Making a high voltage generator a little safer

[Psi]

I found my old high voltage generator the other day hidden away in a box i thought i'd lost and i managed to get it working again.
I'm in the process of building it into a proper case at the moment.

Trouble is i'm not sure how dangerous it is since i don't have any way to measure current at the 50kV output or the ~16kV flyback and im reluctant to let anyone else near it for that reason.

I'd quite like to make it safer so I was thinking of adding a resistor in series with the high voltage output to limit the current. 50Mohm should limit 50kV to 1mA which should be perfectly safe. However I thought i'd ask if any of you guys have experience with high voltage like this.

I don't want to make it safe so people "can" touch it, i just want to make it safe in case someone "does" touch it. I will still be stopping people from getting too close.

So are there any issues I should be aware of when current limiting high voltage?

Obviously the 50Mohm resistor needs to be designed to withstand 50kV.

[Time]

You can measure the current pretty easily if you have time.  What you need to make is a compensated rogowski coil.  You might be able to just buy an industrial current transformer somewhere too.  But a homemade rogowski coil is simple in construction, especially with the mild frequencies that flyback transformers operate with.

As far implementing safety, current limiting is your best bet.  From what I can tell its just transformers and no power conditioning which might contain capacitors which are deadly.

50 kV is pretty serious.  Be careful.

[Neilm]

I do design work with high voltages reasonably often, so I have some experiance of design at these voltages.

When dealing with high voltages, remember what an insulator is : "An insulator is a poor conductor".  It is entirely possible for it to conduct along a 20mm piece of plastic with significant current flow. (Significant at least if it is going into some measurement circuits)

Even with a resistor as low as 50M you can see effects of creepage (leakage) over the surface. Also - check the power rating of the resistor - 1mA through a 50M resistor is 50W. A high voltage resistor that can handle that much power will be VERY expensive. I would suggest several resistors in a series - parrallel configuration. This might be fiddly if you end up with lots of resistors, but probably less expensive than a single resistor.

You will have to calculate the voltage drop over each resistor to check that it is within the resistor rating. Also, check that the clearance is sufficient, both over the individual resistors and between any adjacent ones. Some resistors are designed to be encapsulated in oil or resin, so they will not reach full voltage before it breaks down over the resistor body.

A search on the net for the breakdown voltage of air will give you several values - take them with a pinch of salt. These will have been generated from known electrodes (probably spherical) and in specific air humidity and pressure. A piece of plastic might not be sufficient protection if it is too close - if in doubt keep clear. 50kV will jump further than you expect.

Neil

[Zero999]

I can tell you that the current will be limited to a very low value as I can see that the transformer has a very high leakage inductance: look at the coil spacing and the massively oversized secondary.

I'd recommend measuring the short circuit current, take a fairly low value, power resistor  resistor (say 100R, 5W), connected it across the secondary of the transformer and look at the voltage waveform on a 'scope. It'll probably be around 1V RMS which indicates a current of 10mA and presents a low risk of a severe shock, especially at the high frequency generated by a flyback.

Note, don't short circuit it for too long and keep an eye on the temperate of the driver transistors.

How are you powering it? If it's from the mains you won't find a power supply capable of providing sufficient isolation at this voltage, the danger is, if one side of the HV is connected to earth it will arc across the transformer. If you don't want to earth the DC side, connect both the +V and 0V to earth via  10nF class Y2 capacitors.

[Psi]

I do design work with high voltages reasonably often, so I have some experiance of design at these voltages.
When dealing with high voltages, remember what an insulator is : "An insulator is a poor conductor".  It is entirely possible for it to conduct along a 20mm piece of plastic with significant current flow. (Significant at least if it is going into some measurement circuits)
Even with a resistor as low as 50M you can see effects of creepage (leakage) over the surface. Also - check the power rating of the resistor - 1mA through a 50M resistor is 50W. A high voltage resistor that can handle that much power will be VERY expensive. I would suggest several resistors in a series - parrallel configuration. This might be fiddly if you end up with lots of resistors, but probably less expensive than a single resistor.
You will have to calculate the voltage drop over each resistor to check that it is within the resistor rating. Also, check that the clearance is sufficient, both over the individual resistors and between any adjacent ones. Some resistors are designed to be encapsulated in oil or resin, so they will not reach full voltage before it breaks down over the resistor body.
A search on the net for the breakdown voltage of air will give you several values - take them with a pinch of salt. These will have been generated from known electrodes (probably spherical) and in specific air humidity and pressure. A piece of plastic might not be sufficient protection if it is too close - if in doubt keep clear. 50kV will jump further than you expect.

