MFMP to Run New Glowstick Test (Update, Aug 26th: Fueled Run Now Live Streaming)

The Martin Fleischmann Memorial Project is preparing to start a new test of their latest reactor — the GlowStick GS4 — which is scheduled to begin on August 24th.

A description of the reactor is on their Facebook Page:

“The cell will use a Coorstek mullite tube in differential configuration like GS3: two zone heater coil, one side filled with an alumina rod and the other with fuel.”

“The instrumentation and power control will be like GS3: sine wave AC with variac and PID controller, temperature sensed from K-Type thermocouple. As Alan does not have a suitable power analyser, this is the method available that will yield meaningful power measurements.”

Here’s a diagram of the reactor:


Data will be made available live via the HUGNetLab when the test begins. This post will be updated when the test begins.

UPDATE: (Aug 26, 2015)

A live, fueled run has just begun. The video stream is below, and data can be followed also at Bob Greenyer has said that this setup has the “best calibration ever between “active” and null side”.

  • magicsnd1

    The complete GS4 data set is now available as csv files:

    There are separate files for the calibration, the experiment and the reheat.

  • magicsnd1

    The GS4.2 reheat experiment data has finally been posted at
    Click to VIEW FC0600 at the bottom of the test definitions list. Then click
    on the “From” time and set it to 9/17 03:00 using the slider controls.

    The pressure plot illustrates the reversible LiH decomposition and
    recombination over the range of 425-1050 °C in the core (about 50-100
    degrees hotter than the outside temperatures shown).

    • Mats002

      Thanks Alan! I did not understand the hangout but I now know about the chat function.
      Next time…

  • Stephen

    Is the current GS4.1 reheat test planned to be displayed on Hugnet and Youtube or is it only available on Google Hangouts?

  • Stephen

    Can someone remind me what happens to the other gasses (Nitrogen, Oxygen, CO2 etc) present in the sealed container?

    Could this be part of the remaining gas or are they removed from the gas at high temperatures by reaction with the Nickel and Aluminium (would dry Nickel make a difference)? Perhaps all the hydrogen is absorbed/ adsorbed or converted to Rydberg matter or something but the other gasses are what remain?

    If a steel container was used would it react with the other gasses.

    Do we know the composition of the remaining gas in the device? Or can it be found during degassing or opening?

    In case low pressure is required. Would degassing before or during the experiment after the hydrogen is processed remove some of the nitrogen and oxygen?

    Assuming all the Hydrogen is removed from the gas in some way. I think if we assume the ideal gas law for the atmospheric gasses then a 5 * increase in Temperature would result in a 5 * increase in Pressure assuming no absorption or outgassing and constant volume. If some absorption of the atmospheric gasses occurs or the volume changes due to the reactions then the value will vary from this.

    Would it be worth degassing the current gasses from the device and restarting the test with the hopefully already adsorbed or converted Hydrogen?

    FreeThinkers recent test run reported on LENR Forum fumes to mind here.

    • Bob Greenyer

      They get sequestered to oxides and nitrides of Al that are refractory

      • Stephen

        Thanks Bob, I remember the atmospheric gasses were accounted for somehow but forgot exactly where. I suppose there would be plenty of Aluminium in the LAH and tube to account for the extra gas in the greater free volume

        • Bob Greenyer

          It depends on your dead volume!

  • Preston Brown

    So, what is the conclusion? That 60C delta was real constant for most of the run, is that just tolerance on the sensors or a sign of something? That first run up to 1000C was looking promising; but is it all still inconclusive?

    Also, given the construction what would cause a constant 60C offset – if the resistance of thecoils was different then the offset would vary with input voltage, right? So maybe something else?

    • magicsnd1

      @Preston: We did two separate runs. The second one
      (GS4.2) used a new heater coil and mullite reactor cell, and was
      carefully calibrated before the fuel was added. The analyses done by GED and
      ECCO used the correct calibration data and found no excess heat, with a
      data resolution of about 1% of T (±10°C). The calibration data and charts for both runs are available at

      Your question points out that the construction of the
      Glowstick design leads to some variation in the balancing of the active
      and null thermocouples. This must be accounted for by calibrating the
      two thermocouples against steady-state input power before the fuel is
      added. Here’s a thermal camera image from the calibration of the second reactor. It provides experimental evidence that the difference in TC readings
      is due to the conduction of the longer alumina filler rod in the null
      side of the cell.

  • Obvious

    Please fix the diagram inner and outer diameters. Either the coil tube outer or cover tube inner diameter is incorrect. There is no room for the heater coil as depicted.

    • magicsnd1

      Thanks for spotting that! Here’s an updated reactor drawing with details of the GS4.2 fuel. I also attached the GS4 system block diagram showing the data and control systems.

  • ss dd

    Great work everyone!

    I think there will always be some doubt as to whether we are seeing the Rossi effect until we obtain a decent COP. Rossi on his blog recently answered the question “In the early years of creating the e-cat, what kind of COP did you obtain before yelling “Eureka”?”: “about 1.5”.

