MFMP Runs New Live Glowstick Test [Update: Test Ends after Component Failures]

A new live test of the Martin Fleischmann Memorial Project’s Glowstick reactor has started in California, being run by Alan Goldwater. Earlier this they ran a test on the same reactor that is being tested tonight, but only heated it up to approximately 960 C.

From what I have heard, in this test they will raise the temperature higher — to the levels reached in the test they did in Padua during the ICCF 19 conference, where the reactor reached over 1300 C. It’s going to take many hours before they reach those temperatures, but it will be interesting to follow the progress of this test on the livestream below.

Alan said this about the reactor before the test began:

“After reconnecting the data system, I find that the fuelled reactor still had 13 psi of positive pressure after sitting idle for 28 days. The pressure was 32 psi at the start of this period, so there has been some permeation into the Alumina and/or continued absorption by the fuel. ”

Livestream #1

Livestream #2

Livestream #3

Livestream #4

UPDATE: The test has ended. Here is Alan Goldwater explaining what happened.

“The thermocouple finally failed at 18:00 UTC and appears to have fried the input side of the HUGnetLab board and possibly the PID controller as well. So obviously the test was stopped at that point.”

“After everything cools down, I’ll do some surgical disassembly (hammer and chisel) to see what caused it. The cell will be further disassembled later to extract the fuel for analysis, and the data file will be posted for analysis by the crowd. Thanks for watching!”

  • ecatworld

    Many thanks to those who are entering data into the spreadsheet. I don’t know who is doing it, but it’s not a job for one person for an extended period of time. To keep the ss up-to-date we’ll need round-the-clock monitoring of the YT feed. Any suggestions on how we can make sure we cover the test?

  • pelgrim108

    Watch Live Data stream in of a new replication experiment by Me356
    http://www.lenr-experiment.tk/

    Experimenter on this Forum thread: http://www.lenr-forum.com/forum/index.php/Thread/1275-Reactor-parameters/?pageNo=6&s=1c9cd2e643171a63db006a97ecb093171835c92f

  • Alan Goldwater

    I’ve been puzzling over the test failure yesterday. Two specific questions needed answering:

    1) What caused the short and resulting arcing. Physical evidence points to
    what I proposed earlier – wicking of the melted tie wire into the
    porous alumina cement. The cement itself was more than 2 mm thick
    between the tie wire and the heater coil, and should have been good
    insulation. The wire was fully embedded and couldn’t move even if
    broken. The amount of arcing was surprising, given that the heater
    voltage at mid-coil was only 42 v.a.c when it happened. So big current
    flowed.

    2) Where was the ground return path to sustain arcing over 2 seconds with substantial
    current. The block diagram below shows the same ground path we
    noticed in Padova – from the laptop power supply through the RS485
    connection to the HUGnetLab board. While RS485 data is opto-isolated,
    the HUG-supplied wiring includes a ground connection, which extends
    through the data system to the thermocouples and other input grounds as
    shown by the dotted green lines in the diagram.

    I can only speculate what might have happened if this ground hadn’t been
    present, but it probably protected the laptop from destruction. This
    suggests a straight-forward safety fix, simply add GFCI to the system
    power input as shown in red on the diagram. In the event of a shorted
    heater thermocouple, the whole thing would shut down, better than a
    burnout. But better TC insulation should also be part of the fix, and to
    prevent the suspected wicking of melted metal, I’ll use a chip of
    nonporous alumina tube between the loop in the heater wire and the
    underlying TC tie wire, then embed the whole thing in cement.

    • Obvious

      Have you verified the temperature readings with both devices hooked to the TC at the same time?
      Seems to me that dual measurements would skew the TC voltage a bit.
      (doesn’t fix the short problem, though)

      • Alan Goldwater

        Yes, I confirmed that during testing at 100 C in boiling water. The two inputs were calibrated within 1 degree C of each other, and connecting either or both at once made less than 1 degree change in the readings.

        • Obvious

          Good. I was wondering if maybe the heater was hotter than it looked.

          • Alan Goldwater

            It probably was toward the end – with 900 watts power it would have been around 1300 C inside the core and 1000 outside. The thermocouple junction had moved by then, so it only showed 889 C just before the failure. Full analysis is coming…..

