## High Voltage Generation

Invictos - 1-3-2017 at 21:15

Good Evening everyone,

I need to generate about 1.5 million volts for an electrostatics experiment (I need positrons, if you're curious...)

I currently have a 200 KeV Van De Graaff generator at my disposal. I have two options, wondering if y'all would have any input:

1.) This one's the simpler of the two: I buy another 7 Van De Graaff generators, and put them in series, or I get bigger spheres and get say 4 extra, or some variant of that.

2.) I use the generator to charge a Leyden Jar. I disconnect the generator once it's fully charged. Then, in order to increase voltage, I increase distance between the plates, lowering capacity, increasing voltage. This will almost certainly be done mechanically, with a lot of rubber insulation. See the attached diagram for this one if it'd help!

Thoughts? I'm leaning option 2 because of how expensive option 1 will be, though 2 has its distinctive sketchiness...

Thanks for your time, other ideas are welcome!

macckone - 1-3-2017 at 23:41

This won't create positrons, but it will give you the voltage
you need.

Option 3:
Charge the leyden jars in parallel, then with great care
not to electrocute yourself change them to a series configuration.

Option 4:
Marx Generator and ditch the Van DeGraaff.
This is what is used for really high voltage experimentation.
The capacitors need to be 10KV variety as do the
resistors. With a 10K DC power supply you will need 150
stages but it is straight forward. A narrow beam UV diode
can be used as a trigger to get instant flow across all
spark gaps.

Sulaiman - 2-3-2017 at 01:39

I expect that you will have serious corona problems,

charge can ionise air once the field gradient is sufficient, for rounded objects use 3 kV per mm of radius.

So 1.5 MV = 1500/3 = 500mm

therefore, to avoid breakout at 1.5 MV, the minimum diameter of spheres or cylinders is 1 metre.

One common 'trick' is to use +0.5.V and -0.5.V = +/- 750 kV in this case.

I remember a tall brick building at Harwell, built to house a tandem VDG to generate 1 - 2 MeV particles by this method, it was over 100' tall.

You would probably find a microwave LINAC or synchrotron more practical.
(I helped to commission a 55 MeV microwave LINAC at Harwell in the '70's)

The classic h.v. generator is https://en.wikipedia.org/wiki/Cockcroft%E2%80%93Walton_gener...
though I suspect that 1.5 MV will require a lot of input power just due to leakage/corona.

As above, a Marx generator for +/- 750 kV may be doable at hobby scale https://en.wikipedia.org/wiki/Marx_generator

P.S. due to the very high dV/dt generated by a Marx generator, EXTREMELY long arcs through the air can be formed
(you thought chemistry was a good mix of fun & danger ?)

IF you do generate 1.5 Mv (or +/- 750 kV) the machine will have to be quite large to withstand such voltages,
so the mean free paths will be measured in metres - quite a hard vacuum
(fingerprints on the inside of the machine can add many hours to pump-down time, that was using roughing-, turbo- and ion-pumps)

[Edited on 2-3-2017 by Sulaiman]

wg48 - 2-3-2017 at 03:25

Invictos: Your option 2 is a form of influence machine. The changing of capacitance can be done with rotating discs. One form of influence machine is a self starting Wimshurst machine that would not require your Van De Graaff generator. The one below has 7 feet discs allegedly the largest in the world. Yours would be huge as it might need 50ft discs unless you put it in a pressure tank of .sulfur hexafluoride but the discs would still have to be say 3ft thick.

Start the crowd funding now and call Guinness LOL

Check out Antonio Carlos M De Queiroz (a mad science type of guy) site http://www.coe.ufrj.br/~acmq/bigwin.html with links to lots of HV machines and info.

If you don't mind the high voltage being alternating you could do it with a big Tesla coil like the one below.

From https://www.quora.com/What-does-a-Tesla-coil-do

[Edited on 2-3-2017 by wg48]

Sulaiman - 2-3-2017 at 04:51

I remember Greg Leyh sharing the development of the Electrum on pupman.com
where there is enough information to reproduce an Electrum,
a fantastic project, not for the faint-hearted.