    I’m going to make the wild assumption that a COP of 1.5 is obtainable by using the information in the latest patent. Though it might take a few iterations and we want to make sure that we are able to detect smaller COP values, e.g. 1.1, so that we have a baseline to start with and improve upon iteratively.

    IMO, a power analyser is a must and we will end up scratching our heads every time until we have a clear idea of the energy consumed. Buying one seems to be in the works.

    In addition, the latest experiment by me356 has shown that restricting ourselves to thermocouples might lead to missing a lot of what’s going on. IR measurement seems like another “must-have”.

    Alan has been making many breakthroughs in terms of experiment setup, reactor design, open science etc. However, until he is fully equipped my bets are on me356 to be the first to demonstrate the Rossi effect in a live open experiment.

    It sure would be nice to have the euro and the US team be at the same level of equipment. I already put my money where my mouth is and donated $500 back in January for lugano-inspired replication attempts, as well as another $100 3 months ago to get Alan a power analyzer. Good to see that others have been contributing recently.

  • builditnow

    KeV, I agree. Alan’s work is accurate and polished. For instance, the heating coil winding were very neat and regular, wound with the aid of a lathe.

  • builditnow

    Well done Alan and team.
    I visited Alan and witnessed the experiment directly for some of the time and had deep and wide ranging discussions with all of Alan’s team. I have to say that this experiment was excellently executed in every detail as planned, as I see it.

    The experiment was run on the fuel that Parkhomov gave MFMP at the last cold fusion conference. In many ways this test is identical to Parkhomov’s successful test, so, what is the missing ingredient?

    Was it the fuel container?
    There is a video from the conference where Parkhomov gave his “big reveal” and in that conference he was asked what kind of material the fuel container was made of. The answer was “stainless steel”, but, no answer as to the type of stainless steel. Very thin stainless steel I think with holes poked in it with a needle (if I’m remembering correctly).

    Was it the pressure which decreased to half atmosphere in Parkhomov’s case?
    Is it the identical fuel?
    Was there air in the reactor when it was fueled and sealed?
    Was there cement used in the reactor and what type was it?
    Was it the triac type power supply (I think Parkhomov used)?

    It looks like we need many more experiments.
    Perhaps Alan’s system can be more sensitive to less excess heat as there may have been some in this experiment but it was not clear.

    We may have to look for far less excess heat in order to what factors improve excess heat output.

    • Ged

      I still wonder how much if any effect the lithium’s ability to ooze out of the fuel pellet and get between the mullite and nickel tubes could have. We’ll have to see in the post mortum what it looks like.

      Still, like GS3, there was a clear and discrete state change that pushed the device well above calibration. That change also mostly died out after a day–all of which suggests, if not definitively, that there was a excess heat reaction with a sharp beginning boundary and slow trail off 24ish hours later with a strength of more than 100 W.

      • builditnow

        Ged, thanks for your analysis:
        Talking to Alan 100 watts is very close to the margin or error so the approximate 100 watts is “perhaps” exothermic. Also, the length of time means it could be chemical.

        I’ll talk to Alan about increasing the sensitivity to smaller amounts of exothermic heat. My initial thought is more fuel and insulation using fire bricks. Then add forced air cooling. There is a nice “soft brick” rated at 2600C, easy to drill / cut / carve. GreenTherm 26 LI made by ANH Refactories.
        I have some, they are very easy to work with.

        The advantage of insulation is that less power is required to heat the reactor as a result, small amounts of exothermic heat with cause a larger temperature rise. This way the experiment is more sensitive. At first thought, it seems the experiment could easily be adapted this way.

  • Ged

    Last set of graphs for this run. They don’t really say anything, but they sure are amusing. Starting a little before the power throttling and all the way till shut off (but not completely through cool down, as that is still going, just 8/31 0005).

    Active-Null (square spiral graph!):

    Power v Temp (circular spiral graph, even prettier):

    Obviously, this analysis fails during these ramps, but gotta show it for completeness.

    • Obvious

      Thump. Thump. Thump.

    • Justa Guy

      Ged, these are great plots to spend the time/care doing (as well as all the plots various people are submitting; Ecco & Obvious plotting come to mind, immediately), even if they may look silly!

      The first plot is a great example of what Heat Capacity/Flow is all about, which can be extracted (partially) from the second plot. If perhaps with a little more detail revealed during the experiment (which I witnessed first-hand), you/one could redo the first plot and show that it almost all will sit on that straight line on the plot. There is much recovery/lag of this crude calorimeter, and you/one might need to know the exact ambient temperature of Alan’s Studio Area at any time, to make it happen.

      During the run, Alan mentioned your plotting and his immediate analysis of it to me and hence the additional above comments…

      It would be extremely useful if we record temp/humidity vs time near these experiments and save/analyze that data, also. Alan may even be doing that [unknowingly] with his cell phone or some other fixed gadget in the lab/studio.

      Thanks to all for sharing the plots!

      • Ged

        Thank you for the kind encouragement. Without Alan and Skip, and the other supports such as you and builditnow, there would be no data to gnaw ;).

