Questions about joule thief

[Cecemel]

Hi,
I'm building a joule thief (schematic below), but I'm wondering what happens to the output voltage when increasing or decreasing the 1k resister value or winding more or less turns around the coil.

Regards,
Cecemel

 

[chuckey]

As far as I can see, the circuit is lighting the led (which will need more then 1.8V) from a 1.5V battery. If you change the number of turns, the output voltage could rise, but be present for a shorter period, so the LED might actually dim.
Frank

 

[Tunelabguy]

I had never heard of the "Joule Thief", until I found this Wikipedia article. It is apparently an oscillator that generates pulses higher than the battery voltage.

From that description I would guess that there is an optimal value for the resistor. If you lower it too much the transistor will be biased into saturation all the time and the induced voltage in the winding would not be enough to bring it out of saturation. Therefore it will just stop oscillating and present a fixed voltage to the LED that is too low to light it up.

 

[Audioguru]

Cheap solar garden lights use a circuit similar to a Joule Thief to light 3.2V LEDs from a battery that drops to only 0.8V. They also have a light sensor that activates the Joule thief when dark or activates a battery charger when sunny.

 

[Cecemel]

Luckily I haven't soldered it yet, I will before soldering play around with the number of turns while measuring the output voltage. I've also replaced the resistor with a 2k potmeter so I can ajust it from 0-2000 ohm.

Cecemel

 

[goldsmith]

Hi,
I'm building a joule thief (schematic below), but I'm wondering what happens to the output voltage when increasing or decreasing the 1k resister value or winding more or less turns around the coil.

Regards,
Cecemel

Hi Cecemel
That circuit can be simplified as this :

What is happening ? in fact it can be considered as a boost converter somehow !
If you play with value of 1k ohm resistor it will cause changes in base current hence collector current will be increased then , the voltage across the LED will be increased or decreased .


Merry Christmas + Best Wishes
Goldsmith

 

[Cecemel]

Hi Cecemel
That circuit can be simplified as this :
View attachment 112635
What is happening ? in fact it can be considered as a boost converter somehow !
If you play with value of 1k ohm resistor it will cause changes in base current hence collector current will be increased then , the voltage across the LED will be increased or decreased .


Merry Christmas + Best Wishes
Goldsmith

Thanks for replying and posting your own circuit but I have 2 questions, this component (red in the picture below) I think is an inductor (I'm not really good at electronics), so I think I can just take a small core and wrap wire around it, but I don't now how much turns I should do (picture of the core below). Also, I don't now what this, kinda like a switch looking thing is (green). Then, lastly I have one more question, do you think that that "circuit" (the last picture) I've drawn would work?

Thanks, you, and the other people that have answerd so far.
Cecemel





(Imagine this core without wires) (the multimeter probe is so you can get an idea of the size)


1 volt DC in (red wire)
More then 1 volt DC out (green wire)

 

[goldsmith]

Hi Cecemel

posting your own circuit

That's not my own circuit ! that's just the simplified form of what you've sent before ! ;-)

(red in the picture below) I think is an inductor (I'm not really good at electronics)

What is a transformer ? nothing but two or more inductor which being coupled together !
What is used in your circuit is basically a transformer but it's secondary side is used as positive feedback into the base so what will happen ? it will enforce the circuit to work as an oscillator . ( lovely isn't it ? ha ha )

I can just take a small core and wrap wire around it, but I don't now how much turns I should do (picture of the core below)

It is not a good question to ask how many turns should be twisted on around your core . the better question is what kind of core you've used ? did you find it accidentally on your bench or perhaps you have bought a certain core with known part number ? if so tell me your core type then i'll tell you how many turns would be sufficient . but if no , and you have used a random core another question comes : Do you have any LCR meter ? if you have again you can easily calculate number of turns and i'll tell you how . but another case :
perhaps you don't have LCR meter , and you don't know part number of your core then what ? then i can tell you a known part number of ferrite core so you can buy it from the local electronic shops and then i'll tell you how many turns you should twist around it so your circuit will work !

Now make a choice !

Also, I don't now what this, kinda like a switch looking thing is (green)

The transistor which you've used in your original circuit it can behave as a switch somehow but a fast one indeed !

more question, do you think that that "circuit" (the last picture) I've drawn would work?

Sorry i'm not following you ! do you mean making a transformer like that and then giving that a DC Voltage and then get DC voltage at the secondary ?
If so , sorry that's impossible ! because DC voltage has no "d phi/ dt " and then your transformer will be saturated and it wont work .




Best Luck
Goldsmith

 

[Cecemel]

hi Goldsmith,
I think it would be nice if you could give me a part number of a core, but problem is that it has to be really small (i'm making one of those IR pens so it should fit inside a marker, i will send a picture+ diameter of the inside in a few minutes)

Thanks for helping me out,
Cecemel

- - - Updated - - -

Hi,
Here are the pictures:

The marker:


Disassembled:


The inside diameter is a little more then 1cm, so a core with a diameter of 1cm would be good because the magnet wire also increases the diameter a little bit.

Regards,
Cecemel

 

[BradtheRad]

I made a joule thief from a tiny ferrite bead less than a quarter inch long.

The donut hole was barely large enough to wind 4 (6?) times with 30 gauge wire. It worked.

 

[E-design]

Some time ago, I modified a cheap dollar-store flashlight by inserting the front led assembly from a faulty 4W mains led spotlight bulb and making the inverter by using the same small transformer from the blown inverter pcb.

It turned out to be a very useful light that is capable of operating on a single AA battery as low as 1V with only a slight reduced brightness. Now at last, I have a good use of the flat AA batteries lying around the house. It has become my favorite lab flashlight.

My actual lab measured results were also very close to the simulation results.

 

[goldsmith]

I think it would be nice if you could give me a part number of a core

Hi Cecemel

Perhaps FM series ? FM36 is good for the beginners . but of course you can use any other types of this serial (36 is the diagonal size in mm ) so you can choose smaller ones of this type too but i highly suggest to use big cores to yield a good result and then try to make it compact .

but problem is that it has to be really small (i'm making one of those IR pens so it should fit inside a marker, i will send a picture+ diameter of the inside in a few minutes)

Don't worry about it , i can give you HF circuits which are pretty small and can work at higher frequencies so the inductor and transformer can be obtained by line wiring and without any core !

BUT i highly suggest you to make one joul thief first and if it was successful i'll guide you trhough the final circuit and you can do it with the experience which you've earned since you were making that joul thief !

Best Wishes

Goldsmith