Low Power Consumption Joule Thief

This is a improved version of the "traditional" JT, the idea is to obtain a very low power consumption. I seeked a way to light a 3v LED with as low as possible battery drain.


Addendum:
Direct link to bigger schematic image http://www.instructables.com/file/F8QACNGGAJUTTOO/?size=ORIGINAL

Step 1 Schematic

Here is my schematic and the schematic for a standard JT.

Parts List :
Q1 - BC547C
R1 - 1k 1/4W
SW1 - Switch
C1,C2 - 10nF ceramic capacitor (marked with 103)
D1 - 1N4148
LED - any 3 volt led
a toroid ferite core and wire
see my outhe instructable for details on how to make the winding http://www.instructables.com/id/How-to-make-a-Joule-Thief/step3/Winding-the-toroid-transformer/

Comments

[acmefixer]

I can barely read the numbers in the yellow boxes but everything else is so small that I can't see the parts or numbers. Putting a capacitor across the LED is not the best way to reduce the current. The 1k resistor can be increased and that will reduce the base bias, and the circuit will then draw less battery current.

[neumanngregor]

Click on the image small "i" icon on the upper left and select from left original size, you will get a 3267x2241 big image, all is then big and nice, ready to b printed :) I played with the resistor value too (used a variable resistor for that), but it seemed (yeah i did some tryout in an empyrical way, im not a scientist) to get better brightness and longer battery life until it starts to dim and after some time to switch off.

[acmefixer]

Thank you for the help. I have always downloaded the .PDF, so I've never used that feature. I guess that can't be done with the .PDF file. Usually the .PDF has a good quality graphic, but in this case it's not readable.

You can cut the current in half by using the Supercharged Joule Thief in my blog here.  Just increase the 1.5k resistor to a higher value, like 4.7k or even higher.  The higher the resistor, the less base bias will be, and the less battery current.

[acmefixer]

I've added a switch to allow comparing the output of the conventional JT with my Supercharged Joule Thief.  It shows the efficiency is much higher, giving similar light output for much less power.  Here is the link and check out the other JTs in my new rustybolt.info blog.

[neumanngregor]

Thank you, i gonna get back at my breadboard to experiment a lil' more. :)

[acmefixer]

I'd like to explain about my SJT. A lot of Joule Thiefs use a diode to rectify the output before it gets to the LED. But the LED is a light emitting DIODE, it's already capable of rectifying the output. So it's better to not use the rectifier because it has a voltage drop od 0.6V or more, and that just wastes power. My SJT just rectifies a small amount of the output to supply current to the base to turn the transistor on. Much less power is wasted.

Also, remember that an hour in the library is worth a day in the laboratory.  Read about some other JT experiments.  One of my favorites is Quantsuff's LED web pages.

[larze]

Thanks a lot for the design! I made it and now I'm wondering what I did wrong because the LED doesn't light up. So when your battery gives e.g. 1.5 V, the voltage on the LED side should also be 1.5 V, right? I will also check the coil winding again when I get a new soldering iron... the old one went bad with a big ball of light, a loud pop and a wisp of smoke, huh, and now doesn't heat up anymore...

[neumanngregor]

01. Check all the soldering points that are conducting and if all connections are done like in the schematic.
02. Check the parts (transistor and led).
03. The How - To for winding the is in my outher instructable http://www.instructables.com/id/How-to-make-a-Joule-Thief/step3/Winding-the-toroid-transformer/ .
04. The voltage at the LED shud be equal the voltage rated for him (like 3v 3.2v or 3.4v or 3.6v if it is a white one).
05. The battery voltage shud be over 1 volt for testing.

[larze]

Ok -- I checked the things but it still doesn't work.
1. soldering points should be conducting and connections should be like in the schematics. One change though: I didn't use the switch and the D1 as acmefixer below said it would waste power -- being stingy is one of the points of JT, right? :)
2. Q1 and LED are ok. Q1 is BC547B and LED is for 3.6 V.
3. I wound the inductor again. The first time it was quite messy and with 20 rounds. Now it is with 10 rounds and looks nice. I used 1-core plastic covered wire, I don't know its gauge but it looks thin enough. I was able to wind about 20 rounds on the toroid whose inside diameter is 7 mm.
4. The voltage at LED still shows the same as the battery voltage: 1.5 V. However, another JT I made does the same but it works anyway. This one: http://www.emanator.demon.co.uk/bigclive/joule.htm  Hmm... is it so that the current peaks make the LED light up but don't show in the multimeter??
5. Battery voltage is 1.5 V

Are you sure that the schematics is correct?

[neumanngregor]

The multimeter dont pick the the peaks if it dont have a max value detection or if ur not useing a osciloscope. I used the BC547C. The BC547B has a hFE of 450 and BC547C 800 dk if there is outher differences ... The schematic is correct, i breadboarded it and then after i closed the experiments i did it on perf board. Try adding a small ceramic capacitor in parallel with the resistor connected at the base of q1 to get him a lil' "jump start" ... or replace the q1 with the C version, i dident had my hands on B version and i cant tell you more about it. I have a full bag of BC547C and it worked fine, and with the lowest consumption from all my transistors i had around. try with a 2n2222 if you dont have the C version, it shud work but with big battery drain.

[larze]

Ok: I took off C2 and put it in parallel with R1. Now the LED lights up. Thanks for advice! Why C2? It's late and weekend so the electronics shops are not open... I wonder whether 10 nF would be too much for final use, I'll see with other capacitances.

The other JT circuit had BC550 whose hFE=800, I think I will try out that, too. I also wonder how fine-tuned the component values have to be in order to achieve the best performance of the circuit (or in my case, even some). I saw somewhere that 1 round of coil winding equals to 38 uH but I have no clue whether it really is so. Maybe I'll play around with those as well.

[neumanngregor]

The coil, when i breadboarded around, whas not influencing the battery drain, i replaced with outher JT coils i had even tryed with a small trafsformer (it whas unmarked) and i connected 2 pins from both coils, and the result whud b somthing like a center tapped coil for JT, the funny stuff is that it worked too. The ideea is that the Collector-Emitter Saturation Voltage, Base-Emitter Saturation Voltage, Transition Frequency and hFE shud b near.

http://www.datasheetcatalog.com/datasheets_pdf/B/C/5/4/BC547C.shtml 

[neumanngregor]

In the left upper corner of the schematic there is a icon with the letter "i" (it will link you to the bigger image http://www.instructables.com/file/F8QACNGGAJUTTOO/ then on the left part of the screen you will see "Available Sizes" then click ORIGINAL(3267x2241) <-- thats the actual size of the schematic. :)

[GrotBox]

I'm really interested in your design, but unfortunately, I can't read your circuit diagram; is there anyway that you could post a better resolution version of it please?