            • Obvious

              I nearly duplicated your problem tonight. I was using a 4 channel TC meter, and got scrambled readings, but no damage. The TC insulation melted at a coil wrap. The meter kept on recording normally once power was cut off. ( I havend downloaded the data yet, should be interesting).
              I was using a 500 W J Type T3 bulb to simulate excess heat. Instant failure of the ceramic tube when the bulb was hooked up (I was having issues with the dimmer, so I bypassed it, so thermal shock is probably the culprit.) but the coil was OK. Nice glow with coil shadows, instantly. I had glass frit (melted from some tests) to conduct heat from the coil better to the tube, which liquefies at about 720 C (wets from the coil to the tube, mostly eliminating hot spots). It sealed the cracks in the ceramic, but ultimately dripped onto the internal bulb and it shattered, but the filament kept working(!!). Anyways, a steady voltage increase to the coil (not at all powered) occurred from moments after heat was applied internally, steadily climbing, from about 40 V reaching up to 87 V (AC) peak (Ghost voltage?). I had a voltmeter hooked up to the coil, because I was planning on firing up the external coil once the internal heat stabilized, to see what happened at various outer temperatures (IE what happens when the internal heat is less than the outer heat, since it supposedly can’t go to the higher heat area, (2nd law). I hadn’t plugged in the coil to power yet. weird. But now the whole thing needs a rebuild.

              Glass coated coil

              • Obvious

                Glass coated coil

                • Obvious

                  Internal heat. The bulb is already broken, and is running still, on RH only. No power to outer coil.

                • Obvious

                  Melted HT TC insulation. Glass melted and ran inside, and to bottom in drips.

                • pelgrim108

                  Thanks for trying to replicate. Do you have a place where I can view updates on the progress?

                • Obvious

                  I am still fiddling with designs. The glass coat was a neat idea, but I need to add some Al2O3 powder to harden it up, or encapsulate it in ceramic so that molten glass stays put. This was supposed to see what excess heat might look like, but I ran out of a couple parts at midnight, so I didn’t use a dimmer for the inner heat (bulb) which I had planned.
                  The outer coil runs on a 1000 W dimmer, but that has a crappy range compared to the 600 W one I had originally. The 1000 W unit maintains a constant 10.9 V voltage with current, so the coil actually ran up to nearly 190°C with the dimmer at full dim. And then it only went up to 98 V at peak, cutting off huge amounts of power. (The 600 W dimmer went from 0 to 116 V, so it had a great range but was getting really hot.) I am getting a 4000W 240V SCR heater control to fix this up. 120 V isn’t enough to get enough wraps and the high heat with a stable wire diameter, unless I go to massive wires and quite low voltage, and huge current.

                  Edit: Also both TCs shorted, or the current went through all 4 channels, or both happened (two empty TC channels show high temperature at the short). No sparks were seen in this event.

                • Obvious

                  Graph.

                • Obvious

                  Glass coil graph internal TC. Before destructive test.

                • Alan Goldwater

                  The reactor I just ran and the one for Padua both used a cement made with Al2O3 powder and Na2SiO3 (“water glass”) solution. It sets in a few hours at room temp and cures nicely at 200 C. But at around 1100 C the silicate component becomes glassy like your coating, and then then the bond apparently fails. I have some Jyalucem-20 very hi-temp cement ( http://jyoticeramic.com/content/cement_jyalucem.htm ) and will use that next. It needs 24 hours drying time before curing at 200 C, but should hold up better with a maximum service temp of 1500 C. It’s meant for this kind of application, heaters and halogen bulbs. Let me know offline if you want some to try. alan at magicsound dot us.

                • Obvious

                  This glass is intended as enamel, and stays permanently attached. It is melted on, with no setting or curing. I can barely sand it off, and cant scrape it off except real thick bits, and not at all once the bubbles float out. It’s plenty hard at room temperature. Just a bit too runny at high temperature. If it stayed viscous enough it might be OK, but I’m not sure that it is an idea I will “stick” with. I was going to sprinkle some thickeners onto the hot coil at some point as a test, but it failed first. The dark glass does have a good emissivity, though.

