MFMP: Automated experiment with Ni-LiAlH

    • Official Post
    Quote from Can

    I'm not sure I understand the process there. There's a risk the flask could shatter instead, even if it's made of borosilicate glass.


    There is some small risk of implosion, but I have never seen it happen. You use an Erlynnmeyer type thick-walled vacuum flask of course. The idea of the process is to take a flask full of boiling water and steam, then cap it. The steam will have pushed out the air, so when you plunge it into cold water the steam condenses and dramatically lowers the pressure inside it. Because of the drop in pressure the water continues to boil at lower and lower temperatures. I have a suspicion that the metal particles will retain 'additional heat' for a little while which would promote more vigorous decompression events inside the particles and lead to more fracturing.


    Just a hypothesis of course.

    Alan Smith

    Now it's clearer, thanks. I'm still not 100% convinced; for example I imagine this would work best with a very limited amount of water in the flask.


    BobHiggins

    The calibration seems finished and virtually unchanged from last time, even with the additional descending points. Here is the complete data (average of 100 samples preceding the timestamps):


    Time Stamp Heater Power (W) Tube (C-k)
    2017-05-01 21:38:07.183550+00:00 1.72E-05 17.94234
    2017-05-01 22:35:00.442770+00:00 7.66480778689665 100.0839074
    2017-05-01 23:34:59.973660+00:00 12.6449988464058 150.011068
    2017-05-02 00:34:59.600540+00:00 18.1338170519908 200.3928769
    2017-05-02 01:34:59.139421+00:00 23.8964514275782 250.0262766
    2017-05-02 02:34:58.766310+00:00 30.1809109996351 300.0038539
    2017-05-02 03:34:58.289190+00:00 36.64975614045 349.9707578
    2017-05-02 04:34:57.924080+00:00 43.3730508500335 399.9985218
    2017-05-02 05:34:57.454960+00:00 50.3155602927693 449.9978108
    2017-05-02 06:34:57.089850+00:00 57.3918394773507 499.9888697
    2017-05-02 07:34:56.612730+00:00 64.8733022326514 549.9905458
    2017-05-02 08:34:56.247620+00:00 72.4713587107173 600.037996400001
    2017-05-02 09:34:55.770500+00:00 80.6033403371504 649.9752067
    2017-05-02 10:34:55.413390+00:00 88.5913023331697 700.480553399999
    2017-05-02 11:34:54.936270+00:00 97.4853823143232 749.982511299999
    2017-05-02 12:34:54.571150+00:00 106.154273565126 799.972999999997
    2017-05-02 13:34:54.142040+00:00 115.333362565592 849.957331099999
    2017-05-02 14:34:53.688920+00:00 124.606886352482 899.952266499998
    2017-05-02 15:34:53.307810+00:00 135.2208374344 950.159078399998
    2017-05-02 16:34:52.846690+00:00 144.984181703561 999.9508286
    2017-05-02 17:34:52.467570+00:00 155.740805372694 1049.9104677
    2017-05-02 18:34:52.012460+00:00 166.987543199271 1099.97355
    2017-05-02 19:34:51.631340+00:00 178.783228257576 1149.1512001
    2017-05-02 20:34:51.170230+00:00 190.729405168877 1199.996842
    2017-05-02 21:34:50.789110+00:00 166.858772730629 1099.9311942
    2017-05-02 22:34:50.335990+00:00 144.828057012524 1000.1845243
    2017-05-02 23:34:49.962880+00:00 125.136411235033 900.991367100004
    2017-05-03 00:34:49.493760+00:00 106.432607722423 800.716018600004
    2017-05-03 01:34:49.120650+00:00 72.4039594930798 599.998565100004
    2017-05-03 02:34:48.659530+00:00 43.1024276524585 399.869522000004
    2017-05-03 03:34:48.286420+00:00 17.8808791116521 199.405050200003




    The Ni powder is processed, and dried. It is ready to go into the dry glove box, be mixed with the LiAlH4, and be loaded into the reactor tube. The script for the experiment still must be written. I am still hoping to start the experiment later today; if not, tomorrow morning. I will also post more pictures of the Ni being dried.

