How to Safely Generate High Voltage for Capacitor Testing?

[bob012345]

I need a relatively high voltage less than 10k but essentially zero current on the order of 0.1 micro amps and very low power due to leakage on the order of 200 micro watts. These are rough numbers. My question is can I boost a 9V battery to provide the voltage and will it still be dangerous if the current and power as limited to those max ranges? The application is testing a capacitor. The actual target numbers are around 3500V, 0.06 micro amps and 200micro watts .

Is there a fundamentally safe way to do this? Thanks.

 

[Baluncore]

What is the capacitance value of the capacitor?

 

[bob012345]

It's a parallel plate capacitor about 0.01 square meters and a separation of about 1cm and dielectric is air so about 5 pF is my guess.

 

[Nugatory]

Is there a fundamentally safe way to do this? Thanks.

Are you happy with a pulse as opposed to steady flow? If so a coil and something to break the circuit - think breaker points on an old pre-electronic auto ignition - will do the trick. You will definitely notice the shock if the voltage goes astray and it will be unpleasant enough to discourage repetition, but about as safe as these voltages can be.

 

[Baluncore]

It will take 0.25 seconds, to charge 5pF, to 5kV, with an 0.1uA current source.

Is there a fundamentally safe way to do this? Thanks.

No.
5kV is fundamentally unsafe, without discipline and safe procedures.

You might build a Cockcroft–Walton voltage multiplier using low value capacitors to limit the current.
https://en.wikipedia.org/wiki/Cockcroft–Walton_generator#Operation

You would need to drive the multiplier with a 250 Vpp high-voltage oscillator at about 1 kHz, powered by the 9V battery.

 

[bob012345]

Are you happy with a pulse as opposed to steady flow? If so a coil and something to break the circuit - think breaker points on an old pre-electronic auto ignition - will do the trick. You will definitely notice the shock if the voltage goes astray and it will be unpleasant enough to discourage repetition, but about as safe as these voltages can be.

I need to keep the voltage constant for a long time (several minutes).

 

[bob012345]

It will take 0.25 seconds, to charge 5pF, to 5kV, with an 0.1uA current source.


No.
5kV is fundamentally unsafe, without discipline and safe procedures.

You might build a Cockcroft–Walton voltage multiplier using low value capacitors to limit the current.
https://en.wikipedia.org/wiki/Cockcroft–Walton_generator#Operation

You would need to drive the multiplier with a 250 Vpp high-voltage oscillator at about 1 kHz, powered by the 9V battery.

Not sure I understand. Does this mean I need two voltage sources?

 

[OmCheeto]

Rub a balloon on your head? Just a guess. Have no idea what voltage it produces.

 

[bob012345]

Rub a balloon on your head? Just a guess. Have no idea what voltage it produces.

Probably not enough to get through my thick skull.

 

[Baluncore]

Not sure I understand. Does this mean I need two voltage sources?

Your 9V battery runs an AC oscillator like used in a xenon photo flash tube.
The CW multiplies that AC and produces the EHT DC output voltage.

 

[bob012345]

Your 9V battery runs an AC oscillator like used in a xenon photo flash tube.
The CW multiplies that AC and produces the EHT DC output voltage.

Ok, thanks. I'll look it up and run it in LTSpice. Thanks again.

 

[DaveE]

I'm voting no, not safe. Mostly because you have to ask (which is great BTW).

Working with dangerous circuits is often presented as a circuit question, but it's usually way more complicated that what you intend. There are a lot of secondary considerations that can be hard to control or define. It's more about worst case conditions, failure modes and effects, and most of all, safe working practices.

One of the scariest things I ever did was to measure the impedance of a 1KW fast flow CO₂ gas laser, while it was operating, with a frequency response analyzer. This laser had a 10KV, 10KW linear DC power supply. I literally spent a day at my desk, and just looking at the laser construction, thinking about how to do this thing safely and what to do if things go wrong. Then I spent about another day building special purpose test gear to make connections, limit voltages in fault conditions, emergency kill switches, plexiglass barriers, etc. Then the test took about 1 hour, but I was scared for about 2 days in total. The opposite of "hold my beer"; fear is maybe the most important safety tool.

BTW, the impedance of ionized gasses is a really interesting thing, I wish I had those old lab notebooks. But alas, I couldn't finish that work for business reasons, the project was killed.