Neil

Thanks for that

I can tell you that the current will be limited to a very low value as I can see that the transformer has a very high leakage inductance: look at the coil spacing and the massively oversized secondary.

I'd recommend measuring the short circuit current, take a fairly low value, power resistor  resistor (say 100R, 5W), connected it across the secondary of the transformer and look at the voltage waveform on a 'scope. It'll probably be around 1V RMS which indicates a current of 10mA and presents a low risk of a severe shock, especially at the high frequency generated by a flyback.

Note, don't short circuit it for too long and keep an eye on the temperate of the driver transistors.

How are you powering it? If it's from the mains you won't find a power supply capable of providing sufficient isolation at this voltage, the danger is, if one side of the HV is connected to earth it will arc across the transformer. If you don't want to earth the DC side, connect both the +V and 0V to earth via  10nF class Y2 capacitors.

Interesting idea about the 100R short.
Instead of reading the resistor voltage with the scope i might try the coil of a meter through a resistor. That way i can calibrate it before hand to show me the voltage across the resistor.

I only ever power it from batteries, about 16V. I wouldnt want to get it anywhere near any of my power supplies

The driver is currently two transistors 2N3055 and MJ2955 that self oscillate from a 3rd winding on the transformer.
I have a suspicion that they aren't oscillating at the ideal frequency as i've noticed the spark go through a brighter stage as the voltage falls and frequency drifts. I plan to replace them with mosfets and a micro at some point so i can figure out the most efficient switching freq.

[Zero999]

Instead of reading the resistor voltage with the scope i might try the coil of a meter through a resistor. That way i can calibrate it before hand to show me the voltage across the resistor.

That wouldn't work because it's not a true RMS meter.

Don't worry, you won't mess up your 'scope. If you're worried, connect the resistor to it first, then if the resistor blows up (it won't) the voltage is probably too high for the 'scope so you should try again with a lower value resistor. I forgot to say that you shouldn't use a wire wound resistor which will be inductive, go for metal film on a ceramic former which should have a low enough inductance to not make much difference.

The driver is currently two transistors 2N3055 and MJ2955 that self oscillate from a 3rd winding on the transformer.
I have a suspicion that they aren't oscillating at the ideal frequency as i've noticed the spark go through a brighter stage as the voltage falls and frequency drifts. I plan to replace them with mosfets and a micro at some point so i can figure out the most efficient switching freq.

If you go for MOSFETs make sure you protect the gates as the voltage on the feedback winding could be enough to zap them. Alternatively, try more modern, faster, higher gain BJTs, the 2N3055 is old and slow.

[mikeselectricstuff]

Resistors can be problematic at HV as they can  arc over.   The best way is to limit current on  the low-voltage side if possible.

[Zero999]

High voltage capacitors are a good way of limiting the current of AC circuit but i wouldn't recommend them here because: you'll need more than one in series, the frequency is unstable so you won't know the current and there can' be resonance with the transformer's inductance.

I think measuring the short circuit current first is the best approach, then at least you know if it needs current limiting and you can calculate the internal impedance of the supply.

Another option is to measure the leakage inductance of the transformer and output the frequency when short circuited, then you can calculate the short circuit current created by the fundamental frequency but that won't be as good as measuring it.

[kahvikello]

This is very safe high voltage power supply

[richard.cs]

RS stock some HV resistors, from memory they're made by Vishay, 3W, 10kV and available in 100k to 68M. I regularly use series chains of them in plastic tubes. Obviously this reduces heat dissipation a bit so you shouldn't use them at maximum continuous dissipation when in a tube.

I regularly use these resistor chains for current limiting and bleed downs up to 150 kV.