10kV Cockcroft-Walton Multiplier 230VAC 50 Hz 325VEpk 0.5nF caps safety issues

[eneuro]

I my anty mosquitoes project I need to create 10kV (10mm/1cm/0.010m) sparks 
Quick calcualtion from this Cockroft-Walton Optimum Design Guide shows that I need about 16 stages of Cockcroft-Walton Multiplier directly from power grid 230VAC (325Vmax) 50Hz .
While capacitors rated at >650V needed I'd like to put 2x 1nF in series, so I will have two 0.5nF caps and two diodes in each Cockcroft-Walton  multiplier stage which leads to about 2mJ (0.002 J) of stored energy in all caps.
Equation of output voltage from mentioned guide is:

Eout:float(2*n*Epk-Il*(-n+3*n^2+4*n^3)/(6*f*C))

While n=16, f=50 Hz, C= 0.5nF, Epk=325V it looks like when we take load of 300 Ohm (human body top to feet resistance when skin 1000-100k resistance are broken) it is not possible to get on its output even average 1mA load ?
When Il=1nA then Eout is 10.2kV.
Simulation with such 230VAC 6.5kV 10 stages Cockcroft-Walton Multiplier and bigger 1nF caps shows only 30mV 50Hz voltage pulses on 300 Ohm load 100uA in peaks.

But is it really safe to touch its output and keep while there should be abput 10kV without any load?
However, frequency is quite low and caps are small, so fc is in the order of: 25*10^(-9) so output current under <1000 Ohm might should be very low too-but is it safe?

1 case) We have 10kV on the output and someone touches it in quite dry environment, so lets assume 1k skin resistance
2 case) We have 10kV on the output and someone touches it and keeps handle its output, but in wet enevironment and lets assume only 300 Ohm internal human body resistance
Could you help find safety standards which could apply to such situation: 10kV-15kV Cockcroft-Walton Multiplier 230VAC 50 Hz 325VEpk 32x0.5nF=16nF@650V caps, please?

[c4757p]

If I were you, I would concentrate more on "how do I keep people from ever touching this dangerous thing?" rather than "how do I make this dangerous thing not dangerous?". Or even "do I really need to build this dangerous thing?".

DAS ELEKTRISCHE MULTIPLIKATOR IST NICHT FÜR DER GEFINGERPOKEN UND MITTENGRABEN!


What happens when some nastiness (conductive debris, whatever) puts the full 10kV across one of the capacitors, and you blow it short? You're one step closer to a direct connection from human to mains.

[penfold]

IEC 61010 has a section on this.  I don't have a copy to hand but I know it has requirements for the permissible currents available, the peak capacitance allowed etc.

If it is directly coupled from the mains supply then you're basically stuck as the creapage and clearance for any pollution degree will no doubt be prohibitive.  You're options are: guaranteeing that nobody can touch it, or that if somebody touches it there is no way what so ever that they will receive a lethal shock - that includes due to any component failure within reason.  The best option in my opinion would be to include an isolation transformer, that'll solve many problems you're yet to run in to.

[Andy Watson]

You're one step closer to a direct connection from human to mains.

With the wrong connection of the mains polarity there is already a low impedance path through the diode chain!

[eneuro]

You're one step closer to a direct connection from human to mains.

With the wrong connection of the mains polarity there is already a low impedance path through the diode chain!

NOTE: Notice this post later. While we have grid 230VAC do you mean Load (L) exchanged with Neutral (N) ?
On this schematics ground (N) is connected to those diodes chain, but thats is right it might be danger if someone put it reverse, so in this version without lower safe galvanic isolated input voltage it can not be usefull for everyone, but questions about safe output load current levels under normal conditions in this thing still holds 
Of course failure of caapcitors from first left to right could be dangerous too due to this diodes chain...

Touching this thing is not a normal operation mode of this device, so question is what happends by accident in those two cases if someone (human) will have contact with output of this thing, like It is rather easy to predict if someone puts metal into 230VAC wall socket, there it is not such stright calculation, but this which I've found in this guide mentioned above makes sense while those capacitors in series are very small, so at 50Hz it should not be possible to light on even one LED with 1mA for a few seconds.