Also, (off topic) funny story. The product was called "EVERLASE", it was pretty awful; very unreliable. One time my boss had to go to a customer site that had had several failures. They had painted an "N" in front of the name on the side and called it the "NEVERLASE"... ouch, they were kind of right.

 

[Baluncore]

I'm voting no, not safe. Mostly because you have to ask (which is great BTW).

Powered by a 9V battery.
A 1W single transistor oscillator.
Current limited by a CW multiplier.

For $20, eBay will sell you a "DC-12V High-Voltage Electrostatic Generator Negative Ion Generator" that produces 4.3 kV DC output.
Or, for $15, "DC 12V High Output DIY Air Ionizer Airborne Negative Ion Generator" that will produce 5 kV.

 

[bob012345]

The thing is, how does one go from not an expert to a competent researcher wrt high voltage? Everything I want to try requires something dangerous and I am a very cautious person by nature. Where I experiment has safety people and HV experts so everything I do will be under some kind of supervision with lots of advice and input. I agree there is no completely safe option using HV but I just thought using 9V battery source might be a bit safer that a 3kV plug in power source. Thanks for the advice!

 

[DaveE]

The thing is, how does one go from not an expert to a competent researcher wrt high voltage? Everything I want to try requires something dangerous and I am a very cautious person by nature. Where I experiment has safety people and HV experts so everything I do will be under some kind of supervision with lots of advice and input. I agree there is no completely safe option using HV but I just thought using 9V battery source might be a bit safer that a 3kV plug in power source. Thanks for the advice!

If you are cautious and thoughtful, have enough experience in EE to understand the risks, and have access to people that you trust to ask about this stuff, then I think you'll be OK. You are absolutely on the right track asking these questions, most people don't. It's not the HV PS that kills you, it's what you do with it.

Most of these sort of questions are similar to "Should I buy my teenage boy a (real) bow and arrow set?"* It's not just the bow and arrow that we need to know about. So my generic answer is usually no, unless you provide the other ancillary information.

* or, here in the US, an AR-15.

 

[DaveE]

Powered by a 9V battery.
A 1W single transistor oscillator.
Current limited by a CW multiplier.

For $20, eBay will sell you a "DC-12V High-Voltage Electrostatic Generator Negative Ion Generator" that produces 4.3 kV DC output.
Or, for $15, "DC 12V High Output DIY Air Ionizer Airborne Negative Ion Generator" that will produce 5 kV.

Yes, it can be made or bought and be safe. But this is in a development lab. It's not necessarily the final product specs that I worry about. It's what EEs tend to do while they are working on the prototypes that's the concern.

People always ask, I'm going to build this thing, will it be safe? Perhaps my answer should be probably, if you don't kill yourself doing it.

I will also confess that as a pro EE for decades, I'm suspicious of tinkerer's that are attracted to HV like bugs to a porch light (not necessarily this OP). Why do you need HV? Why not learn about cool stuff at 5V, or 50V? The pros use HV because we have to, and most wish they didn't, it's a PITA. Again, a bias towards stay away, unless you provide further info.

 

[berkeman]

The application is testing a capacitor.

What kind of test?

It's a parallel plate capacitor about 0.01 square meters and a separation of about 1cm and dielectric is air so about 5 pF is my guess.

8.85pF is air dielectric. What is the dielectric? What is the tolerance of that separation number?

How are you going to monitor that 3500Vdc value that is applied to the capacitor? What is the input impedance of your measuring circuit (and hence the leakage current that you need to support)?

 

[bob012345]

What kind of test?


8.85pF is air dielectric. What is the dielectric? What is the tolerance of that separation number?

How are you going to monitor that 3500Vdc value that is applied to the capacitor? What is the input impedance of your measuring circuit (and hence the leakage current that you need to support)?

It's testing concepts and properties of homemade capacitors. The plates will be copper tape. Initially, air is the dielectric with the gap in the 1-2 cm range but ultimately I want the thickness of dielectric thin films like Kapton or Teflon. The whole cap will be sealed in styrofoam as an insulator. The measurement questions are questions I am working on getting answers to.

 

[DaveE]

The whole cap will be sealed in styrofoam as an insulator

Styrofoam??? Really? I'm not an expert, but I've never heard of that being used as a HV (electrical) insulator; except for cables, maybe? It's highly flammable and not really dimensionally stable, because, you know, foam. I would normally see some epoxy like coating/impregnation in this sort of application. But, OK do what you will. I don't know enough to say it's wrong, it just seems unusual. Someone like 3M must sell a version for this, there's about a million choices. Choose well.

