calling Maxwell's Daemon

[Omnibus]

@nul-points,

Whether or not there's anomalous temperature dependence can be established sooner by placing attached and unattached cell in a thermostat and have the voltages measured when thermostat is set at different temperatures. I would do that first. More tedious is to study the long term behavior of both cells at constant temperature. Of course, combined effect of time and random temperature change may be th e reason for what you observe so long time comparison of the behavior of the two types of battery should be performed, independent of the above studies. Both temperature and time anomaly, if they are real, would be quite interesting. So, like I said, the reality of either effect is to be guaranteed and that comparison study seems to be a way to do it. It may reveal also some unexpected behavior of the particular cell you're using, attached or not.

[nul-points]

Whether or not there's anomalous temperature dependence can be established sooner by placing attached and unattached cell in a thermostat and have the voltages measured when thermostat is set at different temperatures...{{SNIP}}
    More tedious is to study the long term behavior of both cells at constant temperature.
Of course, combined effect of time and random temperature change may be the reason for what you observe so long time comparison of the behavior of the two types of battery should be performed, independent of the above studies.

yes, not quite sure what you're intending to say here...
    the present tests aren't at a constant temperature

as you can see from the graph in the first posts, they're cycling slowly with room temperature (presently between approx 15 degC and 22 degC,), so the voltages are already being measured at different temperatures
  i agree that tests will still need to be carried out at different constant temperatures

the ambient test results which i've described above are the second round of thermal response tests i've run with this system configuration

my very first test, run immediately after cell construction, used a USB driven heater to provide an operating-environment temperature for the cells around 30+degC - the intention of the test being to determine whether the system could self-sustain operation with the input of heat additional to ambient

i could see that the voltage-time trend was steadily increasing - and in fact, it was approaching the limit of my highest resolution datalogging channel (2V +/-0.25%), and i had to keep dropping the excess heat input

so, as a result of the emerging behaviour which i observed in that first round of tests, i decided to re-order the tests so that i could check how the system responded just to ambient temperatures - the object being, of course, to see if the system gave any indication that it could self-sustain operation for an indefinite period of time

this test, although 'tedious', does have the benefit of reducing the options for people to argue whether the system operation is truly self-sustaining

when i move on from the present stage of cyclic ambient heat input, i can setup a thermostat-controlled operation for the heater so that i can observe the steady-state operation of the attached & unattached cells

[Omnibus]

@nul-points,

I'm waiting with interest the results of your studies so that we can continue the discussion in more concrete terms. Good luck.

[nul-points]

I'm waiting with interest the results of your studies so that we can continue the discussion in more concrete terms. Good luck.

thanks - the good news is that the DIY cell driving an LED flasher circuit is still showing a net increase in charge, with the present on-load terminal voltage trend >1.9V and rising

the experiment has been running non-stop for over 550 hours now (including a week when it was re-located from South UK to a log cabin in the 'frozen' North!)

the 'not-so-good' news is that the more recent secondary test of an unattached cell is not, at present, giving a consistent outcome:

the first few readings showed a clear correlation with the ambient temperature cycle in the room, then after i accidentally disturbed the cell the correlation has become very tenuous

so it's not clear at the moment if its a type a) or type b) outcome (as in the list i defined earlier)

however, the voltage-time trend is not rising, in fact it appears to decrease slightly compared to the trend of ambient temperature

i'm going to give this 'sidebar' test a full week to see if the outcome becomes more clear, but i don't intend to spend more time on it than that

i'm attaching a preliminary snapshot of the unattached cell data so you can see what i mean


i mentioned previously that, in any case, a test on an unattached cell would be of limited value since the useful internal behaviour of a cell is so intimitely bound up with the external electrical activity - and that test is already in progress with the DIY cell driving the LED circuit

there will also be things to learn from my recently-started control test of some commercial re-chargeables connected to a similar circuit load as in my DIY cell test - at least that new test appears to be proceeding without any issues

[nul-points]

 
correction - the experiment has been running non-stop for nearly 900 hours now (cells first connected to circuit just after midnight of 8th Dec 2010)