IEC 61010 has a section on this.  I don't have a copy to hand but I know it has requirements for the permissible currents available, the peak capacitance allowed etc.

In this thread At what capacitance do HV capacitors become dangerous to human touch? at the end this IEC 61010 is mentioned:

No values are given for single fault condition but acceptable capacitance for 200V is ca 8.5uF, 1kV is ca 680nF, 10kV ca 18nF and 40kV 2nF.

It is not clear what this 18nF there means- at 10kV it has 900mJ, while in my case I have only about 2mJ in the case this thing is not connected to mains and someone touches. Still it is below 18nF while i have

16*2*0.5nF= 16nF

capacitance and each capacitor is under voltage of 650Vmax, not 10kV.
So, it is a little bit confsusing, but it looks like it fits easilly into  this norm 

Second case, where someone touches this thing while it is still connected to mains 230VAC is more difficult, while we need calculate output current from this thing under one of the worst case 300 Ohm load, while someone keeps it touching (by accident) and this thing is charged with 50 Hz frequency 230VAC.
From calcualtions I've made this current should be magnitudes  below 1mA and no shock visible in human body and this waht I'd like to ensure for the moment.
So, the main concern is to validate this calculation of average current under 300-1000 Ohm load, while it should drop very quickly from 10kV to very low voltage and very tiny current at such load.

Failure of capacitors or diodes are other cases and it does not matter when we want to know if output current in normal conditions will be safe or not assuming 50 Hz and 0.5nF capacitances of components in this 10kV Cockroft-Walton HV generator...

BTW: I've found this IEC 61010 3rd edition presentation so we'll try to figure it out and confirm or dismiss:

[Neilm]

IEC61010 has a couple of curves on a graph (Either fig 7 or fig 8 ) to show the allowed capacitances for both normal and single fault conditions. Going from memory, the 18nF is the single fault permissible level. For higher voltages, the total charge is specified. I can't remember what that is.

[Marco]

Why not use a trigger transformer like normal bug zappers?

[penfold]

The standard makes it pretty clear.  It is the voltage to which there is potential exposure measured with respect to earth and the capacitance to earth from that point.

The rules are different if its mains coupled and they are basically, don't do it and don't argue with it.

There are allowances for the lack of isolation from mains voltages for instance regarding EMI controlling capacitors that are often connected from the 0V side of an isolated power supply to the neutral, regulations are more adapted for this purpose and don't lend themselves to what you're doing.

I know the fun in taking risks and do it often but only when I'm quite satisfied that nobody else can be hurt by it, there are too many uncertainties in this world to go round betting other people's lives on matters.

[eneuro]

In this quick MATLabs overview of  IEC 61010 3rd Edition there were something about 0.5mA rms or 2maVDC considered as save currents.
While those Cockcroft-Walton multiplier at low frequencies and at low capacitances in many simulations proved to be very low output current under any even small load, first shoot was try to estimate those currents and see if they can be below 1mA not only when charged, but also when still connected to power source and it looks like they should be far below 1mA.
Simply in normal conditions those small caps and low frequency limit nice possibble output RMS current as expected below 1mA even at 10kV.

However, in the case of direct using of 230VAC mains grid power there are clear issues with reverse connection to Load/Neutral power supply and in the case of failure of some caps we can have really 325Vmax-32*0.5V on diodes chain drop which is not nice and of course it has to be changed or different approach to hunt for mosquitos.
Blue laser uses of consumer electronics with like missle tracker systems from Star Wars to hit mosquitos was shown on TED 4 years ago, so I'm looking for something much more simple 

Using latest NVIDIA GPUs is also another possible solution even to recognize female mosquitos, but... it require a lot of power ! 

