Joule Thief 101

[sm0ky2]

I had thought about this as well but dismissed it as I "assumed" the resistance adjusting circuit would waste more energy than it was worth.  Maybe I was wrong?


Bill

if the inductance and impedance factors are "just right" within the oscillator,
it does not need to be adjusted very often. (depending on the load).
superbrights seem to give the best light -to- energy ratios, which lead to longer run times.

it seems as though the older LED technology uses a lot more energy to produce the same amount of "light".
if it takes say 6 older LEDs to equal one superbright of the same lumens. then you compare the current draw through all 6 in series, or parallel.
compare to the current draw through the superbright, we get more "light" for our $

I speak of light , lightly.,.. as I am currently experimenting along theories that state there is no quanity of a photon.
since a single photon can diverge, each photon can be represented as an infinite number of photons?
as well as that an electron can emit an infinite number of photons during its' existence.

[MileHigh]

And yet the diamond needle dosnt destroy the soft vinyl record like science says it should.
How is it that the hardest substance on earth dosnt just cut straight through those smale plastic bumps in the vinyl tracks?. Over and over you can play your record's without damage from a material that cuts the hardest of steels.--?one for the books MH.

So many things that good old science has the answers for,and yet never produced. 9 times out of 10,it comes from people that are no ruled by book's. For example-the plane. We all know that it is scientifically possible ,but the science guru's never came up with the plane--powered flight. It cam from a couple of brothers that owned a bicycle shop,that said--yes we can. They were the guys that did the job through experiments,and trial and error--not from book's.
In fact,most great discoveries came from trial and error,and the science to explain it came after.
The good old bumble bee was flying long before science worked out how it could do it--this was after some decided that it was an aerodynamic impossibility-now they know better.

There are many thing's left that science and books are yet to explain,and many of those discoveries will be made by those not ruled by book's,or some one elses science and law's.
Funny thing about scientific laws MH,is that they really are not laws at all--there a !best guess! based around only what we have observed so far. These are man's law's im talking about,not those set by nature it self.

If we are to stick with the book's MH,then there is no stone left unturned. But we here,and on other forums,choose to turn over the stones that are still face down. In the next ten years,i feel that you will see first hand how obsolete your books are,and the new discoveries will once again,be made by those that follow no such rules.

Brad

These are just a bunch of tired old cliches that aren't true Brad.  The needle and the record player?  A materials scientist or a mechanical engineer could explain that to you in intimate detail.  The most basic fact is that the needle is not sharp and doesn't cut into the vinyl.  Clearly the vinyl is strong enough to sustain the typically one gram force of the needle that rides on two tiny spots of vinyl as the groove moves past.  You can look it all up in books.  The Wright brothers were science gurus and used the scientific method to successfully engineer their airplane.  They did serious research and experiments, you are grossly underestimating how they achieved their goal.  The bumblebee is just another cliche, and I think that somebody even ran the computational fluid dynamics on a supercomputer to "prove" that the bumblebee could fly.  The algorithms and the number crunching simply weren't around to do that decades ago.

Most of the laws really are laws, and they are Nature's laws.  The laws that govern how circuits work are basically the same laws of physics that people are familiar with in the physical world.  I am just talking basic stuff, not relativistic stuff.

There are also new books being written all the time about new stuff.  That's why you can go out and buy an 8-terrabyte hard drive and turn your vast collection of 6000 Blu-Rays and DVDs into files and put them on your hard drive.

Tons of new stuff will be discovered in the future, and new laws may get written and even old laws may be overthrown or tweaked.  However, when I spin up a flywheel the laws that govern/describe its operation are not ever going to change.  Same thing applies to a coil.

MileHigh

[MileHigh]

Smoky1:

in one side, you can preserve the AC wave properties of the inductor.

on the other, you can simply switch digitally, and the whole circuit acts similar to a simple boost converter.

Look at the basic Joule Thief schematic again.  If it's not switching then you are suggesting that a certain value of base resistance will keep the transistor partially conducting in some kind of meta-stable oscillation.  It would be like when you lean back in your chair and you find yourself at that balance point where you don't know if you are going to fall backwards or not.  Your arms and legs start to wiggle around as you struggle in your own uncomfortable state of meta-stability hoping that you fall forward and recover equilibrium and all goes back to normal.

Look at the L1 coil.  What's going to happen?  The L1 coil is going to be "yanked down" (to ground) by a meta-stable twitching transistor with what looks like a twitching resistance to ground.  L1 is going to see what effectively looks like a meta-stable AC "yank down" superimposed on some DC "yank down."

