OverUnity Research
  
Welcome, Guest. Please login or register.
+  OverUnity Research
|-+  Electrical / Electronic Theory and Learning Center
| |-+  Joule Thief Circuits
| | |-+  Beyond the Joule Thief
Pages: [1]
Topic: Beyond the Joule Thief  (Read 24800 times) Print
MileHigh
Group: Guest
« on: 2009-12-16, 07:03:40 »
So what is the ultimate Joule Thief?  How can you really get more efficiency?

The answer is to abandon the whole Joule thief concept and do it with a microcontroller or a pair of CMOS 555 timer chips.

What is a Joule Thief really?

A Joule thief is nothing more than an inductor connected to your voltage source on one side, and an opening and closing switch that connects to ground on the other side.  The end of the inductor that connects to the switch also connects to a diode to collect the energy spikes and pump them into a load or into a capacitor.

When you reduce it to its basic form, a Joule Thief is a pulsing inductor that gets its energy from a voltage source, typically a battery.  The pulsing inductor discharges it's energy into a load, typically an LED.

A timing source controls the switching of the Joule Thief.  You pay a price in power consumption and have limited control over the timing source when you make a standard Joule Thief circuit.  The timing source comes from the trigger (a.k.a. base) coil and that consumes juice.

Therefore the solution is to switch over to a more intelligent timing source and get rid of the Joule Thief circuit altogether.

The more intelligent timing source could be a microcontroller.  All microcontrollers have built-in hardware timer registers that can control the frequency and duty cycle of a square wave on an output pin.  This gives you the ability to have software control over the timing signal generated by the hardware that is built into the microcontroller.   You could write a simple program that reads some of the I/O bits that are configured as inputs.  You could have switches that control frequency up and down and duty-cycle up and down so that you could adjust your frequency and duty cycle of your timing source "live" while the microcontroller runs.  The microcontroller would consume a small fraction of the power that the Joule Thief consumes in overhead to do the timing function.

Another option would be to use two CMOS 555 timers.  One 555 runs at a variable frequency and connects to a second 555.  This gives you the running frequency. The second 555 runs in "one-shot" mode and gives you an adjustable pulse width to turn on the switch.  This setup would consume a small fraction of the power compared to the JT also.

There you have two options for a rock-steady, reliable, and flexible timing source for switching the inductor current on and off.  Both of them would consume almost no power.  You could have a separate 4.5 source for powering the timing source.

Then, it would be up to you to pick the switching transistor and inductor/toroid setup.  You would have the ultimate flexibility here, pick your transistor, pick your toroid, decide how many turns of wire.  There is nothing stopping you now.  You know that you have a reliable and flexible timing source, and you can mix and match any coil configuration you want.  You could probably fire Xenon flash tubes from disposable cameras, neons, as many LEDs as you want, charge any capacitor at any rate that you want, control exactly how much energy you put into the coil before it discharges, the sky is the limit.

For example, if you want to light a CFL, then you could lower the switching frequency to 70 Hz, just above the human eye's ability to perceive flickering.  Then you could chose your coil/toroid, and then play with the "on" pulse width to put the exact amount of energy that you want into the CFL  for every "burn."  Or you could fire the CFL at a much higher frequency and have a continuously sustained plasma inside the tube.  That may have certain advantages.  Like I said, the sky is the limit.

By using a microcontroller or a dual CMOS 555 timer setup, an astable multivibrator triggering a monostable multivibrator, then you have complete control over efficiency and power consumption.  For every load there is an optimal configuration of inductance and switching time to give you the best performance.  If you also factor in cost, then the optimal configuration may change.

One serious option is to go air core.  Why go air core?  Because all toroid cores burn off energy, they are "lossy."  If you use an air core inductor, then there are no energy losses associated with a ferrite core because there is no ferrite core anymore.

What I described above is the next logical step in experimenting with Joule Thieves - move past them and do a completely new design that does away with the constraining Joule Thief "transfomer" and switch over to a computer-controlled or programmable-555-timer-controlled switching function that drives your choice of transistor and coil.

MileHigh
 IP: [ Logged ]
MileHigh
Group: Guest
« Reply #1 on: 2009-12-16, 07:15:55 »
Supposing you do implement an advanced timing setup for your "Beyond the Joule Thief."

Some of the variables that you can play with are the following:

The source voltage for the coil.  Yes you can use a 1.5 volt battery if you want, but there are no limits now.

The inductance of the coil.  Larger inductors can store more energy, giving you a higher energy spike.