It looks like it is cheaper kill all mosquitos, while I do not care which of them makes this enoying noise 

Inductors are also commonly used in place of a trigger transformer, however are not considered transformers themselves, although similar in operation.
So, maybe those things can be usable.
When I had cheap ceramic capacitors rated at 1kV-10kV 1nF for example than if i had in pulse inductor/ trigger transformer pulse close to 1kV, than it is even easier to get 10kV and I can controll pulses frequency as well as rms output current under any load, so it might be something worth to try   
However, we still need to ensure at 1kV or 10kV that amount of charge is all the time at safe level, so those calculations might have common sense if they are corrent...

[Marco]

I wonder why no one has tried to build a copy of that laser photonic fence yet, it doesn't look that hard.

[eneuro]

it doesn't look that hard.

http://en.wikipedia.org/wiki/Mosquito_laser
Yep, photonic fence idea is quite interesting, while using infra-red light-emitting diode (LED) lamps on a fence post to create a field of light 

"This non-lethal laser is used to determine the size of the insect, and the frequency at which its wings are beating."

This looks like overcomplication and it should be simplified

Maybe this is the main reason they have only working prototypes?

"The idea has been criticized because most areas where malaria is prevalent do not have reliable electrical power"

It works in the range of 30meters, so not so bad, but I think it can be done in diffrent more efficient way by scanning much bigger volume and maybe use photonic fence to select insects to kill, but... catch many of them per second.
They used laser, so I wonna use this particle accelerator technology, so safe HV needed 

but comercialy available 3V hand bug zappers has  sometimes 3 capacitors and 3 1N4007 1000V diodes only after I guess trigger transformer, which creates quite high voltage, so it is not to difficult to do something like this 

[Marco]

The problem is that against a normal background they are very hard to detect. They have retroreflectors on the poles, this will give you far higher contrast.

Also hugely reduces the number of pixels you need, which with some sensors allows you to use high fps.

You don't really need to measure the wing beat frequency, but they want to minimize ecological impact. It's not that hard either, you can meausure doppler shift of a modulating signal.

[eneuro]

They have retroreflectors on the poles, this will give you far higher contrast.

It is interesting and definitely I need to know my "enemy" better to be more efective on "battle field"  http://animals.howstuffworks.com/insects/mosquito.htm
However, for the moment I'm more interested if moscuitos have... a charge, but while they are flying in the air by its moveing parts they might get charged somehow or hopefully I will be able to induce charge, so I think is time to make some experiemnts and design prototype moscuitos accelarator.
If more power needed, no problem I have 2kV transformer from microwave oven teardown

[amyk]

10kV (10mm/1cm/0.010m) sparks

The normal dielectric strength of regular air at sea-level atmospheric pressure is ~3kV/mm, you'll need closer to 30kV (or more) to reliably produce 10mm sparks.

[eneuro]

you'll need closer to 30kV (or more) to reliably produce 10mm sparks.

Yep, It is about 3000kV/m, so while mosquitos size is between 0.003m-0.020m & average weight 2.5*10^(-6) kg (2.5 mg), so maybe it will be much more safe to use wiki: Van de Graaff generator which will allow create much longer sparks 

"Since a Van de Graaff generator can supply the same small current at almost any level of electrical potential, it is an example of a nearly ideal current source. The maximum achievable potential is approximately equal to the sphere's radius multiplied by the e-field where corona discharges begin to form within the surrounding gas. For example, a polished spherical electrode 30 cm in diameter immersed in air at STP (which has a breakdown voltage of about 30 kV/cm) could be expected to develop a maximum voltage of about 450 kV."

450kV gives about 0.015m (15cm) sparks so not so bad 
It looks like it is quite safe device in educational versions if no additional capacitances are connected in pararell.

However, there  Van de Graaff generator  safety are described many traps when playing with such HV low charge sources, while some of them can store significant amount of energy if discharged at high rates.

It's interesting that for demonstration purposes with humans  (at  those very high voltages achieved between the sphere and ground, typically in the range 200 kV to 300 kV) the electrostatic energy stored by the sphere should not exceed 0.5 J  according to this publication
When we calculate maximum capacity we have

Cmax= 2E/V^2

, so it gives at 300kV 11 pF (0.0111 nF) only !!!

Hopefully, 0.5 J energy means that it could accelerate mosquito to the speed of.... 630 [m/s] close to Mach 2