PhysicsProf Steven E. Jones circuit shows 8x overunity ?

[profitis]

Its a throbbing hot-cold cycle in this case @tinselkoala you should not be able to detect a net coldspot if the probe is shoved in the middle of the thing.you should be able to detect heat going OUT(friction,battery losses) not heat going IN. I think nul is just pointing something out as opposed to attempting proof of anything.I suggest an oil bath for the entire core here,just the 2 wires emerging.current musnt be too much otherwise frictional heating will overtake ambient inlet

[nul-points]

@ profitis:  thanks for the suggestions re. alternative core material(s) - not quite so easy for me on this setup as the toroid is providing transformer action in pumping charge back to source (in addition to  reclaiming coil-field collapse energy and cross-feeding that from each circuit to the other's battery.

i may have a sintered-iron toroidal core somewhere, on which i could wind a transformer - i'll try & locate that

a nichrome air-cored toroid coil *may* be feasible if i drop the forward-pulse aspect of the circuit - i'm reluctant to vary the circuit config. too much at this stage, but i may try this later


TK>>> "The last time I saw fluctuations like that in temperature data they were eventually tracked down to two causes: the building's AC cycling on and off, and cleaning staff opening and closing the door into the laboratory"


lol - this is lil' ol' england - no AC here!

maybe you missed the section where i referred to the ambient temperature profile?

(and also how i normalised the probe data against the ambient readings)


profitis>>>"Its a throbbing hot-cold cycle in this case @tinselkoala you should not be able to detect a net coldspot if the probe is shoved in the middle of the thing.you should be able to detect heat going OUT(friction,battery losses) not heat going IN. I think nul is just pointing something out as opposed to attempting proof of anything.I suggest an oil bath for the entire core here,just the 2 wires emerging.current musnt be too much otherwise frictional heating will overtake ambient inlet"

spot-on, profitis - in any previous temperature logging of my coils i've always measured a rise in temperature - the interesting results here, for me, are the temperature drop (small, but measurable) and the cyclic element of the data - perhaps particularly the frequency modulation seen with change in probe location

i'm also thinking now about possibilities for isolating/containing any interesting effect

thanks all
np

[nul-points]

@ profitis:

as a quick comparison with another inductor type, i made a new transformer with a couple of windings on a tubular ferrite core (approx 25mm - or 1" - long, by approx 12mm - or 0.5" - ext diam.,  inner hole diam approx 6mm)

i replaced the toroid transformer used for the previous data readings with this new tubular cored transformer (with solenoidal windings) and repeated the test

this latest data represents temperature readings away from the core, then inside the core tube, then finally away from the core again - each period being approximately 15 minutes, samples logged at 5 second intervals

i've filtered the probe temperature readings with a rolling 7-point window (as opposed to a 5-point filter in the previous data) to try and focus on the average/trend of the temperature more

the data from this test, using a tubular ferrite core with solenoidal windings, shows a more consistent temperature inside and outside the ferrite core (after the settling-in slope)

i've thought of another test to try with the original toroidal transformer, so i'll get some data from that & post  when ready

cheers
np

[TinselKoala]

Well... now your measurements are better and the temperature fluctuation range is about 0.2 degree C. It's really too bad your effect size isn't larger, seriously. And with your sampling system it's hard for me to see how to combine data from multiple runs to average out the noise and improve S/N. I'm thinking about it though. What might be a bit more convincing would be if you could show a change in overall slope while under power and while not. Presumably the thing should cool down below ambient when running and warm toward ambient when not, with some slight delay or hysteresis in both directions, right? So an hour on, an hour off, or maybe two hour periods off and on for a long day, then careful examination of the average slopes, fit by some curve fitting function, comparing the two conditions looking for a change in slope and also just where it changes. If it changes instantly when the power is applied, that's probably  not good. If there's a delay, then an increasingly negative slope when power turns on, that's better. 
Am I making sense?  The issue now is the "chunkiness" of the data, like what is the vertical resolution when you are down in that tenth-of-a-degree range? Is it fine enough to detect the kinds of slope effects I'm talking about? From the looks of your graphs it might not be. So there may be some way to improve the vertical resolution of the temperature data.

[MarkE]

I wonder what the noise floor is with the sensors shorted.