The initial current when the transistor switches off.  This is critical to have control over the initial current, for example you might want to tune it to the optimum current for a given LED.

The operating frequency.  Go crazy here and experiment.  For LED lighting applications the optimum frequency would be about 70 Hz.

The pulse width for the transistor ON time.  This is very critical in controlling how much current is flowing through the inductor when it switches off and for controlling how many L/R time constants you want to charge the inductor with.

The amount of energy in the pulse.  Here you can mix and match your source voltage, the pulse width, and the inductance of the coil to dial up any amount of pulse energy you want at whatever initial current you want.   If you are firing neons, then set your pulse energy to be quite low.  If you are going to drive an ignition coil and you want to create some real fireworks, you set your pulse energy to be quite high.

The list goes on and on.

MileHigh
 IP: [ Logged ]
MileHigh
Group: Guest
« Reply #2 on: 2009-12-16, 07:27:25 »
There are a lot of turn-key microcontroller setups out there that are cheap and connect to a USB port.  They look like a tiny PCB with some I/O terminals and the microcontroller chip.  They are really really cool.  You can compile C code or assembler and then transfer it to the mircocontroller board and then you have your own stand-alone intelligent little beasty running your own algorithms.

The desire is to have a microconrtoller that generates a rock-solid timing source for turning on and off your choice of transistor switch that is connected to your choice of inductor.  It would be great if you could adjust the frequency and duty cycle of the timing source "live" while the microcontroller is running.

All microcontrollers have special hardware registers for an on-board timer(s) that can make one of the output pins toggle a square wave where you have total flexibility to adjust the frequency and duty cycle.  There is(are) typically a "timer out" pin(s) on the microcontroller.  This is the signal that would turn the transistor on and off.

Let's assume that you have eight programmable I/O bits to work with, and you configure them all as inputs.

So here is an example of what you could do with the eight input bits:  (Remember the code running on the microcontroler is in a loop reading these bits "live" and that is how you control the frequency and duty cycle)

Let's assume that the microcontroler has some "e-squared" non-volatile memory where you can store some bytes.  We are going to use this e-squared memory.

Bit 0 - "frequency slew up" - when you push on this button the frequency slowly increases

Bit 1 - "frequency slew down" - when you push on this button the frequency slowly decreases

Bit 2 - "pulse width increase" - when you push on this button the pulse width slowly increases

Bit 3 - "pulse width decrease" - when you push on this button the pulse width slowly decreases

Bits 4 and 5 - "waveform store selection A, B, C, D"  - these are two switches that define which of the four waveform storage registers you are currently using

Bit 6 - Store waveform - push on this button to save waveform A, B, C, or D into the e-squared memory

Bit 7 - Load waveform - - push on this button to load waveform A, B, C, or D from the e-squared memory

Thus you have little microcontroller-based timing waveform generator that lets you save or load up to four waveform presets, all operating "live" with separate controls that allow you to adjust the frequency and pulse width (a.k.a. duty cycle) of your timing waveform in real time.

This could be coded and implemented in a week or less if you know C or assembler.

MileHigh
 IP: [ Logged ]
Fraser
Group: Guest
« Reply #3 on: 2009-12-16, 14:45:11 »
Hi MileHigh,

Nice idea, a microcontrolled boost converter using coils with different cores. Aircore is nice with very little core losses but the losses are still there, manifesting as mag field and eventually RF emmision when propogation extends out of near field and velocity gets up to C.

The programming I could do fairly easily in C, it would be similar to my microcontrolled pulse motor testbed pictured below.

When analysing the system as a whole this approach would only give greater efficiency in applications of larger power when the power requirement of the micro could be pushed into the background.

But from a purely experimental perspective it would yield alot more useful data because one could tweak any parameter accuratelty whilst observing transformer efficiency (and ignoring micro power draw), thus one can map out parameter space and find optimum config.

I will say that your idea would be an improvement on standard JT, because in my testing of this pulse motor I was able to achieve X rpm at lower RMS power into coil than when using a standard ###### like setup. Of course in my calculations I ignored power draw of the micro unit and only analysed power draw of the FET and coil. The FETs gate impedance is high enough to disregard any power input from control side.

Such a testbed would consist of control and FET and then just have four sockets for plugging the trafo under test into, with a little extra work trafo efficiency could be calculated by sampling in and out powers (over carbor resistors into filters into ADC) then the micro can exhaustively hunt parameter space for optimum efficiency. Gradient descent over the multidimensional space would minimise the scan time and this approach would be MUCH faster than a human performing the same hunt.

