Shared Projects by dewhisna

MIDI Interrupter for Dual-Resonance Solid-State Tesla Coil using dual ATmega328PB micros for 20-note polyphony. This is a more compact alternative to the Quad-ATmega328P 24-note interrupter.

This board has been displaced by the better Dual-Resonance Solid-State Tesla Coil MIDI Interrupter STM32F103 board.

This is an implementation of Lionel Sear’s Wall-Clock Joule Thief circuit, found here: https://rimstar.org/science_electronics_projects/joule_thief_powering_clock_ls.htm and described by him in this YouTube Video. And featured in this Hackaday Blog.

The circuit itself makes sense and should work in theory. I’m tired of replacing batteries in my wall-clocks so often and figured I’d give it a go.

The reference designators on this board are the same as in the original circuit at the above link. I did, however, use BC847 transistors instead of the BF546B specified on his original circuit. You should be able to use any similar NPN transistor. For C1 and C2, I opted for tantalum parts to make a smaller footprint. Those are Case-A 1206 size. Their ESR is slightly lower than electrolytics would be, but I believe will be a negligible change. For the diodes, I selected MBR0530 parts since I had them on-hand and they are low voltage-drop schottky type. You should be able to use any similar diode there, such as a LL4148. The resistors are 0805 footprint and the potentiometer is a Bourns 3362 series part.

The toroid I selected was a Magnetics, Inc. ZJ-42206-TC part, found on eBay. I selected that one because of its high permeability and small size (and inexpensive eBay price). Similar spec parts would be the following from Mouser (623-5975001801, 871-B64290L0638X035, and 871-B64290L0638X830).

The toroid mounts on the backside of the board. And assuming I’ve interpreted the toroid windings correctly, the four winding connection points are labeled 1 through 4 with the start-winding connection on “A” and the end-winding connection on “B”. In other words winding 1’s start would connect to “1A” and its ending would connect to “1B”. And so on for the remaining three windings.

UPDATE: This circuit does indeed work, at least electrically. And it does suck the last bit of juice out of a “dead” battery, ending at around 0.5 volts (0.523volts on my first test), just as Lionel said it would. However, it does NOT prolong battery life in general. The big design flaw in Lionel’s circuit seems to revolve around the output voltage limiter circuit, which turns off the Joule Thief by shunting the input through a 1K resistor via an NPN transistor.

This means the circuit draws approximately 1mA continuously. Lionel himself explained that 1mA draw in his video, found at the above links. Unfortunately, a continuous 1mA load is quite large for any small battery circuit that’s expected to run for a prolonged period.

Present testing of this circuit on several different wall-clocks and several identical circuit assemblies has shown an average of about 2-½ months life from a new AA battery with this Joule Thief. That, compared to a minimum of six months to a year with just the battery itself in the same wall-clock without the Joule Thief.

So feel free to build and experiment with this circuit, but don’t expect any “free-energy” nor prolonged battery life from it. I’m planning to investigate some circuit changes to replace that shunt regulation with a switch style regulation and use a PNP transistor to turn on the Joule Thief when charge is needed rather than the NPN shunting it off when it’s not.

Bifilar Pancake coil for experimentation. Each winding has 40 turns of a linearly-increasing radius (Archimedean Spiral). Inner-Spiral Spacing is 12.5mils (25mils center-to-center of alternating winding and 50mils center-to-center of same winding) and the Spiral Trace Width is 12.5mils. The calculated inductance of each winding is 62.857uH and the calculated combined inductance in series is 251.428uH (due to the mutual inductance effect).

Update: After receiving and measuring the board, I was happy to see that the measured values were extremely close to the above calculated/predicted inductance values:

Measured Inductance Values:

• Inner Winding: 65.41uH

• Outer Winding: 65.16uH

• Windings in Series: 288.1uH

Measured Resistance Values (DC Impedance):

• Each winding: 10.5Ω

• Windings in series: 21.5Ω (also included wire connecting them in series)

Measured Winding to Winding capacitance: 383pF

Note: Measurements are subject to the error of my test gear, but seem somewhat reasonable and consistent with calculated/predicted values.

NOTE: This board has been replaced by DRSSTC Driver V2.

