To be deleted...

[itsu]

Interesting guys, indeed a good suggestion Conrad about the supercaps.

I stopped my circuit when the NiMH batteries were around 0.6V / 0.9V, but the led was still on.
It had run for about 30 hours (almost depleted batteries to start with).

I have changed the 2n2222 for a BC547 which cleaned up the oscillation signal and its now running with 2 fully charged 750mAh NIMH's
I had to order the BAT42's as i did not have them, so will redo the test when they arrive.
I will also incorperate the supercap setup as i have 10 10F / 2.7V supercaps, so i will split them in 2x 5 (50F) parallel.

One thing i noticed is that it seems that the led pulled 16mW, while neither battery was supplying that amount of power.
But it is difficult to make accurate measurements with such low voltage/current levels so will need to redo that to.

Another interesting thing would be to check if the resonance frequency has any influence on the workings of the used yoke/toroid.
Maybe at a certain resonance frequency it will co-resonante and lower the losses even further.

 
Thanks,  itsu

[nul-points]

i think that extending operation/increasing efficiency of a circuit using battery energy is a good long-term goal (which is why i initially moved to the 2 battery setup) - however, it's looking like we need to start getting a better measure of the quantities of energy being transferred & converted, to see if & where any gains are being made

i was very encouraged to see the results of the 2 capacitor tests to get some clear indication that the general principle works - and works well

pleased to hear that a BC547 has given you improved switching over the 2N2222, the BAT42s may not make such a visible difference but they should reduce any leakage currents you may have had from the 1N58xx devices

the supercaps will be a good move i think - i've started to experiment with battery i/p & supercap o/p and i think we can still achieve similar efficiency gains with this method (and therefore not need to double up the number of batteries/cells we're using)

i've also been using the circuit at higher voltages (more cells - more LEDs) - a move like this may help you, too, with power measurements

for higher supply voltages i've added a BAT42 between base & common (cathode to base) of Q1, to clamp the drive voltage swing to the base, to avoid any zener breakdown effects - this seems to have improved efficiency at these increased voltages, too

another small change has been to reduce the buffer capacitance and split its location - i reduced the 1500uF electrolytic cap to a 100uF Tantalum and placed it in parallel with D1 & the input supply, and i now only have 0.33uF of non-polar capacitance decoupling the supply to the oscillator on the T1 side of L1

these values may well be suited to my particular L1 & T1 values, so you may need to 'select-on-test', if you wish to try these mods - your comments about resonance are very relevant here, and i suspect that in general the higher frequency of the oscillator will be better

interesting about the LED power measurements - as a *very* approximate rule-of-thumb, your scope traces show the flyback current/voltage spikes as a right-angle triangle so the average E & I values will be (50% of their height x duty cycle), if that helps you double-check the scope power values?

thanks to all you guys reading & experimenting with this circuit, for feedback, ideas & shared results
np

[nul-points]

Responding to your earlier question about comparative illumination levels of the LED between the two tests, Conrad:

I had to make a visual comparison because i haven't yet found the parts i've used  before to compare LED light levels between pulsed and DC operation a few years back - so - visually there doesn't seem to be a marked difference between the tests

Its difficult to give a more useful visual comparison because of course in both cases the supply voltage gradually decreases through the duration of each test anyway - in the case of the swapped devices circuit, the LED becomes brighter again as each swap is made, then decreases as before

Of course, the test conditions are quite artificial, because in normal use as a flashlight the battery supply voltage only decreases very slowly by comparison with the test capacitor


I realised on re-reading your post just now, to answer your illumination question, that i misunderstood your comment about run duration - i originally thought you were referring to the test conditions but i think now that you meant in general - so, yes, i plan to make the variable resistor available as a control - the user can then choose to make the illumination/duration trade-off to suit the needs of the moment

I looked up your electrophorus link to the examples site - very interesting!   I think that there are more connections between 'contact voltage', Galvanic voltage, and electrostatic behaviour than is widely known or accepted in mainstream science.  It seems to me that these are just another way of tapping into the inherent energies which bind matter together, and which we also see becoming accessible through LENR type processes (eg our own Prof Steven Jones, Rossi, and Fleichmann & Pons)

So - cool- electrostatics -  a  very interesting area of study, and not too far separated from the radiant energy collector device of Tesla - kudos!

All the best with your investigations, Conrad
np

[TinselKoala]

Re the issue of judging LED brightness...

I've recently been making some electricity-to-light efficiency tests and it's surprising how much actual brightness change has to happen before the eye can detect it purely visually.

I'm using an Extech LT300 light meter and a box which keeps the light sensor and the light source under test separated by a fixed, standard distance and excludes other light.  Here's the "unboxing" video I made when I first received the Extech meter:
http://www.youtube.com/watch?v=Iulxcqg5USk
(I don't show any efficiency tests in this video, it's just unboxing and testing the meter for operation.)

My experience with this meter and the LED efficiency tests I've been doing have really made me realize that it is absolutely essential to use sensitive instruments, rather than "eyeballing", to measure brightness of LEDs, whether pulsed like with a JT or with steady DC. This is also true of incandescent filaments. The eye is incredibly "non-linear" in response to brightness. Sometimes I've seen 30 percent difference in actual measured brightness when to the eye the brightness levels can look nearly the same.

[conradelektro]

I had to make a visual comparison because i haven't yet found the parts i've used  before to compare LED light levels between pulsed and DC operation a few years back - so - visually there doesn't seem to be a marked difference between the tests

I realised on re-reading your post just now, to answer your illumination question, that i misunderstood your comment about run duration - i originally thought you were referring to the test conditions but i think now that you meant in general - so, yes, i plan to make the variable resistor available as a control - the user can then choose to make the illumination/duration trade-off to suit the needs of the moment

@Nul-Points

My argument concerning LED brightness is the following: if one stores some energy for later use (by swapping the batteries) the LED will shine dimmer. In other words, the total light output is about the same whether on has the LED shine brighter for a shorter time (your test without energy storage) or one has it shine less bright for a longer time (your test with battery swapping).

May be one can just as well forget "battery swapping" by just having the LED shine less brightly (by setting the variable resistor at the base to a higher value or by making the ON-pulses shorter by help of some circuitry at the base of the transistor). I got very good results by carefully switching the transistor with a microprocessor (adjusting pulse frequency and pulse width to some optimum). This only made sense when driving CFLs (compact fluorescent lamps) or many LEDs. This principle is used in modern LED drivers which adjust the Voltage over the LEDs to adjust the light-colour temperature (which changes when the LEDs warm up) and pulse frequency and pulse width are changed  to allow for dimming. So, to optimally drive LEDs one best uses a microprocessor.  http://www.ti.com/general/docs/lit/getliterature.tsp?baseLiteratureNumber=snvy001

Yes, my comment about the variable resistor was meant "generally". The variable resistor at the base of the transistor defines (within limits) the switching, mostly the length of the ON-pulses and as such the amount of energy flowing through the LED (which is the brightness).

Greetings, Conrad