        Knowing the sum total mass of the reactor, the longitudinal thermal conduction coefficient of the mullite and nickel tubes and their heat capacities, and then the ambient temps and humidity as you point out, it should be possible to mathematically compensate for lag times during powerups/downs. This would also smooth out the data in general too, and probably remove the need for double moving averages for trend extraction. It may take something a bit better at number crunching than excel to handle.

        Just heat capacity, mass, and ambient alone would allow a lot of approximate ramp effect compensation, and provide extraction of the data during those events. Even kinetics perhaps, as we have the full behavior of the active side sans hydrogen.

        A few partial derivative heat equation simulations in matlab would do it too, but also provide a complete model description. Interestingly, having such a model would be a sensitive comparison method. But, it is beyond my current equipment and capabilities.

        For now though, builditnow’s suggestions to increase sensitivity would go a long way in improving the methods. More fuel adjustments but even more so automation would be hugely beneficial for further tests.

        I feel like we’re chasing the tail tip of the dragon through some dark cave, barely brushing a touch, but not yet getting a good grasp, let alone taming the whole beast. We’re getting there though, with these brilliant feats of design Alan is cooking up. And one day that trash can calorimeter could add to our arsenal.

  • magicsnd1

    I’ve ended the run at midnight UTC. The total elapsed time is 47:10.
    This long experiment wouldn’t have been possible without the generous help and good company of Skip and Justa Guy, and I thank them for their patience and support.

    The data is all available at A good part of the analysis has already been done by GED and ECCO, who I also thank for the time and skill they brought to the party.

    Best regards to all,
    Alan Goldwater

    • Ged

      Thank you so much guys for this fantastic marathon, once again!

    • Bob Greenyer

      Well done Alan and Skip for excellent execution of not one, but two experiments. Thankyou also to those that dropped in on Alan to offer support.

      Much learning.

      Thanks for taking the relay Justa Guy.

      Thankyou GED and ECCO also.

      And thanks to everyone that took part in the discussion.

  • Sanjeev

    Yes, Bob said he is arranging Li. But not sure when it will be done.

  • magicsnd1

    Now doing temperature cycles to ~500 C (~650 core), 30 minute hold, then back up to ~1050.

    @Sanjeev: I think what you’re seeing is the thermal time delay as the system heats back up. Because of the extra mass of the alumina filler rod, the null side will take longer to rise in temperature. The settling time for this looks to be about 7 minutes before it settles.During that period Tdiff will be higher than the steady-state value.

    • Sanjeev

      I see. I was assuming that since the mass/thickness of alumina between both the TCs and the heater are same, they should respond in same way. But it looks like the mass below the heater also counts and has a large effect.

  • builditnow

    I personally think it’s a good system to run lots of tests, looking for excess heat greater than about 100 watts. With a few modifications it could be made more sensitive (with the addition of insulation and active cooling for instance).

    I discussed with Alan a future thinking, “kicking the tires” idea of automating his system to enable experiments to be carried out production style. The reactors themselves are relatively inexpensive from a materials point of view.

    The estimate is that Alan could do this automation himself in 6 months, with another suitable person, 4 months, 4 people 3 months, (20 people 2 years LOL ). There would be opportunity for others to contribute time such as monitoring experiments, helping run debug tests, buying supplies, collecting donations etc. etc.

    Since Alan is in the Silicon Valley area, there are numerous people here with all the skills needed, it’s really a matter of are the right people interested and willing to put in the time. Of course, people here also earn nice big salaries, so, they are giving quite a gift to donate their time.

    There is potential for a financial benefit, a quality lower priced product for monitoring and control that could then be sold to all sorts of garage startups who likely have a need for a similar as against the currently expensive “professional systems”. An add on board for a Raspberry pi for instance. Such a board does require very skilled design to eliminate the noise from the power control systems from the measurements. Alan does have related experience from designing audio studio equipment.

    If you are interested or know of potential people, you can comment below for the moment as a way of “measuring the enthusiasm temperature” we have at the moment.
    I have now run some errands, so I’ll check back later this evening for comments.

    • SG

      While not physically in that region, I certainly support the effort and feel such an effort would help push our general knowledge forward, and indirectly push this technology into the marketplace.

    • ecatworld

      Builditnow, I put this comment in its own new thread on the main page, to give it better visibility.

  • Sanjeev

    dT behavior during last kick. It suggests that the actual dT is 90C at 1000C active side. The cross talk reduces it as time passes.

    • Ged

      Also, pressure is going up and staying steady again too, oddly.

      • Sanjeev

        May be it will peak and then fall again.

      • builditnow

        No expert here, the temperature may have gone low enough for some of the hydrogen to recombine with the lithium or something else, then released again as the temperature went back up.

        • builditnow

          Perhaps a lower temperature hold might lower pressure so that the rise could have a small “pumping” effect to drive more hydrogen into the nickel.

    • Sanjeev

      (You may need to refresh the page to see the attached chart.)

  • Bob Greenyer

    Well – we are still way above atmospheric.

    – there are cycle kicking as suggested above.

    From the new patent information there is cycling around the formation of LiH – which because we have not got extra Li in there is below 900 and above 1000 internal – so a cycle could be to take it to 900, then kick to where you are now or something of your choosing.