                • Alan Goldwater

                  Yes, very difficult to make anything reliable at over 1000 C. I found that the TC leads need to come straight out from the tube rather than wrapping around. Thus the need for the tie wire, to keep the junction in good contact with the ceramic. I tried all kinds of clamps and fixtures, but none survived thermal cycling except the wire with cement encapsulation. And that technique now seems to have a limit of around 1200 C, where the thin Kanthal tie wire melts. So better cement and small pieces of dense alumina will be tried next. This is a critical piece of technology for a stable and measurable system.

                • Obvious

                  I am going to use those ceramic mini tubes for aquarium filters, and drill a crosswise hole though it to insert the TC. This should make a good standoff, and might also make a good black body hole without the hole drilled in the side. The standoff idea is something like this:

                • Obvious

                  Quick update on the “weird voltage” I had earlier, and the TC short.
                  I rebuilt another glass-coated coil, and it runs within 2% of the original for heat output.
                  It looks like the weird voltage measured on the un-connected outer coil may have been from conduction through the glass, caused by the liquid glass entering the tube through cracks and onto the internal heater wiring, allowing a circuit to the outer coil. Since the glass leak was in the middle of the tube, the outer coil resistor becomes a voltage splitter, halving the voltage. Conduction through the glass from the outer heater wire when activated might also explain the strange shorted TC readings also. The TC does not seem to have actually connected to the metal coil in the forensic examination.
                  This lead to some interesting speculations: That Joule heating of the glass, if carefully done, could result in a more even heat distribution, and an apparent reduction of resistance to the heater coil, similar to Lugano. This would be a sort of “doping of the Inconel” that would appear once the glass or similar material began to conduct. Plain ceramic won’t conduct, but a thin layer of just the right composition of glass-like material, hidden under ceramic, could. Measuring the resistance due to this effect could be Rossi’s “new method” of temperature measurement. A similar material could be added into the inside of a thin TC protection tube, with two non-connected wires, and be an effective RTD.

                • Obvious

                  Coil working nicely. These were heat-power-flux tests, so no fuel.

  • Bernie Koppenhofer

    Too bad he has to go through the same learning curve Rossi went through 3 or 4 years ago, or the same learning curve, I assume, NASA has also conquered. Why?

    • Teemu Soilamo

      I don’t understand. Why not just replicate Parkhomov exactly? Is he not providing the details? If not, why?

  • scientist_1

    only after the hydrogen adsorption.

  • Stephen

    Hi Alan G
    Is the radiation detector you use just a gamma ray counter or is it also a gamma / X-ray spectrometer?
    I’m just thinking if you do detect radiation of some kind at particular times during the experiment the spectrum might be interesting to help find out what is going on inside, it might help identify excited states of isotopes for example or if it also able to operate at 10s or 100s KeV show if X-ray Bremstrahlung radiation is occurring.

    In my work I have often seen that even when test “fail” they give a wealth of peripheral information which we might not otherwise have an opportunity to see.

    Good luck with your remaining tests. I’m impressed with your methodical approach.

    • Alan Goldwater

      Stephen, the detector is a simple Geiger Counter, a GMC320+. I’ll post the data log along with the temperature and pressure data from the experiment, some time next week. A sample of the fuel will be sent for analysis eventually, through service donated to MFMP. Not sure of the details yet.

      • Stephen

        Thanks for that Alan. I’ve been wondering how I can contribute something to your work. Since I’m not an engineer I’m not really able to do a replication. But maybe I can contribute something towards a gamma/ x-Ray spectrometer if it is useful to you. Can I do that through the MFMP site?

        I’ve looked on the Internet and there are a lot of options available with different sensitivities, and some with particle detectors as well but also a large range in prices. But some seem affordable. I’m not an expert so maybe you would want to choose the right one. I think an ideal one should work at low keV (10s keV) as well as the higher energy gamma, and may be if possible detect neutrons as well.

        Maybe if there is someone who knows more about these devices they can recommend an affordable model?

  • Gerrit

    I am happy there are a growing group of experimenters having a try.

    The e-cat looks very real to me, but we’ll have to wait for at least another 7 months and even then we might not know for sure.

    Parkhomov might present more and better data in future, but we’ll still need a widely reproducible setup for success.

    I am also not interested in the failures, but I would be very interested in an unambiguous and repeatable success. Unfortunately the failures are part of it.