.........................................................................................
for option 2, the voltage will rise proportional to distance (ideally)
e.g. capacitor plates 5mm apart charged to 15 kV would only need to be increased to 500 mm separation for 1.5 MV
- why have I never tried this !

insulation will be 'interesting'
and 'stray/parasitic' capacitance reduction will be challenging.

[Edited on 2-3-2017 by Sulaiman]

wg48 - 2-3-2017 at 06:18

 Quote: Originally posted by Sulaiman I remember Greg Leyh sharing the development of the Electrum on pupman.com where there is enough information to reproduce an Electrum, a fantastic project, not for the faint-hearted. [Edited on 2-3-2017 by Sulaiman]

Yes almost anything to do with 2,000,000V is not for the faint of heart lol. It would sort the real mad scientists from the tinkers lol.

Even for an amateur mad scientist

Morgan - 2-3-2017 at 10:46

I always liked this old gizmo.

"Note the deliberately indirect steam path. Quite what this had to do with charging the steam is at present obscure."
"While the Armstrong boilers were considered to be the most powerful electrostatic machines of their day, there were problems. In operation they inevitably filled the room with steam, producing a humid environment hardly suitable for electrostatic experiments which depend on dryness to maintain insulation. One also wonders how the machine could be fuelled while in operation without giving the stoker a severe shock; possibly the boiler would be grounded while this was done."
http://www.aqpl43.dsl.pipex.com/MUSEUM/POWER/staticboil/stat...
http://physics.kenyon.edu/EarlyApparatus/Static_Electricity/...

[Edited on 2-3-2017 by Morgan]

Invictos - 2-3-2017 at 23:27

You're all amazing, wow is all I can say!

The biggest thing I've pulled from this is that these things don't quite scale

The biggest, most out there end goal is to accelerate electrons in a vacuum tube by using this voltage differential, use the Bremsstrahlung radiation in order to convert the energy in the electrons to photons, and then use the photons in pair production. Pair production for positron/electron pairs occurs at 1.1 MeV, so I was shooting for the 1.5 MeV range to account for the energy lost in Bremsstrahlung. Given the scaling difficulty, I may need to aim closer to 1.1 MeV apparently! If I shoot enough electrons down the pipe, some of them should make it at just the right angle to get that energy, right?

Note, I crunched the numbers at 100 KeV for now, and found the power given off by Bremsstrahlung to be 3.31e-20 Watts, while the energy of the particle itself is 1.59e-14 Joules, so it does seem remotely possible. However, 1.1 MeV is really cutting it!

As for some of these ideas...

 Quote: "I've always liked this old gizmo"

I can see why, that's certainly a unique way to go about it!

 Quote: Option 3: Charge the leyden jars in parallel, then with great care not to electrocute yourself change them to a series configuration.

This is an idea...I'd bet if I used some industrial level transistors, it could be done, right? That would increase the voltage an insane amount too, because you're switching between inverse addition to regular...I don't even think you could get a transistor that big, I'd probably be easier off hooking up servo motors to metal rods to use as switches!

 Quote: Your option 2 is a form of influence machine. The changing of capacitance can be done with rotating discs. One form of influence machine is a self starting Wimshurst machine that would not require your Van De Graaff generator. The one below has 7 feet discs allegedly the largest in the world. Yours would be huge as it might need 50ft discs unless you put it in a pressure tank of .sulfur hexafluoride but the discs would still have to be say 3ft thick.

Hey I did a thing! But wow those are some numbers The disk approach is so majestic I love it. Going off of that, one could, in theory, use some sort of conductive fluid for the dialectric, and drain and fill the fluid in order to change capacitance, right? It'd be clunky, but it's a thought!

 Quote: Option 4: Marx Generator and ditch the Van De Graaff.

This has been suggested a couple of times, and I certainly see why! Looking into it, I see a ton of people have gotten quite far with this approach!

 Quote: One common 'trick' is to use +0.5.V and -0.5.V = +/- 750 kV in this case.

That's clever...very, very clever... which leads me into my next silly question: Could I use some weird macgyver style hybrid of the above? Could a Marx generator be hooked up in series to the Van De Graaff somehow? If not, could one at least create the two set ups, one for positive and one for negative...?

I have some resources here, just trying to squeak out as much as I can!
Thanks again everyone!

