Merci beaucoup, Nerzh! 
 Like Lynx and you, I've built a bunch of these BO circuits also.  Fascinating.

Happy hunting to both of you!  pls keep reporting here if you would; too many threads are a tad hard to follow IMO.

On va voir s'il y a quelque chose d'interessant ici;  je pense que Oui.

(Again, my today's vid on this fun stuff is here:  http://www.youtube.com/watch?v=v0WSrcWDy3c&feature=youtu.be )

Hey Prof. S. Jones!

Your French is pretty good indeed!
How comes you learnt this silly language?

Very Best,
Jean

      In my previous vid, I showed the Lynxsteam "cranberry" open-core and results of 107 Lumens/W.  Pretty good -- Lynx noted that there was an aluminum cover on the bottom of the coil holding the 2N3055 which could rob output power due to eddy currents in the aluminum.

     So today I removed this Al cover from the coil axis, placing it outside (along-side) the coil where the B field is small.  The result  shows an improvement in efficacy; primary tapped at winding 75, and I placed 2 ferrite rods in the air-core:

12.8 V  0.332A  ->  4.25W    463 Lumens (with 5 bulbs)   463/4.25 = 109 Lumens per Watt!

Congratulations once again, Lynxsteam!

  Talking about eyes getting zapped.  Last night I finally go my Exciter circuit to light up  the several CFL bulbs that I have.  So, I'm very glad to be having my eyes zapped in this case,  as it also means I'm looking at some good useable light intensity. 
  May have to wear sunglasses now...

  For my exciter circuit I placed two L3 coils that are wound on clear plastic tubes, that fit nicely one inside the other. It takes having two L3 coils, with my set up.  I also use a 1/2' toroid, as the "triger coil", a 2n2222 transistor, 1meg resistor, and In4148 diode.  The two L3 coil tubes are almost touching one another, but separated by the plastic tube. That may not be the best distance, for best results, if I separate the two tubes, I do get a slightly better output.
  So, now that that circuit is working, I'll connect it up to the solar cells, and be charging the 12volts, 4.5ah battery that is running the circuit. 
  I'll upload some pictures soon.
 
 

Here is the newest version of the LJL.  The changes were based on what Joule Seeker (Prof Jones) was finding out about lumens/watt, frequency and tuning.  Thank You Joule Seeker for your encouragement, ideas, and contributions to this area of work.  The design has come a long way from the original idea of a tesla coil based design.  Interestingly though a key aspect of making this work so well comes from Nikola Tesla's work.  And you can't beat the simplicity of Laser Saber's idea of a single reverse biased transistor.
I want to get this tested now and finalize the design.  Then I will publish the design for those wanting to build it, and for anyone around the world who finds it useful.  I will also make these for people not wanting to invest in spools of wire.  I am sure people will find better circuits and better ways to make this in the future, but this is pretty nice.

I didn't mention this in the video.  With 4 bulbs and on the high setting, frequency is 18 khz, on low setting 32 khz.

http://youtu.be/z1H9ckDS_pE

Just a quick question, has anybody tried this setup using a pancake style coil?

  Lynxsteam -- I really like your approach of building these super-coils (your latest design) and making them available -- selling them!  I hope they sell like hot-cakes.  I'm very glad to do testing (and see response below to your questions).

We need some way to get these inventions out the door, and you're taking a marketing approach which I hope works well, very well.  There is nothing wrong IMHO with seeking a fair return on an invention! 

I have encouraged Slider to take his small device (small, one-led and different from yours) into production mode also -- based on your bold step here.  At least I think it is bold and great!

We have on these boards discussed the dangers associated with seeking a patent; but you're not doing that.

Joule Seeker,

Can you put in simple terms what the fantastic results you are seeing means?  For instance you are seeing more lumens per watt than the bulbs are specified to put out.  And we are inputting fewer watts than the bulbs would normally require.  What should someone expect from this type of DC converter?  Dim light and super efficiency, or good light at the same power draw as House grid?
As you report higher and higher lumens per watt it sounds really good, but we need some context or perspective.
For instance 109lumens per watt sounds good, but the bulbs are running with about half the watts they require.  Does that mean we need twice as many bulbs for useful light in a house?

Good questions; I wish I knew all the answers.   Yesterday with your "cranberry" air-core design (admittedly with added ferrite rods) I got 612 Lumens with 5.6 W input (5 bulbs, running at 12.8 V) -- which I think is amazing really.  The bulbs are designed for 1.5 W (each); this is running a little below that, 1.1 W each.

A 40-watt incandescent puts out about 500 Lumens, roughly 13 Lm/W, so 109 Lm/W is much better -- and 612 Lumens is quite bright.

Yes, we seem to be above the rated Lumens per watt typical for LED bulbs.  Two caveats:
1.  this particular bulb is from China and my experience is that they aren't careful in giving accurate lumens values (I've tested a dozen or so Chinese LED bulbs in my light box)
2.  Using Lumens out to estimate POWER-out is a start, but not a compelling result yet IMO.  For instance, it may be that at higher frequency and voltage, these bulbs exceed their rated Lm/W but without any ou-behavior involved.  I suspect that is the case.
So I would like to take the "next step" and actually measure the WATTS out from this device....  I'm working on it and would appreciate any ideas.
 
I would like to dump the AC power-out into a resistance-load, warm water, and measure the temp-rise.  Simple calorimeter approach.  If this loading does not itself change the circuit's efficiency!

As Mags says, it may be that the LED bulb first rectifies the AC input -- this is what I found when I disassembled a CFL bulb; but I haven't disassembled an LED bulb yet.   DC out makes it easier to measure Power out (IMO) -- so this is another approach to measuring P-out.

Thoughts?

Nice work guys!   I've just got a little info to add as I'm not sure how you are measuring Lumens.   I spent years on a forum that is very concerned with accurate measurement of lumen output.   What is not commonly understood is that you need more than just a light meter to get accurate measurement.   The only way to get accurate measurements is with an integrating sphere where the light is inside a sphere.  Measurements done any other way can easily be off by 50% or more.  I don't want to dampen any enthusiasm.  Seeing Lynxstreams vid I was impressed with the amount of light and I'm sure you are onto something good here.   Just be aware of the Lumen measurement situation though.  Integrating sphere's are very expensive but Google search DIY integrating sphere's and you'll find some ways to make your own for reasonable if it is really important to have those Lumens exact.