Joule Thief 101

[MileHigh]

Brad:

There are those that dwell on these forum's that dont have much to say,but there knowledge far exceeds that of those here that often make a stand on what they believe to be true. Vortex1 is one of those extremely well versed in EE,and it's due to experience/bench time.

It's not a question of me "believing it to be true," I know what I am saying is true.  Rather, you are "believing it to be an RLC circuit."  Now if you were wise, you would actually look at what I stated about the Joule Thief and how it operates.  I did that over several postings and you are seemingly ignoring that and made no attempt to rebut it.  I linked to a clip that describes exactly how a Joule Thief operates with a full five minute description.  You are seemingly ignoring that also and making no attempt to rebut that.

If you are going to simply ignore what I said then it's willful ignorance on your part and you don't advance.  Look at what I said, look at what Smoky2 said, look at what you yourself said, and go online and do some of your own research.  Don't just almost blindly say, "Oh, a Joule Thief is an RLC circuit because there is some stray capacitance between the windings" because that is dead wrong.  It's nothing more than an incorrect "drive by" evaluation of a Joule Thief circuit.

MileHigh

[tinman]

Brad:

It's not a question of me "believing it to be true," I know what I am saying is true.  Rather, you are "believing it to be an RLC circuit."  Now if you were wise, you would actually look at what I stated about the Joule Thief and how it operates.  I did that over several postings and you are seemingly ignoring that and made no attempt to rebut it.  I linked to a clip that describes exactly how a Joule Thief operates with a full five minute description.  You are seemingly ignoring that also and making no attempt to rebut that.

If you are going to simply ignore what I said then it's willful ignorance on your part and you don't advance.  Look at what I said, look at what Smoky2 said, look at what you yourself said, and go online and do some of your own research.  Don't just almost blindly say, "Oh, a Joule Thief is an RLC circuit because there is some stray capacitance between the windings" because that is dead wrong.  It's nothing more than an incorrect "drive by" evaluation of a Joule Thief circuit.

MileHigh

The JT works quit fine without inductive coupling between the two winding's,and the reason it dose that is due to the C value of the transistor. When operating at low voltages as the JT dose,and the frequencies involved,the transistors own capacitance plays a vital roll. We know this capacitance exist,so i am at a loss as to how you can say it dose not . As it dose exist,and is part of the circuit,then the circuit !is! an RLC circuit.


Brad.

[MileHigh]

The JT works quit fine without inductive coupling between the two winding's,and the reason it dose that is due to the C value of the transistor. When operating at low voltages as the JT dose,and the frequencies involved,the transistors own capacitance plays a vital roll. We know this capacitance exist,so i am at a loss as to how you can say it dose not . As it dose exist,and is part of the circuit,then the circuit !is! an RLC circuit.

Brad.

Nope, you aren't going to actually show how a Joule Thief is an RLC circuit and show how it operates as an RLC circuit because you can't.  You can't sketch out the circuit or sketch out timing diagrams to back up what you are claiming.  What you are doing is making up a word salad.

Also, the Joule Thief will not work as a Joule Thief, if it woks at all, without the inductive coupling between the two windings.  Saying it works because of "the C value of the transistor" is just more word salad.

The fundamental timing and operation of a Joule Thief is based on L/R time constants and there is no resonance at play at all - the Joule Thief timing and operation is governed by the interaction between inductance and resistance and not capacitance.

MileHigh

[sm0ky2]

we can just "ignore" this 100k Ohm resistor


pay no attention to the man behind the curtain

Since the floor is mine, I think I will mop it shiny with some Mr. Clean...
-------------------------------------------------------------------------------------------------------------------------

@ MH - I'm glad you learned how to use Google to help you learn things.
But you cannot take at face value the first equation you come across.
For the sake of humbling your argument that resistance does not matter,
We will suspend all forethought of Ohms Law.
And only consider the direct equations that apply specifically to an RLC circuit.

Your mistake here, is that you are considering the equation:
Wo = 1/ [(sqrt)LC] This is taken in Radians (not freq.)
What this represents, in terms of an RLC circuit, is the Natural Frequency.
This is the resonant frequency the circuit will assume without constantly driving the circuit.
When resistance is very low (not the case of the JT) it can be taken as an LC tank circuit.
If you have not noticed by now, a JT will NOT continue to resonate after the power is cut.