Sulaiman - 3-3-2017 at 01:32

The problem with putting one voltage source on top of another is insulation
... how do you supply power to something that is at e.g. 100 kV ?

(Cockroft-Walton / Villard / Greinacher) voltage multipliers and Marx generators are fairly easy to configure for +ve and/or - ve voltages

A VDG can be configured to produce +ve or -ve voltage if a h.v. supply is used to spray charge onto the belt.

Tesla coils can also be bipolar, but not useful here I think.

P.S. the idea of charging leyden jars (capacitors) in parallel then stacking them up in series
. is incredibly hazardous.
. is a manually switched Mark Generator

Invictos - 4-3-2017 at 15:57

 Quote: Originally posted by Sulaiman The problem with putting one voltage source on top of another is insulation ... how do you supply power to something that is at e.g. 100 kV ? (Cockroft-Walton / Villard / Greinacher) voltage multipliers and Marx generators are fairly easy to configure for +ve and/or - ve voltages A VDG can be configured to produce +ve or -ve voltage if a h.v. supply is used to spray charge onto the belt. Tesla coils can also be bipolar, but not useful here I think. P.S. the idea of charging leyden jars (capacitors) in parallel then stacking them up in series . is incredibly hazardous. . is a manually switched Mark Generator

I UNDERSTAND MARX GENERATORS NOW

I was looking over the circuit drawings but it never really clicked on why it worked...I see...

Thanks for your time, I've got plenty to look through and consider at this point!

unionised - 5-3-2017 at 02:17

 Quote: Originally posted by Invictos I UNDERSTAND MARX GENERATORS NOW

That's probably going to be the greatest achievement of this thread.

Sulaiman - 5-3-2017 at 02:47

The more that I think of generating 1.5 MV the less practical I consider it;

constant potential difference (d.c.) causes air to ionise at about 3 kV/mm,
1.5 MV is therefore equivalent to a minimum gap in air of 500mm
or a curved conductor of 500 mm min. radius.

Due to inevitable surface contamination, monoatomic layers of water etc.
Electricity will arc along surfaces at much much lower v/m.

with high dV/dt the breakdown of air is much more complicated,
related to streamer growth at the tip of the arc.
e.g. you may have seen Steve Ward's tesla guns https://www.youtube.com/watch?v=R2jlgTA0QtI
< 100 kV in bursts produced those arcs if I remember correctly !

When Marx generators are used, if the risetime is just right ...
http://teslamania.delete.org/frames/longarc.jpg

In short, I doubt that the simple approach of accelerating charged particles in a linear electric field is a viable method of producing 1.5 MeV electrons.
(with the possible exception of bipolar VDG, BIG, EXPENSIVE, but doable)

I think that in an amateur setting, the most likely to succed is a cyclotron https://en.wikipedia.org/wiki/Cyclotron

A LINAC using a microwave oven magnetron may be another viable option ...
http://hpschapters.org/florida/13PPT.pdf
http://users.ox.ac.uk/~atdgroup/referencematerial/The%20novi...

P.S. it seems to me that a 1.5 MeV accelerator would become your only hobby for the forseeable future if you go that way.

Is there a university near you ?
Academics (desperate to publish a paper), and facility operators (I suggest 1 beer per MeV.minute) can be incredibly helpful !
Once an Academic has introduced you to the operator/technician, access is easy (in my case it was for S.E.M. time)

[Edited on 5-3-2017 by Sulaiman]

[Edited on 5-3-2017 by Sulaiman]

wg48 - 5-3-2017 at 06:03

Sulairman is correct the technical difficulties will be a challenge to say the least and the machine will be large so you will need a lot space if its air insulated. I looked for a paper I read about HV construction techniques and generators but failed to find it. Some experimenters at Imperial College London built a million/s volt discharge machine from the info. They used rolls of polyethylene film and aluminium foil to build the caps of a Marx. They used dilute cooper sulphate solution in plastic tubes for resistors. The Marx was about 8 feet high. The Marx output voltage was stored briefly in a very pure water insulated transmission line/s and doubled twice to get to the million/s of volts and increase the discharge speed.

