Joule Ringer!

[b_rads]

   OK - got my aircore LJL to light LED's.  Using 30g magnet wire for a length of 8" and 2 conductor 24g solid speaker wire.  On the primary I wrapped 60 turns which covered the entire secondary.  I had to connect the end of the first conductor to the beginning of the second conductor to get my current draw to tolerable limits.  This gives me 120 turns on the primary for a current draw of 250mA and lit up 2 LED's very well. 

  Has anyone tried this?  It seems everyone has been testing with matching bulbs.  I connected a 0.4w LED night light to the circuit and the current draw jumps up way over 1 Amp.  With or without other LED's on the load.  Anyone have any idea why this happens and does this happen on the ferrite core builds?

Brad S

[Lynxsteam]

Great that you got it running!  Nice amp draw too!  I'll have to try the increased turns on the primary.  More turns definitely drops the amp draw on the aircore.

No I haven't seen the problem with mixed LED bulbs on an aircore.  Definitely mixing CFL or tubes and LED bulbs doesn't work.  There may be some circuitry in the Nite Lite that is dropping voltage which would probably send amp draw up. 

Easy fix - don't use the Nite Lite.

[JouleSeeker]

The best way to describe this Joule Ringer is

Ladies and Gentlemen - KONG

KONG -- lol...  Can you tell us about this sucker, and what kind of power draw you found (compared with your smaller cores)?

I have been thinking along similar (toroidal) lines, I guess -- yesterday I used a high-permeability core, wound in bifilar mode 75 turns for the secondary. (I actually wound this some months ago for another purpose).  I added 8 turns bell wire for the primary to give the approx 10:1 ratio.

I then built the simple Lasersaber 2.0 circuit also used by Lynx, with a 9.5W LED lamp.  Would not light up.  I also built the Lidmotor circuit shown in the photo, with added diagnostic circuitry.  The "white" LED lights up showing current to the base, but the LED would not light up...
Hmmmm.... 

[Lynxsteam]

If you are using a Torroid, that is a very different animal.

Try a couple of things:
1.  Switch leads on the primary. + and C switch, nothing else
2.  Try cutting a little insulation at every turn on the primary where you can poke into it with a probe.  10:1 doesn't work so well with Torroids.  You may find you hit resonance with fewer turns.  Back off a turn (increase by one turn) and then dial in with your potentiometer for best amp draw.

"Kong" is pretty cool.  It does ring.  Left alone (single component transistor) it puts out 50 watts and plateaus.  I don't think I would ever want to run 20 LED bulbs so I am tinkering to see if I can calm it down with capacitors, and an adjustable base tie.  I got it to dimly light one LED bulb on 40 ma.  I have to go shopping for the right resistors and pot.  I will probably try Slider's 3 color LED indicator.  Torroid Joule Thiefs have been exhaustively looked at, I am just doing this to try the one transistor and to see how a large Torroid behaves.  It is not something to replicate because it is a $377 torroid transformer new.

[b_rads]

    My Lux Meter arrived yesterday and I now have three different setups to test.  The LJL with multiple taps for CFL’s, the LJL with speaker wire for LED’s, and a third setup which I will describe.  Have not made the box yet to test the bulbs,
The third setup can be seen below.  I found a transformer that I had salvaged from a speaker system some time back.  This was a Harman/kardon 10 year old system that went bad.  The transformer is a 120VAC 60Hz input with 18VAC 3A output.  This setup is very sensitive to get started, but when it does, it performs very well.  I placed a string of 50 LED’s rated at 4.8 watts, a LOA rated at 2 watts, and a FEIT Bathroom and Vanity rated at 1.2 watts on the circuit for a total of 8 watts load.  Running off a 12 Volt battery, the current draw was 240mA resulting in 2.88 watts.  Using my new Lux Meter I tested the LOA on AC and the meter showed 660.  Running off the circuit the meter showed 1450.  Those numbers are approximations since the numbers were bouncing around a bit.  With these readings being totally unexpected, I performed the test a second time and the results confirmed the first test.

Brad S