Joule Thief

[b_rads]

I'm just mentioning a few general points here, and not wishing to put anyone off experimenting of course. The JT circuit and several other variants of the principles of power invertion via inductance will always show similar results.
Depending on what load we require to work i.e. voltage and current, will itself decide how much current draw from source will be reuired to make it work. To light any type of fillament lamp brightly will take considerable more current draw compared with an LED.  LED's are relatively high impedance semiconductor  devices capable of giving  higher brightness levels with low current and voltage needs. That's why a JT circuit and similar configurations that can provide high output voltages... light several LED's in a series combination and still keep the current through the string at a very low level. Filament bulbs and other devices have far different characteristics thus demand a higher current and likewise greater input power requirements from the source.

I posted a youtube the first of this week of a very simple low power flip-flop.  Running yellow LEDs 2.15V, 36mA took a little over 2mA to run ~ bright white 3.2V, 20mA took about 3/4mA to run.  Both the yellow and white were very bright.  You are correct in that the load can affect the input to the circuit.

@Gadgetmall ~ you must take the brightness of the bulb into account when computing efficiency of a circuit.  @Jouleseeker has shown a lightbox in the LynxLamp thread, it is very simple to build and measure brightness of the bulb.  I have used my lightbox many times on different builds.  His formula for computing the efficiency is not difficult once you gather the numbers.  Would love to see pics, video of your build. 

Thanks,
Brad S

[crowclaw]

@b_rads, Yes LED's have improved tremendously, the high brightness types especially. I use some amber LED's left over from a project I was involved with a few years ago for UK traffic signal control. At 20mA you can't view them directly. I have them built into two outdoor wall lights (six in each) they shine through solid 2" perspex diffused tubes and give a pleasant bright glow at night. And yes they are powered by a JT circuit.

[crowclaw]

Just thought I would repost this dual output JT circuit from an earlier posting_ way back.  New members may like to experiment with it. The two coils have an equal number of turns and share the same former. All values are purely experimental as per standard JT circuits R1 & R2 are optional... several LED's can be series connected  across the output of the diodes, or alternatively each string etc can be separately outputted from each diode (note polarity)and subsequently terminated to ground. I found from personal choice BC547 and 2n222 generally work well and most of my ferrite rings were reclaimed from scrap power supplies wound with thin gauge multi stranded hook up wire...have fun.  See later posting for schematic diagram.

[gadgetmall]

I got it !!  I took the other new rod and wound 1080 turns of number 26 then taped it . Then i wrapped 370 turns #20 on top of it and taped it . I can now light an incandescent filament with 1.2 volts and transistor runs cold ., BUT running over 12 volts blows the transistor . due to high voltage . I blew three 2n3055 and one tip3055 , SO now is where i need to put some resistance in this circuit and probably a neon to protect the transistor . It's KILLER !! about 10.5 it lights an unmodified 120 vac cfl  unmodified leds lamps 60 watt light bulbs and in my opinion i should be able to loop it with a cap  .. no time for pics right now maybe tonight when i figure out how to tame this powerful ls circuit ..

Crowclaw it is lasersabers circuit . I am just modified it to run on one battery . see a few post back for his diagram . he has no base resistor just an 8 inch special ferrite fat rod and one layer then another layer for the primary . It is an awesome circuit one you get the wire length right . right now i think i have too much but i like it that way . when i saw the 110 volt 60 watt filament glow on a rechargeable aa i am happy . i let it run for 15 mins and there was hardly no drop in volts still 1.2  This is unconventional . puts watts out for milliamps . Oh i also have a strong magnet attached on one end . it make it brighter .

Gadget

[synchro1]

I got it !!  I took the other new rod and wound 1080 turns of number 26 then taped it . Then i wrapped 370 turns #20 on top of it and taped it . I can now light an incandescent filament with 1.2 volts and transistor runs cold ., BUT running over 12 volts blows the transistor . due to high voltage . I blew three 2n3055 and one tip3055 , SO now is where i need to put some resistance in this circuit and probably a neon to protect the transistor . It's KILLER !! about 10.5 it lights an unmodified 120 vac cfl  unmodified leds lamps 60 watt light bulbs and in my opinion i should be able to loop it with a cap  .. no time for pics right now maybe tonight when i figure out how to tame this powerful ls circuit ..

Crowclaw it is lasersabers circuit . I am just modified it to run on one battery . see a few post back for his diagram . he has no base resistor just an 8 inch special ferrite fat rod and one layer then another layer for the primary . It is an awesome circuit one you get the wire length right . right now i think i have too much but i like it that way . when i saw the 110 volt 60 watt filament glow on a rechargeable aa i am happy . i let it run for 15 mins and there was hardly no drop in volts still 1.2  This is unconventional . puts watts out for milliamps . Oh i also have a strong magnet attached on one end . it make it brighter .

Gadget

Awesome! Really excited about your accomplishment. I'm convinced it's COP>1 with Zaev demag power. Accounts for the cold transistor. The magnet is the cincher. Good luck on the the loop back to source.