My teenage son has been building Cockcroft-Walton multipliers in his spare time lately, and I'm potentially a little concerned in case he electrocutes himself. His circuit is powered by 4 AA cells in series (6V), and judging by the spark gap he's managed to get to around 6kV using an oscillator and transformer followed by the Cockcroft-Walton cascade. Since he's building this on a simple breadboard there is no insulation between him and parts of the circuit, and he's already told me that he's had a few shocks.

I'm a physicist working in a university, so I looked up the regulations regarding home appliances in Europe and I think it should be ok (we use the same regulations to assess Van der Graaf generator demos). The total capacitance of the cascade circuit ends up at around 1nF, so at 6000V the total charge should be 6uC which I believe should be safe (EN-60335-1) but as I'm not an engineer I can't say for sure.

My question is: should I stop my son from building these? I have always encouraged creativity, and his love for electronics but I don't want him getting injured or worse.

• If he's good at knitting, perhaps you can advice him to knit a grounded shirt of copper wire? (It should work as a faradays cage, but I have obviously not tested the safety of this.) – Oskar Skog Apr 30 '17 at 8:33
• If it's battery powered, and can be held in one hand (which limits the volume of caps and so stored energy (more crude than 60335-1, but easier to apply)) then it's usually fairly safe. The real danger begins with connection to mains. Don't get him to knit copper wire shirts, that's a solution to a different problem. – Neil_UK Apr 30 '17 at 8:58
• Thank you, I had the same impression about the use of batteries making it safe, and he tells me the transformer plus voltage cascade makes the current very low on top of that. – HGros Apr 30 '17 at 9:37
• people have died from tasers (small one-handed battery-operated devices) – dandavis May 1 '17 at 1:26

Getting shocked is good. Dropping solder on clothing and skin is good. Nicking the fingers on sharp metal edges is good. Humans survive by exploring. Humans learn from pain. Otherwise we'll be cowering in the swamps.

As a kid, I felt 117 VAC tingles through the toes, from old power transformers. I learned to sit on wooden chairs and not touch the toes to the concrete floor.

Later while "calibrating" an oscilloscope, I pushed the bared-back of the scope against the metal lab bench, touching the EICO scope tube socket to the bench, and then leaned forward with tummy against the front part of bench through my shirt, and touched my chest again through the shirt to the scope chassis as I reached far around to adjust a "focus" potentiometer. 3,000 volts across the chest. I sat, stunned, for a couple minutes.

But I had a few more lesson about high voltage still to learn.

Let your son see some high-voltage-death videos.

Teach him the "keep one hand in pocket" trick around high voltage.

EDIT: Then there is high current; college prof told of buddy who lost the left ring finger, because the wedding ring ended up in a high current path, making the ring glow red, killing the skin, the muscle, tendons, and the bone.

I've also had MOSFET gate driver ICs blow off the top of the package, during the "bipolar snapback" event when the 1,000 μF storage capacitors inside a huge HP lab supply had the need to unload their energy into the 2 mm × 4 mm silicon of the gate driver. None of us three, hovering quite close, were hit. But after that, I always placed a sheet of paper atop the circuit, to intercept any more IC energy dumps. Energy? 1/2 * C * V^2 = 0.5 * 1,000uf assumed (did not open the HP supply) * 20v * 20v = 200 milliJoules which explains why the DIP plastic top was blown off. And missed our 6 eyes (tho I wore glasses).

EDIT: The gate driver blow-top-off was serendipity, because I took the lesson to heart and realized the danger of stored-energy in 1,000 μF caps. I learned how to tease-the-dragon in evaluating bipolar-snapback, allowing only 1,000 pF right across the Gate Driver, with 220 Ω resistor to the (experimentally variable) Vdd. Using a long-leaded 1,000 pF (3″ leads, 6″ total, or 100 nanohenry) along with the external 1,000 pF and the on-chip well-substrate of ~1,000 pF, during switching events the silicon VDD_GND would collapse and then rebound 5 or 10 or 15 volts above the rated 18 volts. At some level, the slewrate of the ringing (100 nH and 500 pF rings at 22 MHz) induced enough transient charge into the silicon that bipolar-snapback occurred, and the VDD (supplied by 1,000 pF) would be sucked down to 16 or 17 volts whereupon the snapback would self-extinguish. I ran these devices, in/out of snapback at 100 kHz, with no damage, as I diagnosed the transient charge path and realized the layout rules needed change. Serendipity. Energy? 0.5 * C * V^2 = 0.5 * {total protoboard + silicon Cap = 2,000pF} * 31.6volts^2 = 1,000pF * 1,000 (volts^2) = 1microJoule.

