High-voltage AC generation using relays

An intriguing project from the 160-in-1 project kit (seen in Electronic project kits: hands on with a vintage 160-in-1) is #44, the High Voltage Generator II:

Testing with a fresh 9V battery, the claim is true: “If the Batteries are fresh enough you may be able to feel the high voltage pulses by touching both probes at the same time. Happy zapping!” Wired up the circuit as follows:

touched the high-voltage leads of the transformer between thumb and index finger, able to feel a slight tingle. How much voltage is this circuit producing? Measuring with a multimeter reads about 25 VAC. Not bad from a nine volt battery, where previously to feel the electricity you had to hold it up to your tongue. Now you can use your fingers instead.

Theory of operation

How’s this circuit work? The description claims “like a doorbell buzzer”:

The Relay is connected like a doorbell buzzer to obtain square-wave pulses of DC voltage across the series connection of the 100 µF Capacitor and Output Transformer winding. This causes the charging and discharging current of the Capacitor to flow through the few turns of the Output Transformer winding.

…require some explanation. This instructable How to Make BUZZER From RELAY Switch also uses a similar circuit, some alternatively call “relay chattering” instead of buzzing. Taking a closer look at the schematic:

The idea is simple: the relay starts closed (using NC=Normally Closed contact), allowing current to energize the relay coil. This opens the relay, breaking the circuit, therefore cutting power to the coil so it subsequently turns off and falls back to closed. Once closed, the coil is energized again and the process repeats.

The rate of cycling is controlled by the capacitance. Replacing the 100 µF capacitor (terminals 69 and 70) with the 470 µF (terminals 71 and 72) produces a much longer pulse length, you can hear it switch in realtime.

This circuit is also known as a vibrator (electronic); from Wikipedia:

This “vibrator” is essentially a relay using normally closed contacts to supply power to the relay coil, thus immediately breaking the connection, only to be reconnected very quickly through the normally closed contacts. It happens so rapidly it vibrates, and sounds like a buzzer.

Pulses of DC effectively gives AC (see also: pulsed DC), so this current can be fed to a transformer to step up/down the voltage. High(er)-voltage AC from low-voltage DC, could be useful. The output AC voltage can be smoothed back to DC, creating what is known as a chopper: converting DC to DC at a different voltage, as demonstrated in project #93: Mechanical Chopper[sic] (DC-to-DC Converter):

What are these electromechanical choppers used for? The description says:

The mechanical chopper was very popular before the days of Transistors, but now Transistors have replaced these in all new designs. Your parents will remember the days of the automobile radio “vibrator”. The vibrator performed the function of the Relay in this circuit.

The 160-in-1 Electronic Project Kit manual was published in 1982, so this circuit is now very obsolete. Project #91 DC-to-DC Converter contains a more modern solid-state transistor-based design:

Nonetheless, I figured it could be interesting to explore this archaic DC-to-AC conversion technique.

Screenshot from EEVblog2 1000fps High Speed Mechanical Inverter:

which was sent in EEVblog #922 — Mailbag (17:23). A mechanical multivibrator for 12 VDC from a car cigarette lighter jack, to ~120 VAC for an electric shaver, supporting up to about 15–20 watts. The Alec-Tri-Pup:

Anyways enough of history, time to get back to the present.

Prototyping the beginning of a standalone chopper

The 160-in-1 kit projects are cool and all, but I wanted something more permanent and standalone.

Salvaged in Emerson MW8675W microwave oven teardown, a good relay: Omron G5J-1-TP-M 24VDC Contact: 16A20VAC 16A30VDC Coil: 24VDC, 0936YH. Part of Omron’s Power PCB Relay series, G5J. Datasheet:

The coil runs on 24 volts, a 9 volt battery won’t trigger it directly, fortunately I had a 24 volt wall wart on hand. Wiring it up to the coil works no problem.