The resonant frequency of the RLC circuit when it is powered (driven) resistance, as a factor of Damping
as:  Damping Factor = Attenuation (in Nepers) / Wo (in radians).
This is most easily measured in the Joule Thief circuit as the Q factor.
the Q of the circuit = 1/R [(sqrt)L/C]

When Q is low, the circuit is "damped", and losses are heavy.
When Q is high, the circuit is "underdamped" and can oscillate,
but there are inductive losses on the magnetic side.

When all components of the circuit are operating at a resonance
that is also a resonant node of each components SRF
losses are minimized.



Using Kirchhoff's Voltage Law (Vr + Vl +Vc = V(t)): we can reduce the attenuation equation to a value ~ =
R/2L
(I know I said I would suspend Ohm's law, and Kirchhoff is basically the same idea, but this is necessary here)

Therefore, the 2 part equation, for the JT circuit is represented as
a=R/2L
and
Wo= 1/ [(sqrt)LC]

The proportionality between these two factors represents the Damping Factor.
And this can be taken as : 
Damping Factor = (R/2)[(sqrt)C/L]

Therfore, to determine Resonant Frequency, we are left with a Complex Frequency response (s),
part is the Natural Frequency, and the other part is the attenuation.
when s=jW ; where j is the imaginary part of the derivative -- the circuit assumes a sinusoidal steady state.

(peak) Voltage and current levels of the resonant waveform are defined by the relationship:
V(s)=I(s)(R+ L(s) + 1/C(s))

Admittance (Y) = 1/Impedance (Z) (inversely proportional)
Admittance Y(s) = I(s)/V(s) or s/L[s^2+(R/L)s +1/LC]

Now, looking ONLY at current, we find there is a Peak value of the function I(jW)
where (Wo) is also the natural resonant frequency.  Wo = 1/[(sqrt)LC]
It is important to note here, the peak value for Voltage; V(jW) derives a different frequency.

solving for Impedance with respect to frequency we find that:
Z = jWL + 1/jWC + R
By this analysis, we see that at the natural frequency; Wo=1/[(sqrt)LC]
Electrical Impedance peaks at a maximum.
However, Magnetic Reluctance (through the ferrite) at this frequency is NOT at a minimum.
Thus at Wo = 1/[(sqrt)LC], losses approach a peak. (not the maximum configuration, but quite high)


When the complex frequency is taken to be the resonant frequency of the circuit,
and this frequency is also a resonant node of the SRF of all components, such that s=jw
(making the assumption that the base voltage at this frequency is within the linear mode of the transistor)
we find peak (not max peak) amplitudes in both the current, and voltage within the frequency domain.
This represents a condition of maximum power transfer from the battery to the inductor, in a steady-state sinusoidal wave.

the resonant frequency of the feedback loop:
this is the current path through the resistor and coil presenting a reflection at the B-E junction of the transistor.
is defined as:

Wo = (sqrt)[1/LC - (R/L)^2] - note that the resistance value (R) is different from the resistance through the primary current path.

there is a 3rd current path in some configurations, that includes a factor of the batteries internal resistance,
I will not get into much more detail on that particular,
as it can be represented as a loss constant pertaining to the battery.

-----------------------------------------------------------------------------------------------------------------------------------------


**puts up the wet floor sign**

[Pirate88179]

we can just "ignore" this 100k Ohm resistor


pay no attention to the man behind the curtain

Since the floor is mine, I think I will mop it shiny with some Mr. Clean...
-------------------------------------------------------------------------------------------------------------------------

@ MH - I'm glad you learned how to use Google to help you learn things.
But you cannot take at face value the first equation you come across.
For the sake of humbling your argument that resistance does not matter,
We will suspend all forethought of Ohms Law.
And only consider the direct equations that apply specifically to an RLC circuit.

Your mistake here, is that you are considering the equation:
Wo = 1/ [(sqrt)LC] This is taken in Radians (not freq.)
What this represents, in terms of an RLC circuit, is the Natural Frequency.
This is the resonant frequency the circuit will assume without constantly driving the circuit.
When resistance is very low (not the case of the JT) it can be taken as an LC tank circuit.
If you have not noticed by now, a JT will NOT continue to resonate after the power is cut.