Decades back, returning from lunch, was told to go to lab and examine "the debris" on XXXX's bench. There was 6panel wirewrap board (30*6=180 ICs), many ICs with their tops blown off. Turns out a hanging one-end-loose wire had curled over and around and under the front bench edge and **INTO* the hot contact of 117VAC power. Thus management wanted all engineers and techs and rework folk to understand the danger of springy-curly wirewrap wires left hanging.

Ahhhh Assigned to 400-watt Tritek switching supply for couple weeks, for some reason. Just to give me experience in switchers; I was not the designer. Repeatedly, the sacrificial 5-watt 5 ohm wirewound protection resistors --- were exploded, their ceramic cores blasted out of the heatsinkable case and across the walkway between benches, the resistive wire trailing behind like guidance wire for a TOW missile. We learned to not stand in the way.

For safety, and no hum in highgain amplifiers (100dB and 120dB), I learned to use the 9volt "B" 3" by 3" by 4" batteries. The high Rout caused oscillations, almost all the time, until I learned to implement "local batteries" with RC LPF in the VDD to LNA stages. Quite a collection of 5,000uF caps I had.

• been there, (twitch)...done all those (shudder). You've missed the exploding 4.7uf/16v capacitor reverse biased splattering electrolyte into your face (tinnitus). The lessons don't stick until they happen to you, and square that for kids. – glen_geek Apr 30 '17 at 16:05
• +1 for (thoughtfully selected) high-voltage-death videos, perhaps a mention of the few real loss-of-life incidents related to reproducing youtube experiments that have been documented. It's great to experience these things, but there are unfortunate flukes, and not all of them are voltmeters. – uhoh May 1 '17 at 1:22
• +1 LOL I had to laugh, been there done that, then did it again just for good measure. I'm pretty sure we have all had those.. "oops.. that was a bad idea" moments. And some of us have a some scars to prove it. – Trevor_G May 1 '17 at 16:20
• Also pretty sure most EEs have a distinct lack of finger-print and feeling on the tips of their fingers.. – Trevor_G May 1 '17 at 16:23
• 1200 VDC hand to hand. Only once. | RF from various. Too many 230 VAC bites. Surprisingly, 50 VDC on a wiring frame across back of hand is annoying on a very humid day. Telephone ringing is not nice. Uniselector stepping inductive kick is not nice. || Fewer and fewer exponentially with increasing age :-). || People have died with 12 VDC across chest in exceptional circumstances. || 12VDC on flounder fishing spear LED lamp caused user msucle lockup (a friend). | Arc welding (of sorts) from a 250 W 30V PV panel was a surprise. More .... :-) – Russell McMahon May 3 '17 at 4:12

Looks safe from what you describe, as long as he only uses batteries and keeps the capacitors small. CW multipliers increase the voltage but reduce the current so there will only be a few hundred microamps at the output end.

EN60335-1 suggests that under 15kV so long as the total charge of a shock is less than 45 microcolumbs there shouldn't be danger. Your son's circuit looks like it is way below that as Q=CV. Obviously if he starts going to higher and higher voltages he'll need to reduce the size of the caps to stay safe. With 6000V and 1nF any "shock" will feel a lot like a static shock from a door-handle. It's also a similar sort of output that commercial cattle prods have.

Another property of CW cascades is that the output voltage and current depends on the load: the lower the resistance of the load the lower the current goes, which makes them really inefficient but can save your bacon too if you were to become attached to it.

I also agree that he should be supervised, I think that almost goes without saying.

I think the only way for 4 AA (or D etc) batteries to be dangerous in a circuit would be if a circuit like the above was used to charge a huge capacitor. I could be wrong on that though.