Big deal, the L1 has one side tied to +1.5 volts and the other side of L1 is connected to a twitching transistor.  So it will respond to that stimulation like any coil will respond to some kind of signal applied to it.  It's not quite a conventional AC impedance response, it's more of a response to a meta-stable wobbling DC with another meta-stable wobbling AC superimposed on it.   Or more accurately you can say there is a DC voltage source on one side of the coil and a funky skittish meta-stable variable resistance to ground on the other side of the coil.  SO WHAT?  A crazy skittish meta-stable resistor is trying to induce current to flow through the L1 coil.  Are you expecting the parting of the Red Sea or something?

In your scenario the L2 signal that drives the base resistor is just the inverse of the skittish meta-stable voltage seen at the bottom of the L1 coil.  That is effectively the "French tickler feedback" that keeps the transistor in some kind of spastic meta-stability.

The whole "meta-stable transistor in a quivering Joule Thief like some spastic guy permanently balanced on a backward leaning chair" really means nothing.  Basic circuit analysis can explain it all.  There is no mystery to uncover, no mystique, no "hidden knowledge that 'they don't want to teach you.'"

It's just a spastic Joule Thief tweaked into a meta-stable elliptic seizure because it is being tickled that does nothing special at all.  Will the meta-stability get stable and settle down into a "resonance" oscillation frequency.  It might do that you never know.  But SO WHAT?  What's so special about a transistor that is varying at a stable oscillation frequency because of feedback to the base such that it that it effectively energizes an inductor with a regular "yank down" and when the effective resistance of the transistor "oscillates high" some of the energy in the L1 coil gets dumped into the LED to light it up?

Some current will get induced into the L1 coil, and when the spastic transistor is at a temporary high resistance, some of the current from the coil will flow through the LED and light it up.  Then the spastic transistor will start to conduct again, the LED will shut off, and more current will be induced into the coil.  The process will repeat itself over and over which will effectively pulse the LED ON and OFF at a high frequency.  Is there something profound happening here?  The answer is no.

MileHigh

[MileHigh]

But let us "test" a JT circuit, with no diode.

I will just repeat what I said before:  With no diode the L1 coil will simply slam a short pulse of current through the collector-emitter junction of the transistor.  There will be a very short high-voltage pulse of current.  L1 will force current through the presumably switched-off transistor.

Now, this will presumably mess up the normal timing of the Joule Thief where there is an orderly feedback signal back to L2 during the normal discharge of L1 through the LED.  Will it still undergo a normal switching cycle when you remove the LED?  I think the switching will stop completely therefore what I said in the first paragraph is probably more academic than real.

[MileHigh]

Smoky1:

Now I am going to play devil's advocate:

It's just a spastic Joule Thief tweaked into a meta-stable elliptic seizure because it is being tickled that does nothing special at all.  Will the meta-stability get stable and settle down into a "resonance" oscillation frequency.  It might do that you never know.  But SO WHAT?  What's so special about a transistor that is varying at a stable oscillation frequency because of feedback to the base such that it that it effectively energizes an inductor with a regular "yank down" and when the effective resistance of the transistor "oscillates high" some of the energy in the L1 coil gets dumped into the LED to light it up?

So let's assume for the sake of argument that you get better run times for the LED in this stable resonance/oscillation mode.  It's a definite possibility, you never know.

So let's go back to something I said:  The transistor briefly "oscillates down in resistance" and conducts and when that happens the LED goes off and the coil starts to energize.  Then the transistor briefly "oscillates up in resistance" and the coil dumps some energy into the LED to light it up.

So you energize the coil in "small sips" and then the coil dumps those small sips of energy into the LED through "resonance."

When you strip that down to the bare bones, it's just like a DC-to-DC converter that pulses a coil for a very short time at a fairly high frequency and then dumps that energy into an output capacitor.  I think a typical pulsing frequency is around 60 kHz and they only pulse the coil for a fraction of a time constant for the maximum efficiency (reducing i-squared-R losses.)

So just like you can buy a small very high efficiency DC-to-DC converter that switches at 60 kHz, you can buy a small very high efficiency DC-to-current converter that switches at a high frequency and the current output can be set to drive an LED.

So perhaps behind all the smoke and mirrors about a "resonant Joule Thief" the basic operating principle is essentially very similar to how a DC-to-DC converter operates.  The fundamental principle is better efficiency through very small sips of energy that are then sent to the LED.  This reduces resistive losses in the main L1 coil.

MileHigh