Quote from: MileHigh on December 09, 2012, 02:30:26 AM
TK and Bill:

Just to complete the point about resonance from two weeks ago.

How does a 555 timer work?  Current from the supply source flows through a resistance to charge up a capacitor.  When the capacitor reaches 2/3 of the supply voltage a comparitor is triggered and the capacitor is discharged and the process starts all over again.  So the value of the resistance and the value of the capacitance determine the operating frequency for a 555 configured as an astable multivibrator.

Now, would you call that "resonance?"  It's not resonance and it has nothing to do with resonance.  It's just the "operating frequency" of the device that is determined by the timing components.

By the same token, a JT has an "operating frequency" that is determined by the values of certain components.  For all intents and purposes a JT is just an astable multivibrator that charges an inductor and then discharges the inductor through a load that is typically an LED or a string of LEDs.

Again, as a reminder, real "resonance" means that energy is circulating back and forth between an inductive and a capacitive set of circuit elements.  It could be an electrical LC tank circuit, a ringing bell, or when you blow air across the top of a beer bottle to make it sound a note.

Resonance is always associated with observing increased voltage and increased current in an LC tank circuit.  People see that and they think that perhaps "extra energy" or "extra power" is somehow manifesting itself when in fact that's not true.  An LC tank circuit will absorb energy to the point where it is in equilibrium - the energy being put into the storage system is equal to the energy being burnt off in the storage system.  There is a pervasive fundamental misunderstanding of what real resonance means.

So why did Bill's JT circuit make the LEDs jump in brightness when he changed the value of the base resistor?  I don't know the answer but the way to find out why is with an oscilloscope.  Like I said before, the LED will increase in brightness when the JT circuit puts more average power into the inductor.  Sometimes astable oscillators like the JT circuit will have characteristics called "metastability" where for a reason that could be determined with enough diligence and investigation, you will find that the circuit can "snap" into a certain operating mode that can give you increased brightness from the LEDs.

It's a challenge to figure these things out and you learn in the process that there is no phenomenon of resonance at play in a JT circuit.  For what it's worth there is no "lead out" process happening either.

MileHigh

MH:

I would not think that the 555 timer circuit would hit resonance, but I still believe that a JT can.

Check out this video done by Stefan a few years back on an early replication of Dr. Stiffler's work:

http://www.youtube.com/watch?v=qRw_sCzhFnk&list=FL0bTBCRogMzrYTQT3pbhxwg&index=397

I know this is not a JT circuit as it has many more components but, he hits resonance and it has a coil over a ferrite core.  I will play around with one of my simple circuits and my scope to see if I can document what happens at the "sweet spot".  If it looks like anything, I will post a video.

Thanks for the reply.

Bill

PS  I thought of Stefan's video when I was trying to explain the jump in brightness while fine tuning the JT circuit.  It is a non-linear jump just as you see happens to Stefan with his circuit.  You can see how the brightness jumps to a higher level when he says he hits resonance.  This is what made me remember his video.

Bill:

I looked at Stefan's clip where he replicates a Dr. Stiffler circuit.  I haven't watched any Dr. Stiffler clips in a long time but I recall how the majority of them were based on a transmitting oscillator based on some sort of LC tank circuit exciting a receiving LC tank circuit tuned to the transmitting frequency.  So indeed resonance is at play at both ends of those circuits.

When you look at your JT try to focus on the charge and discharge cycle for the main inductor that drives the LEDs.  In theory an inductor charge cycle (transistor on) that is say between three and four L/R time constants, and a discharge cycle (transistor off) that is only long enough for the coil to discharge perhaps 90% of its stored energy will give you maximum brightness with a fairly efficient power consumption.  The whole idea being to keep the pump (the inductor) running with as much energy throughput as possible with a minimum of "dead" time.

There are two "dead" times for the inductor.  If the transistor is on for too long, say more than four L/R time constants, then most of your supplied battery power is just being burnt off in the resistance of the wires and also in the internal resistance of the battery itself.  The other "dead" time is when the the transistor is off for too long.  So the LED flashes and the inductor is fully discharged and doing nothing.

I am not factoring in the persistence of vision here, and just focusing on keeping the main inductor "busy" all the time either charging or discharging.  I am speculating that this will give you the brightest LED display.  However, it might not be the most efficient in terms of how much battery power you used to get the level of illumination.

I am still trying to demystify the JT and I will say it again:  It's just a circuit for charging and then discharging an inductor through a load, nothing more than that.  If you ignore the property of the circuit where it can work at very low voltages, then a CMOS 555 timer doing the same coil charge-discharge cycle will easily outperform a JT circuit and give you full control over the timing of the charging and discharging, the size of the inductor that you want to work with, and the initial current through the inductor at the start of the discharge cycle.  This means that you can decide exactly how much energy you want to discharge through the load and you also have control over the L/R time constant and can decide how much energy you are willing to lose due to resistive losses in the coil wire itself.

MileHigh

Here's a neat little treat. I've been playing with this Arduino inductance tester and the way it operates is very cool.

Take a look at the scope trace below. Using a simple external circuit with an op-amp voltage comparator chip, the Arduino sends a brief pulse to the inductor under test, which is wired in parallel with a fixed and known capacitor to make a resonant tank. (lower trace). The tank responds with a ringing at its resonant frequency when the brief pulse stops. The comparator chip responds to the zero-crossings (voltage reversals) of the ringdown and produces a solid clean pulse at the same frequency, but not decaying, as the inductor-capacitor tank's ringdown. (upper trace). This pulse is read by the Arduino itself and the frequency is used to calculate the final output of the inductance value of the device under test.

The possibilities should be obvious, although I just now thought of them. It should be pretty easy to route the comparator output back into the tank circuit and have a situation where the system seeks and maintains its resonant frequency automagically.
Some of these comparator chips will work on very low power supply voltages, not just input voltages too; the KA393 that I found in a TV circuit board has two independent comparators in it and will work on 2 volts power supply, and as the scopetrace shows it responds to tiny voltage differences, detecting the ringdown until it's almost totally dissipated.
(top trace at 2v/div, bottom at 5v/div, sorry I didn't position it on a graticule line, my bad)

Lets back to joule thief. We all know that ferrite start to sing at audio frequency because of magnetostriction. Suppose we get a magnet with coil around it and connect it in series with the primary of ferrite transformer and we tune it to the loudest sound that ferrite can produce, is it possibe to self oscillate it like Magnetic Resonance Amplifier (MRA) did? I imagine if we tune the ferrite to sing vigorously, the magnet near the feritte transformer will also vibrate and will induce voltage to the coil around it thus supplying additional voltage to the ferrite primary to keep oscillating and singing. This will need a conditioning process like MRA did. When the conditioning process is finished, in theory it powered itself and the ac supply can be removed.

http://www.overunityresearch.com/index.php?topic=1596.msg27361#msg27361


A 4 LED JT works better than a 1 LED JT???


Resonance or magic???