    • Bob Greenyer

      Every step, every slip up, every fall one makes as one tries to climb a mountain is a failure to reach the top. Only getting there is considered a success, but the story of the journey, the documenting of the pitfalls, the dead ends the bad choices of direction or equipment, when documented and shared, can help others to safely and surely reach the summit.

      This is an adventure, achievement of the goal would be nothing if it was easy.

  • Sanjeev

    FYI all, Denis Vasilenko (Firax) will start his new experiment on Saturday. Stay tuned for more. These are fun days.

    • Bob Greenyer

      Great news!

  • Sanjeev

    AlanG,
    Thanks for the efforts, I must say you are very patient with these slow experiments which takes hours and days….

    For the next build, I suggest using two TCs for redundancy, at least one that is type-B, since you are trying to go 1200°C. The TCs must not cross paths with the heater wire and should be secured in place with alumina cement instead of a metal wire.

    I guess these are the obvious improvements that you already know. I also suggest trying a closer replication to Parkhomov’s setup. Using an SS tube to contain the charge, which will minimize hot spots and reactor tube failure. Try to have a single outer tube rather than a split design and try to match the dimensions to Parkhomov more closely, so we have similar pressures inside, which will minimize the risk of leak or explosion due to high pressures.

    He has already arrived at a good design which works, so there is really no point in inventing a new design with new points of failure every time. Now that you have his exact powder, the chances of success are more, so lets save time by learning from his design.

    • Alan Goldwater

      Good points Sanjeev. As you can see from the photos, my design does use
      alumina cement to encapsulate and protect the TC. The design has worked
      well three times now, but only for a single heat cycle. A better cement
      should help and further tests will use one.

      The image shows an arcing track through the porous alumina cement, probably caused by migration of the melted Kanthal 30 gauge tie wire. Who would have guessed this could happen?
      http://magicsound.us/MFMP/burnout_autopsy.jpg

      Parkhomov’s most recent design uses three alumina layers instead of my two. He basically
      builds a small tube furnace type heater, with alumina inner sleeve and
      outside cement coating. Then he inserts the alumina tube cell to hold
      the fuel, with the TC sandwiched between it and the heater inner tube.

      It’s a good design and worth emulating, but it separates the core from
      the heat source with another layer of alumina, so it needs higher power
      density to reach the temperature required. If you recall, his last
      attempt to reuse a heater also failed before completing the test.

      At these temperatures there is no easy solution. The Silicon Carbide
      heaters Bob Greenyer proposes will handle the temperature, but mounting a
      thermocouple will be even harder. And type B thermocouples are $400 a
      pop and are quite fragile. We destroyed one in a test at HUG in
      February.

      • Sanjeev

        If the tie wire was really Kanthal then its a miracle that it melted at such low temperature… or did the temp went above 1200°C at that spot ? Its puzzling, and if it did reach 1200, the heat must have come from inside, which is amazing if true, but its a speculation at best.

        Anyway, this design has revealed a point of failure, which needs to be eliminated somehow. If you wish to avoid the expensive type-B, then the only way is to use a type K at a cooler spot and infer the core temp via measurement. This can be hooked to the PID. A second, “sacrificial” K-type can be at the hot spot and if it burns out (which it will at 1200C), nothing will be affected.

        The last experiments by MFMP and others have revealed the issues one by one.
        1 – Heater burn outs (Solution : use SiC, which you say you will)
        2 – TC burn outs (Solution: use B-type, or K-type at cooler spots)
        3 – Pressure leaks (You have solved this nicely, no more an issue)
        4 – Reactor tube failure (Solution: use a thin SS tube to avoid hot spots and alumina melting).
        5 – Instruments burn out (Solution: proper isolation)
        6 – Junk data (As happened in Padua with power data, solution : Proper checks)

        What has happened is, most of the experiments did not reach the stage of “sweet spot” and failed earlier due to one or more such reasons. You already have a robust design , it just needs some more tweaks before it can become a reliable apparatus. I agree that Parkhomov’s setup is not free from such issues, but he has more success rate. He went through all these issues one by one and arrived at a solution. We need to take it and start from there. Of course, if these problems can be solved in a better way , that will be ok too.

