This build requires the fondling of components charged to HAZARDOUS VOLTAGES.
You need only get close to, and not phsyically touch these components to get a jolt, which may lead to Ventricular Fibrillation.
You can die (or shit your pants, sometimes both) while building this device. We are not liable for any losses or injuries, and cannot answer any support queries regarding this build.
Edit: Yes, you can still die, even if you know how to program an Arduino.
Legality of this build
While hand-held stun-guns are illegal in Australia, cattle-prods and stun batons appear to be available online legally from several suppliers, including at least one Australian camping/survival store. In fact, they'd probably be considered a tool here. How else are we supposed to keep the 'roos off the streets?
As this is a difficult build and you can't exactly conceal such a device in your pants without seeming at least a little suspicious, I figure it should be okay'ish to post about.
That said, getting caught with one in public may still be analagous to carrying an unregistered firearm!
You have been warned.
I've devoted quite a lot of time to R&D'ing Pulsed-Arc (Micro-TIG) welder this past year.
Due to the extended hours spent researching and prototyping in my workshop, people have accused me of 'not having a life' and 'always working'.
I disagree, so I took some time off to create a people deterrent.
It's worth noting that these devices are probably very illegal (if you're caught with them) in Australia (like everything else)
So, don't get caught with one. And definitely don't keep one in your car.
I guess taking one camping might be fine though.
We're stepping up the four or so volts available from a standard'ish 18650 Lithium Ion cell, to just about ten Kilovolts at the output spark gap through several stages.
First, two boost converter circuits are wired in parallel to double the output current (Diode-ORed). This bumps the four'ish volts at several amps available from the Lithium cell, to just over 13 volts at 700 milliAmps required by the Royer Oscillator.
The converters used are based on Linear Technology's LT1308 single-cell chips, junkbox parts I had on-hand from a previous project.
Next, the 13'ish volts is fed into a relatively unmodified off-the-shelf cheap-as-shit CCFL inverter, designed for 300mm CCFL tubes. These are commonly used in cars, PC cases, and other places where people feel the need to compensate for their small cocks.
Most (if not all), cheap fixed-brightness inverters are designed around a Royer Oscillator (resonance witchcraft), this drives the transformer at it's natural resonant frequency while utilising a relatively small component count.
The Royer Oscillator magically turns the ~13 volts at 700mA applied to its input, into just under 900 Volts AC at a couple of milliamps at its output - They're usually made for 12volts, but we're being naughty here.
Finally, we feed the ~900 Volts AC into a twelve stage Villard Cascade (a type of Voltage Multiplier). The gives us ~10 kilovolts across our output sparkgap, with a fire rate of 5 ~ 10Hz.
A One kiloVolt commercial spark gap has been added in series with the output sparkgap. This helps protect against direct short circuits, and also ensures that at least 1kV will be discharged into our
victim research subject, in the event of direct skin contact.
This design was thrown together very roughly in one full day using existing and/or recycled parts; I really didn't expect to write about it, but figured it might make an interesting read for some.
In order to complete this write-up, I'm building and documenting a second unit from scratch - Apologies for the poor'ish quality photos, I'm a little short on time.
Yes, the startup "whizzz" is just for show! (intimidation plays a big part with these things)
Blue LED means the Boost Converter is active, and the push button is unlocked for use.
Red LED means the Boost Converter and FET Drivers are switched off (Standby mode, saves ~220mA). Pressing the push-button once will wake up the device once in standby.
Alternating Red/Blue means a discharge is in progress
All designs are provided as freeware for non-commercial use
[Download] Firmware Binary - Microchip HEX format
[Download] 2SK4033 N-Channel MOSFET (boosted uC card)
[Download] MCP1252 Inductorless 5V Buck/Boost Reg (boosted uC card)
[Download] PIC12F1840 8Bit Microcontroller (both uC cards)
[Download] MCP73833 Single-Chip Charge Controller (charger card)
[Download] LT1308A Single-Cell Boost Converter (Boost converter cards)
[Download] MUR1100E High Voltage Power Rectifier (Villard Cascade)
[Download] A71H10xx3820 1KV Surge Arrester (Villard Cascade)
01 - On the microcontroller card, I'd advise adding an external 47K pull-up resistor between the switch (before R7) and VDD, as the PICs internal weak pull-ups are more than weak
02 - The huge currents pulled from the Lithium Ion cell may cause some protected cells to enter shutdown mode after a few seconds of usage (the cell output is switched off until the load is removed). Increase R2 on the Boost Converter cards to 12K, and decrease the lead spacing of the output sparkgap (or just limit in-air discharges to no more than a couple of seconds!)
03: The "DandeLiIon" charger card has two embedded PCB fuses; if the polarity of the battery or charge inputs are reversed, a protection Diode will conduct, blowing the fuse. Get the polarity right!
04: Once the unit is powered down, a hazardous (painful) voltage may still exist across the output spark gap. The is due to residual charge in the Villard Cascade capacitors; you can discharge the caps by shorting the output onto something metallic
No known issues (yay!)
IMPORTANT: If your CCFL inverter has an Electrolytic bypass capacitor across it's input, ensure it's rated for at least 25 volts!
Exceptionally shitty units, such as this one, don't have any bypass capacitors installed. Adding one won't hurt in this case and may actually improve performance
Important: Apologies for the possible confusion here. The only FETs I had on-hand have an enhancement voltage (VGSth) higher than the nominal supply voltage (3v7)
As I'm short on time and suppliers are closed for the long weekend, it was far quicker to re-engineer the board to include a small MCP1252 boost converter for the FET's gate drive (only good for slow'ish switching, which is what we're doing).
I have included PCB artwork with and without the converter circuit for your convenience.
TL;DR: I messed up. The board shown in the images below may look a little different, but the wiring is effectively the same.
Don't forget to flash the PIC before installing it! You can buy a Pomona 5250 SOIC clip for about $20 from Mouser. It can also be done in-circuit with the same clip.
Important: I ran out of thermal adhesive. In it's place I used good quality thermal grease (CW7270) in the centre of the board for heat transfer, and Araldite around the edges to hold the board down.
Also, ensure you file the ground feed-throughs as flush as possible with the board (without compromising the ground connection); the more heat we can suck out of the PCB and dump into the heatsink, the higher our charge current will be.
Don't bust your balls on the above though, we'll never be able to obtain the MCP73833's 1-amp charge current with a simple homebrew PCB. The charge current is currently set to 450mA.
There are many different ways of skinning this cat. For example, the pins from a fluorescent starter would make for an excellent pair of contact probes - I'm keeping it simple here with some solderable galvanized fencing wire.
It's held up to daily discharges for the past few months without issues, aint broke, not fixing
I'm not terribly good at mechanical stuff.
Don't get into any trouble - and if you must, please keep me out of it.