Akula 3 Volt Flashlight

[ion]

IMHO, the schematic from the Wesley video with the grayed areas and big X through the inverter is a less viable schematic than the schematic offered by verpies,  Vadik Guk and others. Is there really a good reason to use that schematic. What is the rationale?

The schematic I question has an FET or transistor and resistor doing little more than loading the supply and producing the sine wave ripple. This seems to be wasteful of input power.

Most of the other Akula schematics show the mixing of two frequencies in the transformer. I will pursue that approach.

I get a lot of interesting audible ferrite singing at certain settings of the 4069 frequency adjust pot.

Of course since no one has a runner, it is only an opinion, to each his own.

My build attached.

[TinMan]

ION
In regards to the 3:1 transformer winding's,is this going by resistance ratio,or by the number of turn's?

[ion]

ION
In regards to the 3:1 transformer winding's,is this going by resistance ratio,or by the number of turn's?

Transformer ratios are traditionally specified as a turns ratio unless otherwise specified e.g. inductance. I have not ever seen a resistance ratio specified.
The transformer I intend to use has not been wound yet, shown is one I had on hand with about 2:1 turns ratio.

[TinMan]

Transformer ratios are traditionally specified as a turns ratio unless otherwise specified e.g. inductance. I have not ever seen a resistance ratio specified.
The transformer I intend to use has not been wound yet, shown is one I had on hand with about 2:1 turns ratio.

That was my first thought,but better to check first,as im still catching up on everything here about the device. I expect to recieve my core's this week in the mail,along with a few other item's.

[Itsu]

After extensive tests of the TKLABS V1.4 setup (see first diagram) in which i used severall E-core and pot-core transformers with different inductance's (600mH/100mH, then 57mH/26mH and finally 5.7mH/2.6mH) in all possible hookup configurations, i finished with this testing because never any sign of extended run times, after removing the supply voltage, was seen.

Severall screenshots and video's can be found on this thread on overunity.com showing the results:
http://www.overunity.com/14524/3v-ou-flashlight/msg401832/#msg401832



I now modified the circuit to the so called Vadic Guk circuit (see 2th diagram) which does seem to have a more logical reason for the used transistor/MOSFET as mentioned by ION above (see picture 3).

There are some minor differences as i used the components when possible from the old circuit like R0, R1 and R2 (Valik diagram) are still 0.33 ohm, 3.3K and 180 ohm.

The transformer used presently is a pot-core 4.7cm diam. Al 9500, and W1 is 35mH, W2 70mH (when clamped together tightly).
I have i copper tape strip first on the former (gap 2mm) then W2, a layer electrical isolation tape, then W1 (the same direction as W2) and finally a copper tape strip shielding all this (gap 2mm).

First runs show i am missing my trigger point which was the blue trace on the first diagram, so i use now pins 4/5 of the 4069 to trigger on.
The leds (both show different behaviour) stay on longer after removing the supply voltage as they did on the old diagram, but after a few seconds they fade.
The MOSFET/transistor drain/collector signals are much more noisy now due to the inductance it is now switching.

I will have to try again all sorts of possible combinations for hooking up the transformer and tuning all parameters.

A short video of the running thing is here:  https://www.youtube.com/watch?v=DiPQuVYxJB4&feature=youtu.be


Regards Itsu

[ion]

BTW, the attached schematic (circuit.png)makes a bit more sense than most of the schematics I've seen as far as the operation of the coupled inductor, the transistor VT1 acting as a synchronous rectifier switch to utilize the hi Q ringing of the tank circuit (W2 and copper foils) and add to the 470uF. This is more in line with the Vadik Guk circuit and original Russian circuits we began investigating many months ago here.

As shown, with a low frequency drive for the transistor, and un-synced to the main oscillator, it is hit or miss (asynchronous) rectifier, hoping to  close on the peak of a positive excursion of the ringing frequency and open at the peak of the last positive, it will add 1/2 cycle of the ringing energy to the 470uF storage cap. It will zero to a null, operating randomly (asynchronously) unless the LF oscillator can lock to a subharmonic of the main oscillator. Not shown on this schematic is a trimmer for the second oscillator, which should be like the others as I believe it requires careful adjustment.

I think the idea of the copper foil shields was to tune the resonance of W2 and possibly some other odd effect You could add a small mica trimmer across W2 and see if you could tune that winding for resonance. I will try the same. Note the drawing of the waveforms on the schematic, as W1 rings down W2 is shown to increase in amplitude.

I hope this does not add to the confusion and perhaps can add some insight regarding my understanding or best guess of the operation of the circuit.

P.S. circuit values need not be changed as both 34063 oscillators are fairly similar, except perhaps the trimmer for the second oscillator.

[Grumage]

Dear All.

Have a look at this !! 

https://www.youtube.com/watch?v=6B79UJGoNJE

Simplicity !!

Cheers Grum.

[ion]

Anyone notice the HF ringing on the ramp? It seems to exhibit very little decay.

A high impedance voltmeter hooked across the cap would give much info regarding rate of discharge vs. power drawn by the LEDs.

Remember that LED's can look very bright when hit with a very short duration high current pulse, the retina has memory. The pulse repetition rate need only be around 20 pulses per second to appear to be lit continuously.

[Hoppy]

Remember that LED's can look very bright when hit with a very short duration high current pulse, the retina has memory. The pulse repetition rate need only be around 20 pulses per second to appear to be lit continuously.

I think there will improved running times as inductance of the coils is increased and performance optimised by increased coil resistance (adjustment of wire gauge). You make a very good point as LED's only need sharp micro amp pulses to get them glowing. As the LED threshold voltage is reached the storage cap starts to pulse around the LED threshold voltage level, thereby further reducing load current. Coil design is clearly important to optimise performance. However, I'm fairly sure Akula is exploiting this effect by using a high cap value (and possibly a button cell battery or two) and most importantly short video run times, as his LED load for the 30W device is large compared to the other LED self-runners demonstrated.