3v OU Flashlight

[Vortex1]

Itsu:

Your waveforms look quite normal for TKLABS v1.4 schematic.

The modulation and cutoff of the switcher resulting in the sine wave seem reasonable for a secondary loaded with a 100uF capacitor. This is like the switcher looking into a near dead short at high frequencies (not considering leakage inductance and esr of the capacitor.)

If you want more "action" try the Vadik Guk or verpies schematic version posted above. The switcher chip will operate in a more normal mode without such a difficult capacitive load on the secondary.

Regards, ION

[TinselKoala]

This looks just spot on.
As a matter of interest, have you checked the native resonance frequency for the core?

~A

Spot on -- nearly --, but at 1/20 the frequency of Akula's scopeshot. Also the embedded sinus is showing adequate supply voltage to the 34063: it is switching cleanly rather than following the sinusoid during the bursts.

[verpies]

Guys,


Let's think about this more analytically.

The two core halves apparently are vibrating mechanically, creating a variable air-gap between them.
The variable air-gap means that any coil wound on this core  possesses variable inductance (dL/dt).

This variation is in all probability sinusoidal because elastic bouncing and mass-spring vibrations are inherently sinusoidal.

The variation in this inductance can be sensed on-the-fly with high frequency LC oscillator formed by one of the windings resulting in frequency modulation (FM) or by measuring the rise-time of the current ramp (di/dt) when this winding is charged by a constant voltage source.

When a constant frequency PWM modulator chip is supplied with smooth voltage source and subjected to a constant load while its output is connected to such variable inductance winding and its feedback loop is correctly configured - its driving output will experience pulse width modulation (PWM) proportional to the inductance of the winding.

If the PWM modulator chip is operating at relatively high frequency, the integration of these variable pulse widths will yield a low frequency signal that is proportional to the air-gap between core halves. 
It can be argued that this LF signal also represents an instantaneous width of that air-gap ...or relative position of core halves.

All very straightforward and conventional like in a linear inductive encoder.

[itsu]

Inductance of the transformer coils reduced to 56mH and 27mH.
As always, the core halfs firmly forced together by electrical tape with a piece of paper inbetween the 2 outer legs of the e-core.

Highest inductance to W2 as the other way around gave no stable signals no matter how i varied the voltage and/or the pot meter.
No sign of any prolonged functioning after removing the supply voltage (here again 2.3V @50mA)
The transistor (MJE13009) still in use.
Note the pin 13 sign wave now increased to 53Hz.

Next will be further reducing of the inductance to 5.6 and 2.7mH

Regards Itsu

[magpwr]

Inductance of the transformer coils reduced to 56mH and 27mH.
As always, the core halfs firmly forced together by electrical tape with a piece of paper inbetween the 2 outer legs of the e-core.

Highest inductance to W2 as the other way around gave no stable signals no matter how i varied the voltage and/or the pot meter.
No sign of any prolonged functioning after removing the supply voltage (here again 2.3V @50mA)
The transistor (MJE13009) still in use.
Note the pin 13 sign wave now increased to 53Hz.

Next will be further reducing of the inductance to 5.6 and 2.7mH

Regards Itsu

hi Tinselkoala and itsu,

There is one small noticable difference between Akula's version and Tinselkoala along with itsu.
I have attached the edited scope shot to show where i spotted something like "resonance within resonance" at these area.Upon observing carefully i can't help noticing for Akula there is much more of resonance within resonance base on scope shot provided from him.
I only spotted the difference only after itsu provided his latest scope shot which i compared against Tinselkoala's base on my memory then compared against Akula version.
Got to rush off bye for now.....