NOTE: in the strobe light pic, the green sector is the delay between reed switch signal and coil fire. The red sector is the pulse width of the coil fire. Both signals are dynamic in order to achieve constant values in degrees regardless of rotor RPM, of course the degrees can be adjusted using the tiny joystick button to the right of the screen.
-------------------------------
gui_in_dark.jpg
Beyond the Joule Thief
* gui_in_dark.jpg (40.77 KB, 512x509 - viewed 15712 times.)
whole_board.jpg
Beyond the Joule Thief
* whole_board.jpg (50.97 KB, 547x371 - viewed 15446 times.)
strobe_mark.jpg
Beyond the Joule Thief
* strobe_mark.jpg (51.09 KB, 547x426 - viewed 15179 times.)
strobe_lights.jpg
Beyond the Joule Thief
* strobe_lights.jpg (43.5 KB, 547x559 - viewed 14710 times.)
2.jpg
Beyond the Joule Thief
* 2.jpg (51.58 KB, 500x472 - viewed 14903 times.)
Last Edit: 2009-12-16, 14:59:04 by Fraser
 IP: [ Logged ]
MileHigh
Group: Guest
« Reply #4 on: 2009-12-17, 05:44:00 »
Fraser,

That's a pretty cool looking setup that you have there.

About the fancy timing setup, for anybody that already has a signal generator then it's trivial.  You just trigger a 555 adjustable monostable multivibrator an you are home free.

With respect to the microcontroller power consumption, there are ways to make it really really low...  How low can you go?

First slow down the clock, you don't need a 10 MHz clock, why not 100 KHz?  A cautionary note - if you waveform timer is on the same clock (most likely) then you don't want to slow down the clock too much.  You don't want the timing interval to become too granular.

On the brighter side, you can make that problem go away.  The solution is to have no forground code running at all so that micro is basically asleep 99.99% of the time.  You assume that you have another available timer inside the microcontroller to generate an interrupt, and all of your code runs on the interrupt.  Otherwise the microcontroller is in sleep mode.  If you can generate an internal interrupt at a ridiculously slow rate of ten per second, then the microcontroller is consuming almost no power at all.

It's like this:

sleep
sleep
sleep
Hardware Timer Interrupt:  Check for a key press?  If true -> Then go run your key press decoding routine.
Else -> You are getting sleepy....
sleep
sleep
sleep

Of course while the micro is sleeping the main waveform generation timer is still working and outputting your pulse waveform.

With this strategy the main power draw for the microcontroller waveform generation system would be to drive the base current into the NPN transistor.  If you have no load on the micro to turn on your switching device then I will guess the microcontroller will consume an average of hundreds of microamperes or less - chicken feed!

MileHigh
 IP: [ Logged ]
ION
Group: Elite
******
Posts: 2122

It's turtles all the way down

« Reply #5 on: 2009-12-17, 14:20:33 »
MileHigh

Agreed a microcontroller approach would be a nice universal tool for tuning various core / winding configurations. It would also use just microamperes of current to accomplish the timing etc.

The biggest problem I see is getting clean switching of the drive transistor. This requires that current be injected quickly into the base (gate) upon turn on and removed quickly for fast turn off or else there will be considerable switching losses. This current has to come from somewhere, and the faster you repeat the process, the more energy you expend, be it bipolar or FET.

So when dealing with a JT circuit, the goals of the circuit must be clearly defined. Also required in the definition is whether the design is a cost no object..one off for personal use, or a least cost production type design that might sacrifice a little efficiency. The two are not the same.

Most of the circuit topologies posted by the JT team at OU do not address this critical issues. That is why I wish we could clearly state our goals with this device. Then I could offer some designs that will satisfy the requirement.
 IP: [ Logged ]
Fraser
Group: Guest
« Reply #6 on: 2009-12-17, 17:30:31 »
@Milehigh,
Nice ideas for minimising micro power requirement there.

@All,
For quick tests I have a pulse generator Tektronics PG501 and an arb func generator Tektronix FG504 I can trigger the func gen in pwm mode from the pulse gen, but your fingers get tired after much knob twiddling lol. So micro is more work up front but makes life easier in the long run.

I agree before optimum solution can be designed then solid requirement specs are needed. Of course the one requirement spec that would be nice to achieve is COP>1. Myself I have a sneaky feeling that COP>1 system will have to reach out far and wide to harvest relatively large volumes of space. So a pot core which contains field almost perfectly might not be good for this, although it would give excellent and best trafo efficiency. Something to definetly consider if the boost converter (JT) is a smaller part of a bigger system.
 IP: [ Logged ]
WaveWatcher

Group: Elite
******
Posts: 1798
« Reply #7 on: 2009-12-28, 16:04:28 »
A simple blocking oscillator is normally good enough for what I do. I love to work with micros but there is one blocking oscillator function that is not inherent when using a micro.... phase lagging synchronization with ambient signals.