    I just started experiment, "HClEtch". The Google Drive data folder for this run will be:


    https://drive.google.com/drive…B0NkRqSDhkWDg?usp=sharing


    The experiment will last about 46 hours. I am going to try a program I created that will automatically copy the data files to the Google drive as they are ready. Hopefully, there will be a steady stream of data throughout the night. This run includes the features that we have discussed. It will have heating to 1200°C, drop-backs to 900°C with a short transition back to 1200°C, and at 1200°C there will be a pressure reduction to low pressure to see if I can make Li boiling occur.


    Now that this is started, I am going to go have dinner.

    The units of measure for the pressure are PSIA (pounds per square inch absolute). So 0 PSIA is a perfect vacuum. On the mountain here, atmospheric pressure is about 10.6 PSIA. Remember also that this is the programmed maximum pressure. If the pressure is below this value, the regulator doesn't do anything.


    This experiment is with 1.3g of fuel. Of the original 7g of AH50 Ni that went into the etchant, after drying, 4.76g was recovered.

    Rigel

    Thanks! So far, so good. The automation is turning this run into a lights-out, hands-off experiment (literally - I will be going to bed in an hour or so). The vacuum system maintenance was automated for last time and this time the automatic file mirroring seems to be working. This should keep the stream of data files flowing through the night. Next we need to automate the plot updating. Automation relieves the burden of the experiment. Without it, I would not be keen to run more than one experiment per month. I want to get to at least 4 per month going forward.

    Good luck with the test Bob and thanks to Can for these great inputs. I think we are learning a lot from this robust methodical approach in your testing. This approach is exactly what is needed.


    May I ask if your fuel mixture is densely packed or loosely packed and if you see there is any advantage an ond of these approaches?


    I'm enjoying following these tests, seeing the growth in real reliable data and better understanding and am looking forward the developments in this and future tests.


    Stephen.

    Some loading appears to be occurring:




    As an experiment I'm starting to also upload files into this Mega.nz shared directory:

    https://mega.nz/#F!XpQwQIrL!UWg9jMHh4I__eK33-d5x8Q


    This is much simpler than having to study and rely onto the plot.ly API. I'm looking to automate the process so that eventually new graphs will be automatically uploaded there as new data files from Bob Higgins arrive.

    Really nice work, both Bob and Can. I appreciate not only your efforts but willingness to share the experience.


    I see the pressure is somewhat volatile between 3:00 and 5:30 and then it pretty much flattens out. At about 9:10 another event.


    Are these programmed into the protocol and are being done under control, or is the pressure oscillating due to some reaction?


    If not programmed, what are your thoughts on the cause?


    Again, thanks both!

    Quote from Can

    There is a valve which discharges excess pressure (e.g. from Hydrogen decomposing from the LiAlH4) from the cell when it exceeds a certain predefined threshold.


    Thanks for the clarification and the graphs! Having them all in one pdf is really nice to compare events. Really nice work! :thumbup:


    So loading perhaps has started at about 10:10 as we see a steady fall of pressure.


    Nice to show the pressure oscillation at about 9:10. At the same time, the temp was increased by 100 C. One would expect the temp / pressure correlation to be seen and one does. A good sign that the system is measuring as one would expect.


    Bob,

    Is this a MFMP project or one you are conducting personally?


    Again, thanks and nice work!

    can,

    Good morning and thanks for the great graphs. I see the prototype file mirror application has been doing its job though the night.


    @Bob

    Can is correct. The system is controlled with a USB programmable back pressure regulator that I designed/built. There is an electrically operated solenoid valve that vents to vacuum under programmed control:

    On the outlet of the solenoid valve, between the valve and the vacuum reservoir, there is a 1/8" MNPT nipple that has inside a 10 micron orifice to slow down the flow - it is a flow resistor. Not marked, but on the lower right of the plumbing is the pressure sensor that feeds back the pressure to the DAQ and the back pressure control electronics. The Labview program controls the back pressure by programming an 8-bit DAC for the desired back pressure. For a given setting, the DAC output is static, and an analog feedback loop controls the opening/closing of the solenoid valve (so there is little computer load for control). The system limits the pressure to a maximum of the programmed value, presuming the gas is sourced from inside the reactor.


    I could add, and have considered adding, another valve to control a supply of gas and regulate pressure as well. With such a system, I could even switch gasses between H2 and D2 (for example) under computer control. If I can get the system to show excess heat, I eventually want to have it setup for gas sampling as well with the samples fed into my ultra-high vacuum system for RGA analysis to measure the evolution of the gas composition. MKS even makes a high resolution light gas RGA that can separate the hydrogen and helium isotopes in the analysis.


