Thank you for starting the thread Steve.
Anomalous to myself could be common knowledge to others and indeed is why a thorough shake down and assessment is needed. I began down this Lasersaber SJR Looper/Akula road with the thinking that a 1K resistor over a cap with an LED was going to trump whatever circuit resistance would be 'in the way' from something built. But....
The running of a mains LED night light far better than a single white LED, an LED that jumps in brightness and recovers after flickering to what should be the end, a coil being enhanced by mains proximity and negatively being impacted by a 27MHz R/C controller are just a few of the reasons that my own thinking is changing.
To cut to the chase, my own 'best' with such a circuit is now a 3 minutes run, as shown in the video below:
[youtube]ZtH6k3ctFQk[/youtube]
Circuit details:
Core is from Nebraska Surplus, as kindly supplied by Steve and matches the one Lasersaber showed.
3300uF x 3 electrolytics on the front end (from the mid 90's so are what may be termed vintage)
3.6V ex cellphone li-ion battery as initial power source (measured and it's old too so never gets past 3.8V)
SMD '1P' 2N2222A.
The coil is an Akula wind, as determined by watching the video where Wesley translates the unravelling of a coil.
I'd like to know if i'm in the ball park ? - 160 winds of 24AWG first layer, 130 winds 24AWG second layer. Gapped copper pieces as per the video.
Then, a 3rd winding of 200 turns 34AWG, connected as per Lasersaber's red coil connections.
The rest of the circuit is per Lasersaber's SJR Looper and what would be the blue coil is connected to the bolt at the top...so it's a hybrid.
2 coils to LS's spec didn't run at all unfortunately and hence why the Akula type was tried. They did however run on my version of LS's circuit, which intends the ability to run any type of core that doesn't have a middle bolt through it.
Here's that video:
http://youtu.be/nIcvbRzhdzoHere is a list of the uploaded videos so far of relevance.
Bobbin Building - How I build the winding bobbins without owning a 3D printer -
http://youtu.be/g3fZROqybcYA good view of the present setup -
http://youtu.be/KBYiKgFruGUAmbient energies - the effects of a mains lamp being brought near to the coil and of the negative effects of a 27MHz R/C controller -
http://youtu.be/OGoa21FKrtI2min 40 sec run - this shows the sudden jumps upward in LED intensity that the circuit does all on it's own in the first minute of running. Then, the recovery back to stable light near the end, after it has faded and seemed to flicker out !
http://youtu.be/eZ7NCJgzFYASummaries and thoughts:
Transistors tried include C1815, MPSA14 (Darlington), C2236, but the '1P' SMD version of the 2N2222A that arrived in the mail today work best by a long way. 40 seconds was the record with a MPSA14, now it's up to 3 minutes.
Using this coil, the LED jumps in intensity in increments in the first minute of running.
At the rundown point where it flickers and should go out, it can recover to solid light output for several seconds again ?!?!
The circuit oscillates at 32kHz initially, falling to a plateau of 16kHz, then onward down til stopping. All is quite linear on the downward section, until near the end where it hunts all over the place around 2kHz.
The coil has an audible ring. It makes growling noises when no load is connected.
The circuit will natively run as a Joule Thief type on a 1.5V button cell.
A 9V battery doesn't make a whole heck of difference compared to 3.6V.
With no capacitors in place, the coil will sound just like an old camera flash circuit.
Run times appear unaffected by increases of mains night light capacitance above 4.1uF
The oscillation speed may be one factor, reducing it to about 40Hz would seem a good use of available energy for LED's.