Granted, this is not normally needed but it is an interesting feature of a simple blocking oscillator.
-------------------------------
"As far as the laws of mathematics refer to reality, they are not certain; as far as they are certain, they do not refer to reality." - Einstein

"What we observe is not nature itself, but nature exposed to our method of questioning." - Werner Heisenberg
 IP: [ Logged ]
4Tesla
Position: Full Member
***
Posts: 146
« Reply #8 on: 2011-04-23, 07:21:08 »
The most efficient oscillator IC I know of is the LM3909 (LED Flasher / Oscillator).  Has anyone tried using with JT?  I have a couple.. might give it a try.
http://www.datasheetcatalog.org/datasheet/nationalsemiconductor/DS007969.PDF

Tesla
 IP: [ Logged ]
Farrah Day
Group: Guest
« Reply #9 on: 2011-05-29, 10:12:03 »
You have to remember that the original Joule Thief was a simple cct designed soley to utilise the remaining energy of an effectively 'dead' battery source. The cct could run on batteries that would no longer power other ccts and/or devices, hence the name, 'Joule Thief'.  And for this purpose it works very well.

I guess it is the simplicity of the cct and the interesting results that have been obtained through various experimentation and reconfiguration that has led to its popularity.
 IP: [ Logged ]
ION
Group: Elite
******
Posts: 2122

It's turtles all the way down

« Reply #10 on: 2011-05-29, 15:13:04 »
Totally agree, the JT or blocking oscillator is an efficient means for draining the tiny bit of remaining energy in a battery.

But why folks are so excited about this is beyond me. Draining the remaining energy in a nearly dead battery will not solve the worlds energy problems, since the battery was already bought and paid for with resources.

There are even more efficient ways of removing the last bit of energy for utilization assuming you had an infinite supply of nearly dead cells, you could wire them in series to obtain the voltage and current required to operate the device in question. Use reverse protection Schottky's across each cell to limit reverse charging, should a cell go dead before others.

The goals of any switchmode converter should be defined before construction, and such devices are only required when a voltage or current transformation is required from the source battery. Otherwise a piece of wire is the most efficient energy transfer means between source and load.

There are far more interesting aspects of flyback converters or blocking oscillators. The fact that energy is relatively slowly stored in an inductor and released nearly instantly is intriguing. The possibility for using it as a pump to entrain ambient energy during the quick release portion of the cycle should be studied more closely and in a scientific manner.

I don't see this happening in a scientific manner anywhere. People are busy studying how many LED's or neons they can light or relative brightness of an LED. This is most unscientific, considering that a cheap $3 meter from Harbor Freight will be far more meaningful with a simple R-C filter and shunt resistor method.
Last Edit: 2011-05-29, 17:07:21 by ION
 IP: [ Logged ]
Grumpy

Position: Hero Member
*****
Posts: 2709

tExB=qr

« Reply #11 on: 2011-05-30, 02:04:34 »
an open core (like a rod) flyback transformer is supposed to produce a pumping effect at the end of the rod

For Bibhas De fans, this would be the magnetic companion wave, not the electrical companion wave.
 IP: [ Logged ]
Grumpy

Position: Hero Member
*****
Posts: 2709

tExB=qr

« Reply #12 on: 2011-06-01, 04:52:42 »

There are far more interesting aspects of flyback converters or blocking oscillators. The fact that energy is relatively slowly stored in an inductor and released nearly instantly is intriguing. The possibility for using it as a pump to entrain ambient energy during the quick release portion of the cycle should be studied more closely and in a scientific manner.

I don't see this happening in a scientific manner anywhere. People are busy studying how many LED's or neons they can light or relative brightness of an LED. This is most unscientific, considering that a cheap $3 meter from Harbor Freight will be far more meaningful with a simple R-C filter and shunt resistor method.

Would be nice to see some measurements of permittivity and permeability change during the release.
 IP: [ Logged ]
« previous next »
Print
Pages: [1]  
« previous next »
Jump to:  


Powered by MySQL Powered by PHP Powered by SMF 1.1.11 | SMF © 2006-2009, Simple Machines LLC Valid XHTML 1.0! Valid CSS!
Forum and Banner by OUR Team