    Regarding MFMP... I am an MFMP volunteer/contributor. We have common purpose and I support MFMP in some way on nearly a daily basis. However, everything in this experiment is my construction. I have received a few materials from MFMP, and of course, a great deal of shared information! For example, Bob G. got me a sample of Parkhomov's Ni powder. Brian Albiston sent me a sample of the micro-encapsulated Li. Alan Goldwater sent me a sample of FT catalyst. Ryan Hunt (Hunt Utilities Group) loaned me an RGA (not used in this experiment). We all share experiences and materials as we can. All of my designs are open.

    BobHiggins

    It did work an in the end I also managed to cobble together something that can automatically upload somewhere the graphs. However that directory is not guaranteed to exist indefinitely so do keep saving/backing them to your experiment folder (no hurry though).


    Changing subject, can you comment on the rationale behind the current experiment schedule?

    Is this experiment mostly a test to see if only changing the powder type (etched, degassed) yields any change compared to the previous runs?

    can

    I am saving your graphs to the experiment folder so that they don't disappear (I also have all of the files doubly backed up). I could give you permissions to load into the experiment Google Drive folder directly if you think you want to do that.


    Since we are in a search mode in a multi-variable space, I decided there will be too many experiments if only one parameter changes in each experiment - at least until we get some positive feedback (other than "this didn't work.") In this experiment the Ni preparation changed, and also the excitation schedule (protocol) changed. I have mentioned the change to the Ni preparation (and I am going to add some photos today from the preparation of the Ni). The protocol adds stepping to 1200°C, deep drops to 900°C with a sharp return to 1200°C as stimulus; and after a while at 1200°C, there is a pressure drop to 1 PSIA to try to stimulate Li boiling and again stimulate XH. If we see XH, we can go back and separately break apart the changes to see which one is important to follow for the next experiment. I have more of this same fuel remaining, so it would be a quick turn-around to run again with an exact replication or reproduction of only part of the steps to see which needs greater exploration.

    BobHiggins

    I'm asking because for example the GlowStick 5.2 experiment famous for the "signal" had deeper and more frequent temperature cyclings, although the "low" steps were probably not deep enough to ensure significant formation (and subsequent decomposition during the "high" steps) of LiH. There is a graph that shows them well on quantumheat.org:


    http://www.quantumheat.org/ima…Parkhomov_Replication.png

    http://www.quantumheat.org/ind…cookbook-is-in-the-signal


    I've mentioned it in earlier comments, but I don't think that just allowing the Li to evaporate will work as a stimulus. Still worth trying, though.


    Latest graph attached in this comment.

    Also here: https://mega.nz/#F!XpQwQIrL!UWg9jMHh4I__eK33-d5x8Q

    Quote from BobHiggins

    I decided there will be too many experiments if only one parameter changes in each experiment

    Have you considered using a statistical design of experiment setup such as Taguchi?


    I have taught seminars on this and used it in real life R&D situations. The positive is that you can test for x number of variables at one time and statistically

    determine cause/effect. The down side is that the experiment plan must be clearly designed ahead of time and then followed explicitly. It does not allow for

    changes to the protocol "on the fly" .


    You may be familiar with the discipline (or not), but I have found it very useful when having many variables to test and need to know the quantitative effect of each on the variables. However, since the "Result" one is looking for is excess heat, if none is seen in the experimental array, then you will obtain no statistical value. I do see it being VERY useful if one can obtain the excess repeatedly and one is looking at what variables effect various parameters such as so called "COP", start up time, etc. Just a thought anyway.

    @Bob

    I am familiar with DOE and have seen it applied to some benefit when the problem's variables can be identified and the experimental outcomes can be quantified. However, as you say, in this case we are not yet seeing any quantification of result - either XH or radiation. Until we start seeing some quantify-able metric, my understanding is that DOE is not appropriate. It is a good thought, and if we start seeing XH or radiations, I would appreciate some guidance in setting up a DOE. Finding the right assistance is one of the wonderful aspects of community involvement.

Subscribe to our newsletter

It's sent once a month, you can unsubscribe at anytime!

View archive of previous newsletters

* indicates required

Your email address will be used to send you email newsletters only. See our Privacy Policy for more information.

Our Partners

Supporting researchers for over 20 years
Want to Advertise or Sponsor LENR Forum?
CLICK HERE to contact us.