• +1 but do cattle prods give the same "user experience" as a static shock from a door handle? – uhoh May 1 '17 at 1:24
• worstcase doorknob sparks: 150pF and 15KV = 34 millijoules. See lots at web.archive.org/web/20070210193125/http://www.jci.co.uk:80/… Highest ever measured was over 50KV, that's in northshore Alaska winter conditions, after having slid across the long bench seat in a truck. – wbeaty May 1 '17 at 6:28

Safe is a relative term, what is safe for one person may not be for another, and as a professional I can't really tell you he is 100% safe. It does sound like the energies involved are fairly trivial, but that does not mean he won't hook that up to a power transformer an hour from now to get a bigger spark. Also, even with these values he could over-volt a capacitor and cause it to fail rather violently. Safety glasses would be a good idea for this, and other projects.

Working with electricity always has risks associated with it. They can be from electrocution, explosion, burns, starting a fire, chemical exposure and a few others. That just comes with the nature of the job.

Should you stop him? Well you could try, but you may be better to have him properly educated on the risks and safety measures he should be taking to limit his vulnerability to something going wrong. Those measures should include a strict restriction that "We don't play with electricity when nobody else is around!".

There are a number of simple, and complex guidelines on-line.

Perhaps spend some time with him while he is experimenting. You may enjoy it, and I am sure your son would appreciate it.

ADDITION: As a parent though, I would be taking measures to ensure that where the lad is working is properly equipped. The right tools, equipment, ground fault interrupted power outlets, lighting, ventilation are all important.

DANGER! The heartstopping-ness of cap-discharge is determined by energy, not by coulombs.

In general, it's bad to let your capacitors get into the 10-joule range and above. That's for discharges across your chest, of course. Significant heart danger starts at 20joule discharges. Below 10joules, the main problem is muscle contractions, getting sliced by slamming sharp objects etc.

0.001uF and 6KV gives 36 millijoules. Pretty safe, if a bit painful.

Still, the cardiac effects depend on energy density, not just joules. If you stabbed a sharp-terminal capacitor into your ribcage, the energy delivered to the pacemaking system would be orders of magnitude higher than if the same capacitor terminals were touched with two hands.

When working with cap-discharge systems, only use one hand. That way the accidental discharges won't go across your chest. Or better yet, always remain very fearful, and paranoid that you'll make a mistake and receive a bad zap. Some serious respect (if not sheer terror) goes far in encouraging proper research beforehand, and to avoid developing any ignorant risky behaviors when dealing with kilovolt capacitors.

• Interesting. Do you know why the UK regulations 60335-1 say that below 15kV the charge must be less than 45uC and above 15kV the energy must be less than 350mJ? This determines whether the output is considered live, but I guess you can go a little higher and still be safe (I haven't read a solid number in the regulations). – Kurt Newman May 1 '17 at 7:07
• Wow. 10J is a lot of energy, I'm very surprised to read that it could be safe. – HGros May 1 '17 at 9:06
• +1 for "Some serious respect (if not sheer terror) goes far in encouraging proper research beforehand" – Trevor_G May 1 '17 at 16:14
• @HGros not "safe," just non-fibrillating. Didn't that paper say that injuries start at 1J, from extreme muscle contractions? – wbeaty May 1 '17 at 18:34

There were a few hints about "power it from a battery, period". There is a good reason not to do that with experiments where a few kilovolts are generated even if a normally safe mains-connected power supply is used, and it has not really been mentioned here.

High voltages in an amount (maximum sustained current, stored energy) that normally couldn't do much permanent damage to you still brings at least these hazards:

• it can break down the primary-to-secondary insulation in even a sturdy power supply - in the worst case permanently, so things that aren't supposed to be directly mains connected now are. Even a power supply built perfectly to safety standards is likely to have an insulation failure if you manage to put more than approximately 3-4kV against earth into any of its terminals by mistake

• High voltage strikes arcs really easily. If you manage to strike an arc against anything that has live mains on it (could be an imperfectly insulated connector near your setup, like a cord cap not being inserted into a socket all the way, or something exposed through a ventilation hole in a power supply case...), this arc is now a conductor perfectly capable of conducting anything connected to it (if you are lucky, the arc only lasts until there is a zero crossing by the mains AC.).