        • Bob Greenyer

          Good point Sanjeev – there are a good number of groups looking to start testing more developed versions of this design – like SKINR and Open Power Association – so it is just as well we are sorting out all the bugs – and having the keen eye of you and others on solution finding will help us get there faster – thank you.

        • Bob Matulis

          Hi Sanjeev, a while back some people mentioned the use of laboratory tube furnaces for testing. Would heating from the exterior of the core eliminate hot spots and be a plausible option in testing? http://www.carbolite.com/products/furnaces/ If the oven temp is set to 1200 C over time it would become evident if less power was required to maintain temperature. Thanks, Bob

          • Sanjeev

            Yes, for those who can afford it, a tube furnace is a ready made test bed with everything in place, except the reactor core. If the local hot spot is due to lenr, then the method of external heating will have no advantage, the core can still fail if it gets too hot at a single point. That’s why Parkhomov used the SS tube inside the core for uniformly spreading the heat that came from lenr. The lenr generated heat will be the major source once the reaction starts, the heat from external heater will go down as the power gets automatically reduced, so protection of the core due to lenr heat is the point here.

            • Bob Matulis

              Thanks for the reply. You mention the hot spot at a single point. Since the method of heating is not introducing heat in the inside wouldn’t a hot spot failure be indicative of LENR (compared to a method that introduces heat internally)?

              • Sanjeev

                I guess it would indicate that some anomalous event happened, thats all. If the hot spot failure is accompanied by a drop in input power then it will be a more believable evidence. But will still not be good enough to rule out any chemical origins. To rule that out, the setup must show a drop in power for many hours. And that’s why it is a necessary requirement that the core must not fail as soon as the lenr starts.

        • Alan Goldwater

          Sanjeev, your summary is a good one. I have to point out that this is primarily an engineering problem, where the usual trade-off is time vs quality vs cost. An old joke say “quick, cheap, reliable – pick any two”. This case is no exception, and with a modest budget cheap and reliable means sacrificing quick. We solve the problems as they arise by looking for clever and robust solutions to challenging design problems. Every test leads to progress, as you others have pointed out.

        • Obvious
  • Ged

    Can always build your own, if you know a better design and set up to try. The materials are pretty cheap–nothing to stop you there hopefully.

  • Zeddicus Zul Zorander

    I feel your pain bro…

    I do think you underestimate how difficult this still is. Although we now have a reasonable understanding what is needed to replicate the effect, the material control, the way to heat up the reactor and dozens of other parameters are still very difficult. This is were Rossi has the advantage.

    It also shows how good Parkhomov’s replication was. Did you know he also had dozens of failures before he came up with a working replication?

    These MFMP guys are really, really competent in what they are doing. And although I’m convinced Rossi has a working product, tor me the ultimate confirmation of the Rossi effect is their replication. I have every confidence that they will succeed at some point in time. And at that time also, the whole world will be able to replicate their success because that is what their goal is all about.

    Even though superficially a replication seems quite straightforward, MFMP shows us it really is difficult to define all the parameters for a repeatable successful replication. But when they figure it out, things will really speed up. This is what I believe.

  • James Thomas

    “I have not failed. I’ve just found 10,000 ways that won’t work.”
    ― Thomas A. Edison

    “Success is not final, failure is not fatal: it is the courage to continue that counts.”
    ― Winston S. Churchill

  • hempenearth
    • http://magicmusicandmore.com/ Barry

      Great vid hamp, thanks.

  • Ged

    Lucky; too bad I missed the fireworks 😉

  • magicsnd1

    The thermocouple finally failed at 18:00 UTC and appears to have fried
    the input side of the HUGnetLab board and possibly the PID controller as
    well. So obviously the test was stopped at that point.

    After
    everything cools down, I’ll do some surgical disassembly (hammer and
    chisel) to see what caused it. The cell will be further disassembled
    later to extract the fuel for analysis, and the data file will be posted
    for analysis by the crowd. Thanks for watching!

    • Ged

      Thank you for the run and putting on the show! Sad the little thing died and took equipment put with it. We need to find a way to prevent all those expensive losses.

      • magicsnd1

        Yes, it needs a better (and safer) way to attach the thermocouple. Expense-wise, the PID controller is about $70. HUGnetLab data board was donated by HUG and I have a spare on hand. If it had been a NI Labview system, that would be another story!