A large Tesla coil can be used I recall an article in Scientific America. It used a Tesla coil to power an evacuated long glass tube to accelerate electrons. A filament at one end of the glass tube rectified the alternating voltage, an x-ray tube. It could sterilize seeds. I don’t recall if it was the x-rays or the electrons that achieved that. You could build a bipolar coil and mount the glass tube between them. The coils will have to be long perhaps 8ft or more.

An other problem from your description of what you’re doing is: Brem photons are broad band. Meaning the electrons kinetic energy is converted in to several photons and only very occasionally in to a single one and then that would probably require very high deceleration acceleration. Probably the level of deceleration/acceleration you get when they hit/interact with a proton/electron or even better the nucleus of a heavy atom.

I assumed you where going to fire the electrons at a suitable target that would generate the positrons. But I don’t think that will work or is very improbable. You would need to hit nucleus that absorbs the negative electron then decays and emits a positive positron. Do any do that? Yes if they are traveling very fast you can make a shower of partcles but thats a lot higher than 1 to 2 MV of acceleration.

[Edited on 5-3-2017 by wg48]

[Edited on 5-3-2017 by wg48]

unionised - 5-3-2017 at 06:13

There are not many projects where a DIY cyclotron looks like the sensible option: this is one of them.
WGTR - 5-3-2017 at 07:21

How much beam current do you need (at 1.5MV)?

There are a number of ways to make low power, fairly compact modules at these high voltages, but the materials are not cheap, and the whole high voltage section would need to be encapsulated in a dielectric.

Invictos - 5-3-2017 at 14:41

 Quote: Originally posted by WGTR How much beam current do you need (at 1.5MV)? There are a number of ways to make low power, fairly compact modules at these high voltages, but the materials are not cheap, and the whole high voltage section would need to be encapsulated in a dielectric.

Whatever current we could reasonably get! We're trying to get pair production to occur, and I don't believe current is a factor in that (correct me if I'm wrong)

As for the target, we were planning on using tungsten.

Also, as this thread is taking a bit of a...pessimistic.. turn (darn Sulaiman, you and your knowledge...pft...), I've got a question about plan B that I'm going to put in another thread in a minute here...

Again, y'all are perfect

WGTR - 5-3-2017 at 16:32

Is this a personal curiosity, or a funded project? It's not a trick question; it would help you the most to have a solution tailored to your needs and resources.

Much of my electronics work has been in general power electronics design, up to about 50kV, but others in our facility work specifically with high voltage. The power side sounds doable on a benchtop to me, but without understanding all the details of your application it's impossible to say for sure. It would require some rather cleverly enforced isolation techniques.

The beam current question comes back to me not understanding the application very well. Are we talking about pA, or several amps? Does the current need to be continuous, or pulsed? If pulsed, then with what pulse length? If current requirements are in the pA or nA range, then power supply design becomes much simpler (and safer for you to use).

Dow Corning is my go-to place for dielectric potting gels, but generally their stuff is going to run several \$100 per gallon. If it's what the Dr. orders, though...

Sulaiman - 6-3-2017 at 06:20

even if a potting gell is the ultimate goal,
unless you are confident that no future repairs or modifications may be required

Invictos - 6-3-2017 at 19:09

 Quote: Originally posted by WGTR Is this a personal curiosity, or a funded project? It's not a trick question; it would help you the most to have a solution tailored to your needs and resources. Much of my electronics work has been in general power electronics design, up to about 50kV, but others in our facility work specifically with high voltage. The power side sounds doable on a benchtop to me, but without understanding all the details of your application it's impossible to say for sure. It would require some rather cleverly enforced isolation techniques. The beam current question comes back to me not understanding the application very well. Are we talking about pA, or several amps? Does the current need to be continuous, or pulsed? If pulsed, then with what pulse length? If current requirements are in the pA or nA range, then power supply design becomes much simpler (and safer for you to use). Dow Corning is my go-to place for dielectric potting gels, but generally their stuff is going to run several \$100 per gallon. If it's what the Dr. orders, though...

Sure! This is a hobby grade project, so our budget is pretty much duct taped together...Also, our requirement as far as current goes is as low as we can reasonably expect to detect! And Sulaiman, a mineral oil or something is probably a smart idea to try it all!

