KA4EP Torch (kacher torch)

[d3x0r]

Okay, so this started as akula lantern4 and Lasersaber joule looper (tesla torch)...

Like a kacher; a low inductance drives a high inductance, and the self-feedback tunes oscillation.


L1 is Primary
L2 is Secondary
D1 is base diode
D2/D3 are load diodes
C1 is Power supply
C2 is the topload capacitor


B1 is battery (optional external power source)


Topload would be equivalent to a tesla coil topload.

[d3x0r]

Top load of kacher coil is a capacitor that allows a high voltage to be produced at the load.
The low end of the secondary that goes to the base of the transistor has a diode to ground, this reduces power draw significantly by dumping excessive low voltage back to the low side of the circuit.  It doesn't seem to bother an NPN transistor to have a negative voltage on its base, but it's a point of recovery that can be used.


The output beyond he capacitor, the positive part goes back to the positive power side of the circuit.  This turns out to be a very tiny current, but again, it's otherwise a loss that can be recycled into the system.  The other side generates a low voltage that drives LEDs from negative power back to ground.


The tuning of the top-load capacitor will depend on the ratios of inductance and voltage driving the system.  If a large ratio is used, then there will only be a small current flo w and the capacitor should be small.  This allows a large potential to develop in that capacitor.   If the capacitor is too large, it will flatten out the voltage.
If a small ratio is used (1mH:3mH for instance) this capacitor should be larger since a larger current will develop in the secondary.  It is possible that the capacitance becomes irrelavent at certain transformer ratios; however for all tested scenarios it has always reduced current draw to have a capacitor there.


To the left is the power supply capacitor/battery/.... It is meant to be a very large capacitor that has a stored energy in it for the system to run.  If a battery is used, a capacitor should remain there to allow the load/output to recharge a low impedance source.


All diodes should be as fast as possible... but may be slower depending on the transistor.  I did have a successful version that used a 2n2222 and 1n400x diodes. 


The original Brovine Kacher circuit has some resistors around the base to power and ground, a high resistance connection from power rail to the base may be required for starting the system; I have occasionally had to touch the diode on the base to start the system manually.

This circuit easily generates -100V from 2.7V input.  The voltage against the collector is not very high, unless a high voltage load is created and a low ratio coil is used.


The voltage against the base is also almost always very low, because as soon as the transistor starts conducting, the secondary attached to the transistor goes low.


If the circuit is configured to supply a high voltage chain of LEDs, you may find that connecting fewer LEDs will increase their brightness; however this will increase the current requirements of the circuit... the greater the load, the lower the required current, but the less current flows through the load; Up to a point... at a certain point, further series LEDs added will increase the current draw, but not lower in brightness.  C2 May be changed to tune this appropriately. 


There is no effective resonance in the system... topload capacitor tuning is purely a direct relation to current generated by the secondary, and requirements of the load.  You might be able to enable a resonance in the secondary by removing D1; this will immediately create a higher current draw.

[d3x0r]

Low voltage version...
Meant for up to 2.7V input (supercap input)

[d3x0r]

TL494 external oscillator

[d3x0r]

While measuring this circuit, I ended up finding that there was a continuous 15mA draw from my battery.  I wanted to see how long 2 AAA batteries would last in this, and it turns out it will be something like 3 days (76 hours). 
I found that the capacitor used will affect the input current used from external sources.  C1 can be reduced in size if a external source like batteries are applied.
The scope shot shows a continuous draw from the battery into the capacitor.  THere is a peek draw when the transistor starts conducting and the primary in turn starts conducting, which is the lead-in spike; this action causes the secondary to generate a low voltage to the base of Q1 immediately stopping the conduction of L1; After this the system needs to relax, and it's actually the start of a second oscillation of the system that starts it conducting again.

(approximate events.  Phase 1 generates a low current through the topload to the real load through D3, phase 2 is a relaxation that provides a slight positive potential increase for C1 through D2)
1) Q1 base goes high
2) Q1 collector goes low (down from supply voltage); starts conducting
3) Q1 base goes low; secondary has induced current from C2 to Q1 base; excess current is drained to ground
4) Q1 collector goes higher than supply voltage; primary continues forward conduction after Q1 is opened (stops conducting).
5) when the secondary stops having a current, there is a positive potential in C2 that causes the secondary to start another current, this generates a positive on the base of Q1. (goto 1)

I've used several NPN transistors for kacher circuits, and really since this is no different, I see no reason that an appropriately scaled higher current transistor cannot be substituted for other configurations.  This could be an air core, or small ferrite core...


----
Attached image the yellow trace is the current draw from the battery into C1. 
Tried to break out a separate graph of the current to label phases (T1/T2 ...) and figure out 1) when the base actually goes high, and 2) why there was a constant current draw. 
I subsequently learned that the type of capacitor used for C1 in conjunction with a battery/power source can have an impedance that requires a consnant supply, but there are capacitors that immediately relax, in the middle low current phase, when the current is going through D2. 
The base goes high because the topload has a resisdual charge in C2 from the prior phase... and that is actually a delay from phase 2 completing...