Most Joule Thief circuit schematic specify the use of 1 or 2 coil.

Question 1 - Instead of using a transformer or winding a toroid coil. Is it possible to use ready made inductor, those that look like surface mount resistor.

Question 2 - Is the 1N4148 diode neccessary in the above schematic ?

 

Q1 - Yes, you can use this type of a inductor.
Q2 - No.

 

Let us know whether it worked. I just tried it in LTSPICE with the inductors coupled (as a center-trapped transformer) in both phases, and as un-coupled inductors and it did not work in any case.

 

1st picture, using 2 pieces of 220uh inductor coil. Located next to the LED & next to the brown capacitor. Marking on inductor is 221

2nd picture, the same picture above, lighted up with 1.5V AA battery

3rd picture, using 2 pieces of 100uH surface mount inductor, that look like SMD resistor. Does not light up.

Hence, my question can the flat grey rectangle SMD inductor replace the toroid coil in Joule Thief ?

 

I am using either MMBT2222A or MMBT3904 NPN transistor instead of MPSA06 specified in the schematics.

Thank you, Jony & Dick for taking the time to answer my questions. Looking forward learn from you guys.

 

Thank you for your test results.

3rd picture, using 2 pieces of 100uH surface mount inductor, that look like SMD resistor. Does not light up.
Hence, my question can the flat grey rectangle SMD inductor replace the toroid coil in Joule Thief ?

All other things being equal if the inductors have sufficient inductance they should behave about the same.

 

Some web article mentioned that both inductor must be parallel & close to each other, in order for the magnetic field to interact & do its magic for the Joule Thief to work.

However, the 2 piece of inductor is position in an L shape on my PCB & not coupled together.
Could this be the cause of failure ?

The 220uH inductor used in picture 1 & 2, actually look like a tiny toroid coil or transformer. But the 100uH inductor is just flat & rectangluar.
Why the shape & construction of both inductor are so different ?

 

The youtube video show moving two inductor closer will light the LED.
Moving inductor further apart, LED will switch off.

 

I would think the magnetic fields need to be aligned so if the construction and or mounting of the inductors makes the fields unalligned, then you will get poor results.

 

If coil #221 is constructed like I believe it is......the flux can be shared. But not in the SM type.

You have shown 2 different circuits.

 

You do not need magnetic coupling between the two inductors according to the simulation. I think that the 10 nf capacitor is what gives the necessary additional phase shift to make it oscillate.

In what might turn out to be a similar circuit the role of the capacitor is more obvioius.

 

Hi,

I did a quick analysis with both L's of 100uH and C2=1uf just to see what would happen, and it did show that it could oscillate at around 160kHz. I also used a linear amplifier in place of the transistor just to see if the circuit had the ability to oscillate given some favorable conditions.

One thing to check for, and not just in this circuit, is the reverse voltage on the base of the transistor. If it goes too low it could damage the transistor badly. This leads me to believe that the two diodes in the original circuit may have originally been connected to the base of the transistor to limit the reverse voltage that could appear there for some parts of the oscillation cycle. This could be checked in simulation first.

 

If coil #221 is constructed like I believe it is......the flux can be shared. But not in the SM type.

You have shown 2 different circuits.

The 3 picture are using the same PCB, only the inductor type is different. Schematic specified 100uH, which was shown not to work in picture 3.
221 inductor used in picture 1 & 2, is similar to below picture

100uH inductor used in picture 3, is shown below:

What is the difference between the 2 type of inductor ?

Can I use the flat 100uH inductor in a Joule Thief circuit ?

Please advise. Thank you very much

 

The difference appears to be that one inductor has a much higher inductance than the other. Many of those chip ferrite inductors are made of material that is lossy at high frequencies because they are used as RF chokes, and those in particular would probably be harder to use in an oscillator. I don't know how to tell the difference by looking at them.

A hobbyist once wrote that he had built the version of a Joule Thief that uses a center-tapped inductor on one of those plastic plug-in type breadboards. He said that after he got it working he replaced the inductor with two jumper wires and the LED still lit up! Clearly he had made a radio transmitter that happened to have an LED in the output stage .

From this, yes it might work with 100 nh inductors but it is more likely to work with higher inductances, particularly those what allow the oscillation frequency to be low enough for the transistor being used to spend most of its time on or off than in its analog region.

 

The difference appears to be that one inductor has a much higher inductance than the other. Many of those chip ferrite inductors are made of material that is lossy at high frequencies because they are used as RF chokes, and those in particular would probably be harder to use in an oscillator. I don't know how to tell the difference by looking at them.

A hobbyist once wrote that he had built the version of a Joule Thief that uses a center-tapped inductor on one of those plastic plug-in type breadboards. He said that after he got it working he replaced the inductor with two jumper wires and the LED still lit up! Clearly he had made a radio transmitter that happened to have an LED in the output stage .

From this, yes it might work with 100 nh inductors but it is more likely to work with higher inductances, particularly those what allow the oscillation frequency to be low enough for the transistor being used to spend most of its time on or off than in its analog region.

Ok, thank you, I will try to rebuild the circuit again

 

When you replace the coils with the surface mounts........you changed at least 2, if not 3 properties.

First you changed the magnetic orientation of both inductors. Second you changed the flux density of both inductors. And third.....you might have put one of the surface mounts in backwards.....or anti phase.

If you understand field orientation.......and have enough leakage from the surface mount....with proper board placement....the surface mounts should work.

 

When you replace the coils with the surface mounts........you changed at least 2, if not 3 properties.

Thanks for pointing out, this is new to me. Does this apply to all schematic that use a coil right ?

First you changed the magnetic orientation of both inductors.

The 100uH flat chip inductor on my PCB was at right angle. Some internet article did suggest placing them close together in parallel. Hence, I am suspecting this was the cause.

third.....you might have put one of the surface mounts in backwards.....or anti phase.

Are you refering to the inductor ? The 100uH flat chip inductor got no polarity, look the same from either side or end.

If you understand field orientation.......and have enough leakage from the surface mount....with proper board placement....the surface mounts should work.

Do not understand "field orientation" & "enough leakage from surface mount". Please explain.

Isn't leakage a bad thing ?

Thank you very much.

 

See post #11 concerning magnetic coupling

 

You do not need magnetic coupling between the two inductors according to the simulation. I think that the 10 nf capacitor is what gives the necessary additional phase shift to make it oscillate.

See post #11 concerning magnetic coupling

Thank you for sharing your simulation results & unique conclusion.
If magnetic coupling between 2 inductor is NOT needed, then 3rd photo in post #4, should light up but it did not.

Most internet articles on Joule Thief circuits emphasized the importance of the toroid coil or transformer & the coil to be coupled correctly. My understanding of these articles are that a magnetic flux is generated in the coil & when the magnetic flux collapsed, a kickback energy is released back to boost the battery voltage to a sufficient level to light up the LED.

Don't mean to be rude or disrespectful, but is it possible there are some bugs in the simulation.
Thank you very much.

 

You are absolutely right to suspect the results of computer simulation.

I duplicated your circuit with real parts. It works very well with 1.21 volts in, the only voltage I tried. The transistor is a 2N4401 and the resistor is 4.7k -the closest I could come to 5.6k. The capacitor is polyester film.



In the photo the inductors are 5 cm apart and at right angles (for minimum mutual inductance). Regardless of orientation or how close together the two inductors are there is no visible change in the luminance of the LED.