        So an important lesson was learned at reasonable expense, pointing out a key advantage of open science done this way.

        • Ged

          So very true. It’s a wonderful learning process for all.

    • Bob Matulis

      Keep up the good work! New info for improvement.

  • magicsnd1

    I’ll try to answer some of the question backlog. This test is a reheat of the same reactor I ran at the end of April. It’s physically identical to the one we used in Padua, and has the same fuel mix but from different sources – Hunter AH50 Ni and Consolidated Chemical LiAlH4. I’m using roughly the same temperature profile as the Padua test.

    The data display is near-realtime, from a Linux-based HUGnetLab system and extra calculated display like power vs temperature can’t be added in real time. That and more will be done after the test is finished.

    The “explosion” mentioned was a momentary arc from the center of the reactor, where the heater wire crossed over a very thin (30 gauge) kanthal wire holding the thermocouple in place. I suspect the cement around it cracked, allowing the tie wire to touch the heater, and immediately vaporize. The DAQ and control systems seem to be operating after that, so the experiment is continuing on the scheduled temperature profile, targeting 1000 C about an hour from now.

    • MontagueWithnail

      Thanks, should have known there would be a good reason.

  • Ged

    Wait, what do you mean “exploded”? I just went to lunch and can’t scroll back in time with my phone. Seens it is smacked up against a PID holding temp at 700 C for outside core, thus making voltage noisy as the heater turns off and on to hold temp. (if I am interpreting right and there is a PID involved)

    Can see a healthy glow on one side though.

  • bfast

    Question, what would “excess heat” look like on the graph?

    • pelgrim108

      Average Heater Voltage going down and Core Temp going up.

      • MontagueWithnail

        Agree, but the different scales are re-adjusting to keep all the lines in the same visual range so I don’t think it will be easy to notice from the graph. Maybe Alan can be persuaded to add a new line of core temp as a function of V – perhaps with some averaging to account for the more erratic nature of the voltage – then we can see if/when the relationship starts to change.

        • pelgrim108

          I agree with you, and would also like core temp as funcion of V line in the graph.

  • Bob Matulis

    Question: What is being done differently in this test compared to the one recently done in Italy? The test done in Italy did not show any excess energy produced. I am hoping for positive results – good luck on this test!

  • magicsnd1

    I’ve started a new Youtube stream (#3) for the GlowStick2 restart test now running: http://youtu.be/oNBLgO8FhEk

  • Gerard McEk

    I hope you will succeed now Alan, it is about time for a geat success for MFMP!
    Just a question: The glowstick seems separated into two halves, what is the reason?

  • pelgrim108

    Alan, can you please enable the text chatting ( chatbox ) for your next live stream on youtube.

  • MontagueWithnail

    Good luck Alan, I hope the experiment is successful – by which I mean that you are able to complete the test – without any failures of the equipment or significant experimental error – and draw solid conclusions. I also desperately hope you find excess heat, and it is appalling bad scientific form for me to say so, but I’m not professionally a scientist so I don’t care : )

    The resolution on the stream is much better than last time but I think the occasional screen shots would still really help people to follow and understand.

    • Ged

      I second you on all points. Kick butt, Alan!

  • LCD

    Off topic a little, has anybody done or is there a plan for a pre/post isotopic analysis of parkhamov yet?

    • Josh G

      There already is. He sent his ash for analysis, and it came back showing isotopic changes. But that was from only 1 lab. He apparently sent samples to two other labs (I think), and those results haven’t been released yet AFAIK. Or maybe he hasn’t gotten those results back. The MFMP glowstick right now is to continue ‘baking’ the Parkhomov fuel they tested in Padua. They want to heat it for awhile, probably a few days I guess, and then send it also for isotopic analysis.

      • Ged

        It’ll be interesting to see if the behavior of a restarted reactor is any different to the virgin run.

    • Bob Greenyer

      Good news – Peter from NeoFire has received the Parkhomov Nickel I sent him and will do a ToF Sims on it – so we will have isotopic data for the nickel to compare to Parkhomovs independent lab data. He is also hoping to do Auger electron microscope analysis also.

      SKINR intends to do ICP-MS

      • LCD

        Any news