If you need safety agency approvals, they will definitely ask about this.

edit: Scratch that last comment, maybe the whole reply, I forgot about the "homemade" part. I was thinking about a different world.

 

[Rive]

I literally spent a day at my desk, and just looking at the laser construction, thinking about how to do this thing safely and what to do if things go wrong. Then I spent about another day building special purpose test gear to make connections, limit voltages in fault conditions, emergency kill switches, plexiglass barriers, etc. Then the test took about 1 hour, but I was scared for about 2 days in total. The opposite of "hold my beer"; fear is maybe the most important safety tool.

Done right.

Is there a fundamentally safe way to do this? Thanks.

Based on the displayed knowledge, you better not exceed the limitations of a Van der Graaf generator: the kind used in school experiments.
That will give you plenty of references about the safety measures too.

 

[bob012345]

Styrofoam??? Really? I'm not an expert, but I've never heard of that being used as a HV (electrical) insulator; except for cables, maybe? It's highly flammable and not really dimensionally stable, because, you know, foam. I would normally see some epoxy like coating/impregnation in this sort of application. But, OK do what you will. I don't know enough to say it's wrong, it just seems unusual. Someone like 3M must sell a version for this, there's about a million choices. Choose well.

If you need safety agency approvals, they will definitely ask about this.

edit: Scratch that last comment, maybe the whole reply, I forgot about the "homemade" part. I was thinking about a different world.

Thanks. I hadn’t thought of the styrofoam igniting. Now I’ll be ready just in case…

 

[berkeman]

It's testing concepts and properties of homemade capacitors.

Then why do you need to use a high voltage? That is only used in breakdown testing, which would seem to be irrelevant to simple testing of a capacitor.

 

[bob012345]

It

Then why do you need to use a high voltage? That is only used in breakdown testing, which would seem to be irrelevant to simple testing of a capacitor.

It is related to breakdown testing. I want the highest field strength without breakdown.

 

[Averagesupernova]

Are you prepared to control the humidity for these tests? Breakdown will vary based on humidity.

 

[bob012345]

Are you prepared to control the humidity for these tests? Breakdown will vary based on humidity.

Not control but be aware of. It’s not going to be an academic scientific study as much as a hobbyist project.

 

[berkeman]

I'm still curious about this:

How are you going to monitor that 3500Vdc value that is applied to the capacitor? What is the input impedance of your measuring circuit (and hence the leakage current that you need to support)?

 

[bob012345]

I'm still curious about this:

How are you going to monitor that 3500Vdc value that is applied to the capacitor? What is the input impedance of your measuring circuit (and hence the leakage current that you need to support)?

The plan is to design the power source such that it is so current limited and has safety features ensuring the voltage does not change even if the cap breaks down. I would think the cap would also have some leakage current aside from any measuring circuits?
I don’t know yet how the measurement would be incorporated into the setup. If you or anyone can point me to a book, a course or some source of information on HV systems that might cover such topics I’d appreciate that.

 

[DaveE]

The plan is to design the power source such that it is so current limited and has safety features ensuring the voltage does not change even if the cap breaks down.

This makes no sense. That's not how arcs work. If it's current limited the voltage must drop. You need to do some more thinking about this.

edit: oops, that's not really the right link. But there's lots of stuff on the web about arc discharge I-V curves.

 

[bob012345]

This makes no sense. That's not how arcs work. If it's current limited the voltage must drop. You need to do some more thinking about this.

I’m not studying or trying to measure arcs. I’m trying to avoid them. Obviously, if the cap discharges, its voltage drops but not that should not force the power supply to collapse. The current limit I meant is on the power supply so it doesn’t draw a lot of current because the subject of this thread is building a somewhat safe power supply.

 

[berkeman]

I don’t know yet how the measurement would be incorporated into the setup. If you or anyone can point me to a book, a course or some source of information on HV systems that might cover such topics I’d appreciate that.

You plug one of these into your DVM:

I don't remember the probe's input resistance, but if it is $10M\Omega$ then it will draw $350\mu A ~at~ 3500V$

Obviously, if the cap discharges, its voltage drops but not that should not force the power supply to collapse.

As @DaveE says, that is not how a current-limited power supply works. An arc looks like a short circuit to that power supply, which causes the output voltage to drop out. Otherwise, the supply is not "current limited"...