However, there is a problem: this relay only has a Normally Open (NO) contact, there is no Normally Closed (NC)! How can we solve this? My idea was to use a MOSFET to switch the relay, acting as an NMOS NOT gate, where the load resistor is the relay coil itself. Had a STP60NF06FP N-channel 60 V Power MOSFET (datasheet) on hand, built this:

When the relay closes (note: the G5J is a SPST relay, but KiCad only had a DPDT relay symbol so that’s what is shown here), the gate is pulled high, turning off the relay coil. And it works in practice:

Except, the transistor gets quite hot, so does the relay, sparks fly, it oscillates too fast with no cap, missing flyback diode, and other design flaws. Good enough for a prototype, but let’s see if we can improve and build it on a PCB.

Soldering the circuit

Soldering the relay and MOSFET, and a connector salvaged from an oven:

on the reverse side:

The relay coil is wired to power and ground through the MOSFET’s drain and source. We can test the switching action using the white and black wires here, touching them to the MOSFET’s gate, and measuring the continuity on the contacts (on the other side) with a multimeter. Powered up, the relay closes. Touching the white wire to the gate opens the relay, as we expect.

The gate could now be wired through the relay’s contacts, but first…

Spark suppression

When the power is removed from the coil, usually a spark occurs:

sometimes even bigger (but blurrier):

due to the inductive load of the relay coil. This can be addressed with a diode, as explained in StackExchange: Why is there a diode connected in parallel to a relay coil? I used a 1N5404 diode I had available, rated for 100 V peak reverse voltage (datasheet), towards the positive:

the sparks seemed slightly smaller, but still present. An additional technique is described in project #40, Capacitor as a Spark Suppressor:

The “proper size capacitor” is problematic. StackExchange: Preventing relay contacts from sparking on disconnect has some equations to calculate the proper capacitance needed based on the relay coil’s inductance in henries, but I haven’t a means to measure this value yet. In the 160-in-1 kit they use 0.1 µF for the 9 V relay but the 24 V relay I’m using is completely different.

Adding a capacitor is a good idea for another reason: controlling the rate of oscillation. Recall the 44. High Voltage Generator II schematic:

The capacitor is in parallel with the coil. Made these changes:

Measured 14 VAC output across the coil (in parallel with the diode and capacitor). However the sparking is still an issue, and the components or wires begin to fail. Should’ve placed a spark-suppression cap across the contacts! This is where it started to go downhill.

And the plastic on my $1.90 test hook multimeter leads began to melt:

Not good, but still curious to see if the AC voltage can be stepped up.

Transforming AC

Why not test with a beefy transformer, see if we can step up this AC voltage?

Not captured on camera, but I was able to step up the voltage to about 100 mV (with a lower input voltage). The relay or other components at this point are probably irreparably damaged. Sometimes it would stop oscillating after being plugged in for some time. Unsoldered the diode, the capacitor; repeated. Got it working again momentarily but still too finicky.

This is not a good circuit.

Indicator LED

Just for fun, as I had a bicolor surface-mount LED already on this board, and because the output voltage was unsatisfying low, I wired it up to the AC output, the two LEDs back-to-back with the hope that you could see it alternate red/green as the AC switches direction:

Half-successful: captured the green LED emitting light when the circuit was powered up, but then not thereafter (surprisingly, the LED didn’t burn out).

More investigation would be required to diagnose and solve these problems.

Conclusions

Here’s the final circuit I ended up with:

Overall, this project was a failure. The basic principle of generating AC from DC using an electromechanical relay was demonstrated, but the circuit design was flawed and/or the components stressed beyond their rated limits, before I could successfully connect the AC to a step-up transformer, then (optionally) convert back to higher-voltage DC.

Electromechanical choppers are largely obsolete today, probably for good reason. Or if you need one, likely better off using an all-in-one device which has all the components built to work together, or make sure you know what you are doing. This is always good advice. Oh well, no big loss, you can’t win them all, live and learn. Now I know what not to do, and if I was to build a DC-to-AC converter in the future, I’d consider going with a transistor-based design (project #91) or learn more before building. If anyone has hints to how to improve the mechanical circuit to make it functional and safe, I would be interested. But for now, I am moving on to other projects.

Anyways, to conclude I’ll leave you with this real-world Russian chopper circuit schematic from ChipDipvideo: Electromechanical chopper: