MFMP to Start New Glowstick Test, May 28th — Live Thread

There’s a post on the Martin Fleischmann Memorial Project’s Facebook page announcing a new live test which is to start later today.

Alan Goldwater of the MFMP gives this schedule:

The current plan is to start the test at 4 PM local (23:00 UTC), to accommodate a wider range of time zones for the second half of the run. The early parts of the ramp should show the pressure peak from hydride decomposition at around 02:00 UTC tomorrow, followed by absorption of hydrogen. We expect the interesting stuff to start at around 850 C, at 10:00 UTC. I’ll post the link for the Youtube stream when we start the run

Calibration runs have already been completed; here are a couple of charts which will be useful for reference:


We’ll be following the action here on ECW on this thread. More information when it comes available.

Here’s the live thread:

Part 1

Part 2

Part 2 (cont.)

Part 3

Part 3 (cont.)

Part 4

Part 5

Part 6

Part 7

Thanks to Sanjeev for starting a spreadsheet to track data from this test. Here’s the link, and it’s editable by anyone:

  • magicsnd1

    The GS3 calibration files from 27May2015 have been added to the Google Drive archive, at:

  • magicsnd1

    The data files are now available in a public archive at:

    The large “raw” files are in five segments of 12 hours (x4) and 4 hours. All are .csv format and can be copied and downloaded. The geiger counter data will be uploaded later today. The gamma spectrometer files need processing to remove background level, and will be added tomorrow.


    • Bob Greenyer

      Here is a very Rough, partially labelled 15minute average based chart.

  • Mike Henderson

    I have set protection (locked) two of the tabs in the Google Sheets workbook that was used to manually collect and display data in Alan & Skips test this weekend. A small group of users can unlock it including myself and Frank Acland. Let us know if you need it unlocked. It is viewable and I have save my own copy for backup purposes.

  • Ged

    Some curve fits would definitely suss that out nicely. The full data should give us a better view of that strange happening too. I haven’t looked too closely at that point cause of all the bumping going on and chaos, but there could be a lot of good data hidden in it as you suggest.

  • Gerard McEk

    When looking to the glow (part 7 1.01-1.03 hrs) while cooling down, it seems to me that the empty reactor glows longer than the fuelled. Can it be that because right hand (’empty’) reactor contains an Aluminia rod and therefore has more mass?
    Further: When the voltage (brown line) is turned down after the highest temperature, there seems to be a delay in temperature response, After that the temperature showed no delay. HAD?

    • Obvious

      Apparently the left side was the empty one.

    • Omega Z

      Yep, That is one of a couple issues.
      This should have been done with 2 separate, but Identical reactors.
      Both should have contained Nickel. The active device having the Lithium hydrate being the only difference.

  • timycelyn

    “…………. A 10% delta T is still within the bounds of measurement errors.”

    Well, yes and no. In the abstract, this statement is a quite reasonable one in the right situation.

    However, I really struggle with this here. I’ve commented above, but to summarise i’d say: bloody strange error! Somethinng that smoothly cuts in unde a steady state condition to smoothly change an output parameter by 10% over a failrly long period of time…..

    Not saying it isn’t an error – time will tell – just that I’ve not seen an explanation yet that fits the facts yet.

  • Gerard McEk

    James, I believe it is up to MFMP to decide whether or not excess heat was generated. I must admit that it looks like that, but I think we need more evidence, like the power values, comparison with bur-in tests etc. I think we should leave it to Alan end Skip (after their motor trip) to do that analysis and I look very much forward seeing it.

  • Nigel Appleton

    A really good experiment – thanks guys!

    If nothing else, it has IMO at least validated the reactor design wrt hydrogen leakage and heating coil longevity; which in turn opens the door to testing different power regimes and fuel composition.

    A question – is it possible for different thermocouples of the same K-type to exhibit different response slopes to temperature?

    • Bob Greenyer

      K-types are very well understood and affordable. The whole experiment was in their operating range.

      A post test calibration will be done.

      In addition – a test with both TCs on a joule heater might be useful.

      The next *GlowStick* generation will have a lower fuel area to volume ratio.

      • Stephen

        I wouldn’t want to make it too complex but I wonder if it would help to have a second or third redundant thermocouple on each side. It could give confidence in the values if both show same. If one fails then the other may still remain usable. If spaced they may give thermal flow information. If three on each side voting could be used if one was showing bad results.

        • Bob Greenyer

          this is a good idea.

      • Alain Samoun

        You mean, for the same amount of fuel: A bigger ID tubing and a lower length occupied by the fuel in the tube(cell)?
        Was the fuel filling the active cell in this test?

  • Bob Greenyer

    If real, it could be higher because of the way heat generated from the core will “look” different to heat generated from the heater according to David Roberson.

    The power was shut off deliberately because Alan could not program the steps down in the PID.

  • timycelyn

    One other interesting point – after power off in the final cooling, below 200C, the experiment once again slowly becomes cooler than the control. On the way up this was a more definite offset around -8C, but at 150C on the way down it has restablished itself at around -2C.

    Tends to point away from mechanical movement within the device changing something, which was one of the worries on the way up when things kicked off….

    • timycelyn

      Another comment on ‘the morning after’ looking over the data again is that when the temperatures of control and experiment reversed at 600C on the way up, it wasn’t that fast – it was a process that took around 30 min or a little more, and allowing for the scatter in the data caused by the temperature controller cycling, quite smooth.

      Again, this is not easily explained by coil shifting or thermocouple moving hypothesese. By definition, anything mechical is more likely to be abrupt, wheras anything chemical or atomic is more lilely to have a slower, smooth characteristic; ie any explanation not requiring mechanical movement.

  • Skip

    Thanx for all your positive comments, folks. It’s been fun for us too, even with the odd and long hours.
    Tomorrow Alan and I are going for a motorcycle ride, (our other common interest) to Alice’s Restaurant. When we get back, we’ll upload the data and maybe start to prepare for the bookend calibration run.

    • ecatworld

      Thanks to you and Alan, Skip. You did a great job of running the experiment and keeping us all informed. Very interesting and worthwhile test!

    • Bob Greenyer

      Enjoy the ride – you fully deserve it!

    • Gerard McEk

      Thanks Alan and Skip for your hard work. I have enjoyed the show. I look forward to your analysis of the collected data.

  • Axil Axil

    It seems ill advised to destroy a perfectly working and sucessful reactor. Most of these reactors explode or the heaters breakdown. I would think it is advisable to run this reactor using a test plan that takes advantage of its pristine and undamaged condition. MFMP can alway break it down to look at its innards at some latter juncture. Run more tests with this reactor.

    • timycelyn

      …… and the longer it is run, the more prononced will be the isotopic or elemental shifts on analysis when it is finally dissected….

  • Enrique Ferreyra

    Congratz to the MFMP team, awesome job, awesome attitude, keep it coming!

  • R101

    Alan might want to close that Team Viewer window 😉

    • Bob Greenyer

      er – yes… and re-issue a code

  • Ged

    When we think about HAD, we also have to remember relative magnitudes. For a very loose thought experiment example, let’s say we take from the highest COP range of 1.3. We could at most be getting 30% of the energy heating the reactor as coming from the reaction, and the other 70% pushing the reactor to, say, 800 C is coming from the heater wires.

    So what happens when you cut power? Now all that is heating the cell is that 30% from the reaction, would that 30% power be enough to bring or sustain a reactor at 800 C? Never, of course. That amount would only allow several hundred C lower temps (e.g. say, 240 C), so the reactor will start cooling immediately. and if that cooling also dims and then kills the reaction, then there will be an inflection. What would be seen in the data, then, is the trace starts dropping upon heater switch off, but at a slower rate than it otherwise would (a different exponential decay function, as if the “room temp” end point was higher by the amount the reaction was providing, and this point would be lowering as reaction efficiency decreased, making it a constantly moving target), until the reaction cooled enough it completely died, and then there would be a faster temperature drop rate back to the natural exponential decay curve that is dependent on the room temperature “floor”.

    But we will never see the temps just hang out fine and dandy for awhile; not without some really good insulation or a really strong reaction that can keep the temperature “floor” closer and kick itself into self sustained mode at that “floor”.

    So, I wouldn’t be surprised if at a small COP of just 1.3, a hypothetical reaction did not show much if anything by the way of HAD, unless serious curve modeling is done. I don’t think we should be looking at HAD for testing the null hypothesis here, due to that insensitivity versus other metrics. But, I could be totally mistaken.

    • Ano Nymous

      Brings up the question, in a larger reactor holding a larger amount of fuel, would the sintered mass of nickel have an insulating effect (trapped hydrogen?) sufficient to maintain a reaction at the center of the fuel. Maybe there just isn’t enough fuel volume to support HAD.

    • Bob Greenyer

      You make a very clear and reasoned argument. This clarity of understanding is also what is driving several researchers like Bob Higgins and Brian Albiston to think about surface area to volume ratios – if heat and energetic/ionising (potentially triggering) radiation can leave by the surface of the fuel – if that surface is too large relative to the mass of the fuel – the ability to self sustain or even be active is curtailed.

      This is a very good area to study – it is why our next generation of *GlowStick* will have a larger internal diameter.

      • Stephen

        Good point. I wonder if EM strength is important for SSM and if the EM varies with the different gauge wire used in the tests for given temperatures. It seems 600 degrees may have been the critical temperature for excess heat. I guess these are is all tuning issues which will be overcome in time.

        • Bob Greenyer

          well – 600ºC external is around 1000ºC internal