The resonant frequency of the RLC circuit when it is powered (driven) resistance, as a factor of Damping
as:  Damping Factor = Attenuation (in Nepers) / Wo (in radians).
This is most easily measured in the Joule Thief circuit as the Q factor.
the Q of the circuit = 1/R [(sqrt)L/C]

When Q is low, the circuit is "damped", and losses are heavy.
When Q is high, the circuit is "underdamped" and can oscillate,
but there are inductive losses on the magnetic side.

When all components of the circuit are operating at a resonance
that is also a resonant node of each components SRF
losses are minimized.



Using Kirchhoff's Voltage Law (Vr + Vl +Vc = V(t)): we can reduce the attenuation equation to a value ~ =
R/2L
(I know I said I would suspend Ohm's law, and Kirchhoff is basically the same idea, but this is necessary here)

Therefore, the 2 part equation, for the JT circuit is represented as
a=R/2L
and
Wo= 1/ [(sqrt)LC]

The proportionality between these two factors represents the Damping Factor.
And this can be taken as : 
Damping Factor = (R/2)[(sqrt)C/L]

Therfore, to determine Resonant Frequency, we are left with a Complex Frequency response (s),
part is the Natural Frequency, and the other part is the attenuation.
when s=jW ; where j is the imaginary part of the derivative -- the circuit assumes a sinusoidal steady state.

(peak) Voltage and current levels of the resonant waveform are defined by the relationship:
V(s)=I(s)(R+ L(s) + 1/C(s))

Admittance (Y) = 1/Impedance (Z) (inversely proportional)
Admittance Y(s) = I(s)/V(s) or s/L[s^2+(R/L)s +1/LC]

Now, looking ONLY at current, we find there is a Peak value of the function I(jW)
where (Wo) is also the natural resonant frequency.  Wo = 1/[(sqrt)LC]
It is important to note here, the peak value for Voltage; V(jW) derives a different frequency.

solving for Impedance with respect to frequency we find that:
Z = jWL + 1/jWC + R
By this analysis, we see that at the natural frequency; Wo=1/[(sqrt)LC]
Electrical Impedance peaks at a maximum.
However, Magnetic Reluctance (through the ferrite) at this frequency is NOT at a minimum.
Thus at Wo = 1/[(sqrt)LC], losses approach a peak. (not the maximum configuration, but quite high)


When the complex frequency is taken to be the resonant frequency of the circuit,
and this frequency is also a resonant node of the SRF of all components, such that s=jw
(making the assumption that the base voltage at this frequency is within the linear mode of the transistor)
we find peak (not max peak) amplitudes in both the current, and voltage within the frequency domain.
This represents a condition of maximum power transfer from the battery to the inductor, in a steady-state sinusoidal wave.

the resonant frequency of the feedback loop:
this is the current path through the resistor and coil presenting a reflection at the B-E junction of the transistor.
is defined as:

Wo = (sqrt)[1/LC - (R/L)^2] - note that the resistance value (R) is different from the resistance through the primary current path.

there is a 3rd current path in some configurations, that includes a factor of the batteries internal resistance,
I will not get into much more detail on that particular,
as it can be represented as a loss constant pertaining to the battery.

-----------------------------------------------------------------------------------------------------------------------------------------


**puts up the wet floor sign**

Wow, that is a lot more about this circuit than I even knew I did not know.  Thanks.

MH:

I always thought the JT was a tank circuit as I have always tuned mine to either the brightest light, or the lowest mA draw...these were never at the same resistance.  I thought, as Brad and others have said, that a coil has capacitance?  I have several JT's here that I can cut the input power to and the leds will continue to glow for more than a few seconds...not as bright as when the power was on but, certainly bright enough to see clearly so, that tells me the energy had to be "stored" somewhere right?  I had no other caps in the circuits which I am describing.  I always "assumed" that the stored energy was in the inductor and, if a device can store energy than it has capacitance right?

Those that know me know I am no electronics wiz by any means.  I have played and experimented with many variants of these circuits for about 7 years or so now, and I too am convinced that even the most basic JT has capacitance.

Am I wrong here?

Bill