Driver board for a Dual-Resonance Solid-State Musical Tesla Coil. This design features a zero-current-switching phase-lead circuit, an over-current detection/lockout circuit, an under-voltage lockout circuit, and a PLL-based adjustable dead-time delay. Requires regulated 15VDC power supply. It’s designed to be used with the Dual-Resonance Solid-State Tesla Coil Full Bridge board.

Drive two of these with the FPGA-Based Dual-Resonance Solid-State Stereo Music Tesla Coil Interrupter.

The debut performance for this board was during Shock-or-Treat Halloween 2018, and can be viewed at: http://mediadrop.dewtronics.com/media/media/shock-or-treat-halloween-2018.

Note: This board is still in the prototyping/testing phase and hasn’t been fully qualified yet. Build at your own risk.

Errata: Oops, just realized I goofed the via hole size. They were supposed to be 25mil, but are 35mil instead. The vias are 46mil diameter, so it’s still within fabrication specs, but just barely. I hate cutting things that close to minimum tolerance specs.

Also, resistor R5 should be a 330-Ohm instead of 470-Ohm to allow the Blue LED to be a little brighter, as Blue LEDs require more current. And, the 1K-Ohm R9 should be at least a ½-Watt part and preferably 1-Watt, or you will risk burning it up from the high-voltage peaks of the current-transformer in the zero-current-switching circuit, which occurred during the first static-load test performed on this board. Unfortunately, finding such a large wattage part in a 0805 footprint will be difficult, if not impossible, meaning a large through-hole part will have to be graft in.

NOTE: This board has been replaced by DRSSTC Bus Supply V1 and DRSSTC Bridge V2 as a modular pair.

Full-Bridge power circuit for running a small desktop Musical Tesla Coil from 110/120 VAC. This design features a Soft-Start power-up circuit and relay controlled power-off voltage bleeder circuit. Note: Soft-Start and power rectification circuits are NOT compatible with 220/240 VAC! Recommend using 2oz Copper.

This board is designed to be driven by the Dual-Resonance Solid-State Tesla Coil Driver board.

The debut performance for this board was during Shock-or-Treat Halloween 2018, and can be viewed at: http://mediadrop.dewtronics.com/media/media/shock-or-treat-halloween-2018.

Note: This board is still in the prototyping/testing phase and hasn’t been fully qualified yet. Build at your own risk.

ERRATA: Do NOT build this circuit as-is. The soft-start circuit has a design flaw that causes SCR (D29) to false trigger early and to “bounce”, which causes the Soft-Start Relay RL1 to chatter and fuse its contacts due to the high inrush current, ruining the relay. The high inrush current will then fry one or both mains input rectifier diodes D32 and/or D33, and then blow the fuse.

If building, do not install the following components: RL1, D29, D30, D31, R5, R6, R7, R8, R9, RV1, C2, and R14. In the place of R14, install a NTC Thermistor Inrush Current Limiter, such as Ametherm MS22-10008 (DigiKey: 570-1003-ND) – note that this can be installed between pads 41 and 44 of the RL1 footprint for closer pin spacing, but give it plenty of air space and ventilation, as it will get hot. You can also use a variac to slowly bring up power to the entire circuit – but you’ll need to start it at around 70 volts to keep the Bleeder Relay RL2 from chattering.

Also, resistor R10 needs to be a sufficiently high wattage part. It will have about a 155-160Volt drop at 3.3mA, which is 530mW. So, use at least a 3/4Watt resistor for R10 or risk it leaking its “magic smoke”.

Note that at circuit power-down, resistor R15 will get a little toasty. This is normal. However, your driving circuit should monitor the ‘fault’ pin (pin 3 of J1) and shutdown all AC input if the fault condition (i.e. Relay RL2 open) exists for more than a few seconds, as the 2.5W rating of R15 is only sufficient as a bleeder resistor for the system. A continuous load resistor of that value would have to be around 120Watts. If the 2.5W part was left in the circuit very long (longer than the required voltage bleed down period), it would either burn in two or catch fire.

Also, cut the copper-pour trace for the ground fill where it runs underneath the fuse holder. Failure to do this will cause a short between the 120VAC mains and ground under the 120VAC-input end of the fuse holder when the solder-mask gets breached beneath it. Apparently, I should have made the area under the fuse a keepout-area for that copper pour. It is, however, a self-correcting problem, because when it shorts, the trace, and select parts of the fuse holder, will vaporize. But it is a rather loud and startling bang, and you’ll probably have to reset the circuit breaker for your power feed.