Invictos - 11-3-2017 at 17:12

Hey everyone, reopening this can of worms...silly question: What would happen if you were to make a massive Marx generator while using Leyden jars as the capacitors?

The dielectric breakdown of something macroscopic like glass or plastic is probably stronger than the breakdown of a store bought capacitor, right? We have absolutely plenty of space, so things like 20cm spark gaps aren't an issue, and assuming we can charge them faster than the corona puts out, wouldn't we be alright?

Isn't that what this is, albeit on a less industrial scale? Would this be able to scale? : https://www.youtube.com/watch?v=5-nmtNg1i10

[Edited on 3-12-2017 by Invictos]

jpsmith123 - 11-3-2017 at 17:53

I'm wondering, are you just looking to make big sparks, or do you need an electron beam?

[Edited on 12-3-2017 by jpsmith123]

Invictos - 12-3-2017 at 05:38

 Quote: Originally posted by jpsmith123 I'm wondering, are you just looking to make big sparks, or do you need an electron beam? [Edited on 12-3-2017 by jpsmith123]

We just need a high voltage in order to get a short burst of an electron beam. (I think that is a valid answer to that question )

unionised - 12-3-2017 at 05:45

 Quote: Originally posted by Invictos H The dielectric breakdown of something macroscopic like glass or plastic is probably stronger than the breakdown of a store bought capacitor, right? [Edited on 3-12-2017 by Invictos]

If it was, guess what they would make store bought caps from...

Invictos - 12-3-2017 at 06:10

Quote: Originally posted by unionised
 Quote: Originally posted by Invictos H The dielectric breakdown of something macroscopic like glass or plastic is probably stronger than the breakdown of a store bought capacitor, right? [Edited on 3-12-2017 by Invictos]

If it was, guess what they would make store bought caps from...

I'm sure they use Kapton or something similar, but I bet half an inch of plastic is stronger than the hair's thickness they try to cram into those caps!

After all, how many people would be buying five gallon buckets for hobby electronics?

Morgan - 12-3-2017 at 09:56

There are these film caps that probably wouldn't have a long life or duty cycle but might work for some things. I bought a dozen just to fiddle with.
"30kV DC 1000pF High Voltage Film Capacitor" for \$1.84 a piece on eBay.
It's interesting how tricky some of the designs are for "film capacitors" on the Wiki page.

jpsmith123 - 12-3-2017 at 13:04

High voltage will make sparks, but that's not the same thing as an electron beam.

You might have to do something like what's described in this paper:

But you'd have to scale it up to get the voltage that you want, which would probably involve replacing the spiral generator with an appropriately designed and driven air core transformer, and increasing the size of the device to prevent HV breakdown across insulator surfaces.

Quote: Originally posted by Invictos
 Quote: Originally posted by jpsmith123 I'm wondering, are you just looking to make big sparks, or do you need an electron beam? [Edited on 12-3-2017 by jpsmith123]

We just need a high voltage in order to get a short burst of an electron beam. (I think that is a valid answer to that question )

Invictos - 13-3-2017 at 04:29

Quote: Originally posted by jpsmith123
High voltage will make sparks, but that's not the same thing as an electron beam.

You might have to do something like what's described in this paper:

But you'd have to scale it up to get the voltage that you want, which would probably involve replacing the spiral generator with an appropriately designed and driven air core transformer, and increasing the size of the device to prevent HV breakdown across insulator surfaces.

Quote: Originally posted by Invictos
 Quote: Originally posted by jpsmith123 I'm wondering, are you just looking to make big sparks, or do you need an electron beam? [Edited on 12-3-2017 by jpsmith123]

We just need a high voltage in order to get a short burst of an electron beam. (I think that is a valid answer to that question )

I'll look into that! And we have electron emitters and a tube, we're just using the voltage difference to actually accelerate them

macckone - 20-3-2017 at 09:33

https://www.sciencenews.org/article/signature-antimatter-det...

Spark created positrons.

jpsmith123 - 20-3-2017 at 20:12

According to the following linked paper, relativistic electron beams have also been detected during storms; so maybe it's not the "sparks", per se, that are making positrons, but bremsstrahlung from the high energy electrons?

http://www.atmos-chem-phys.net/11/7747/2011/acp-11-7747-2011...