in the time axis, why no data points for many hours between first and second point?

          • Ecco

            Because the full raw data isn’t available yet. Right now one can only access about 15 minutes chunks of live data a time; past data cannot be downloaded from the plot.ly interface with the stream set up by me356.

          • Ged

            Oh, the trend is very strongly there, in this graph as well. The R^2 may be lower than before, but probably not by much. This is a very good curve fit, and thus strong trend, so far as graphed.

          • Bob Greenyer

            Actually it is not – but is planned for next test. In the future – there will also be a second geiger counter to help distinguish signal from noise

        • Ged

          Thank you, Ecco!

  • Sanjeev

    Would installing more wires, say, 4-6 of them, increase the signal ?
    If yes, I suggest doing it next…

    • Ged

      I agree. If each wire contributes a set amount based on available sites of activity, then adding more wires should increase the effect. Likely additively, though always possible there could be some synergy. This is the joy of a wire design–great idea.

  • Sanjeev

    Would installing more wires, say, 4-6 of them, increase the signal ?
    If yes, I suggest doing it next…

    • Ged

      I agree. If each wire contributes a set amount based on available sites of activity, then adding more wires should increase the effect. Likely additively, though always possible there could be some synergy. This is the joy of a wire design–great idea.

  • Ged

    Looks like the power-down-over-time trend has finished and bottomed out. That is, the reactor has reached its maximum state for this power in. Again, almost identical to last time’s behavior.

  • Ged

    Looks like the power-down-over-time trend has finished and bottomed out. That is, the reactor has reached its maximum state for this power in. Again, almost identical to last time’s behavior.

    Edit: One thing I am liking about Me356 methods is how long he lets the reactor run at steady state.

  • Obvious

    I wonder if this design makes something akin to a low pressure sodium light bulb. A quartz tube might be worth a try to get a look at what goes on. Unless the alumina is neccesary for this effect.

  • Obvious

    I wonder if this design makes something akin to a low pressure sodium vapour light bulb. A quartz tube might be worth a try to get a look at what goes on. Unless the alumina is neccesary for this effect.

  • Stephen

    Do we have a good idea of the free volume and do we know what the expected pressure would be at 400 deg C if all the Hydrogen in the last load of LiAlH4 was in the form of H2 Gas?

  • Stephen

    Good luck with the current test… Do we know what the current set up is? Is it using Titanium wire this time?

  • Stephen

    Good luck with the current test… Do we know what the current set up is? Is it using Titanium wire this time?