Hi People,

I do agree with HartiBerlin.

Appart from a natural laziness, it is the extreme 'dangerousity' of these CFL lamps
(containing mercury) that makes me hesitating to build this devices.

What about mere "old fashioned" fluorescent lamps (neon lights)?
Would it work with this Circuit?

Very Best,
Jean

Ok, done with the experiments.  I am not going to make a table.

I don't know why I ruled this out before, but winding the heavily insulated primary turns right on the secondary is the way to go and even easier.  Get the biggest insulated gauge stranded wire you can find and wind 10:1 ratio on the secondary evenly spaced.  My secondary is 20 awg and 368 turns.  I put 36 turns of 12 awg on in the opposite winding direction and I get 1.8 amps for 4 LED bulbs.  That's 5.4 watts each.  With 6 bulbs I get 24 watts total or about 4 watts each into the 7.5 watt bulbs.

I tried tesla parallel - higher amp draw no more brightness
I tried several strands in parallel - higher amp draw no more brightness
I tried out farther from the secondary and total power is reduced.
I tried putting the turns at the middle and at the ends with poor results, falls out of oscillation.

Simplest is best.  So simple, this circuit could be wound on a stick and work nicely.  For fluorescents the primary turns are  halved so put a tap at 18 turns for this big 368 turn coil.  For the smaller original LJL use 14 turns for the 680 30 awg wire, or double for LEDs.

Thanks, Lynxsteam -- and for this informative new vid on how SJR 2.0's are put together:

http://www.youtube.com/watch?v=ROwdzpbISB4&feature=uploademail

Here's a Video titled "Joule Ringers that Don't Ring" showing the E-core and the Aircore side by side.  It demonstrates the flexibility of both circuits, the power draw, and bandwidth capability of both.  Each has its positive aspects.  For pure efficiency the E-Core is amazing with little downside.  The Aircore adjusts to load better, acts more like a household AC supply circuit and starts a little nicer.  Now for Lumens/watt testing which JouleSeeker is undertaking and will report back when done.
 
My guess from watching both is that the E-Core is twice as efficient at providing lumens per watt.  But both devices are very efficient and would easily utilize a medium sized solar panel to provide lighting for several of the rooms in your house - for Free!

I also show the big "Kong" torroid making lots of noise and putting out too much power in a very narrow bandwidth.  This Torroid would probably benefit from the original Joule Ringer circuit.  I will not pursue that however.  If anyone wants it let me know.

http://www.youtube.com/watch?v=-KzugLhMB9A

See that picture above?  Now imagine the primary is not on that PVC tube and instead wound directly over the 20 awg magnet wire.  I wound the primary completely over the secondary.  The turns are 368 secondary, 107 primary of 14 awg electrical wire.
I sanded off a stripe of insulation on the primary so I could tap at each turn.

With one 7.5 watt LED and 12 volts I found a spot at 68 turns where amps dropped down to 280 ma.  That spot is a little over half way across he secondary.  Either side of that turn ma go up in a sort of linear way.  Turns 66-70 ma are right around 300 for one bulb.

1 bulb, 68 turns, 280 ma, 3.36 watt input, 99.4 Khz - brightness maybe 50% - useable

6 bulbs, 68 turns,  848 ma, 10.2 watt input, 47.7 Khz - brightness maybe 30-40% - not very useable

14 CFL unmodded, brightness 70% , 107 turns 1.02 A 29.2 Khz  I suspect that amp draw could be less with more turns but I ran out of turns. 

The other interesting thing is the circuit with no battery hooked up puts out .225 volts, about a 1/4 of a volt with base disconnected.  Base connected about .100 V, 1/10 of a volt