Coaxial Marx Generator for EMP / High Voltage / Forums

Forums

4hv.org :: Forums :: High Voltage

« Previous topic | Next topic »

Coaxial Marx Generator for EMP

1 2 3 next

Move Thread

LAN_403

Dr. Shark

Thu Apr 03 2008, 02:01PM

Registered Member #75 Joined: Thu Feb 09 2006, 09:30AM
Location: Montana, USA
Posts: 711

Returning from hibernation, of course I have been struck by multiprojectitis straight away. I have 30 doorknob capacitors (30kV 590pF or so) that I bought for a Marx project

but never really used. Recently I came across this thread

where EVR says

If you are truly interested in creating an EMP type device, then the easiest thing to do is start with a marx generator, and here are the steps:

POORLY INEFFICIENT EMP
1. Marx generator - Build a 10 stage marx generator with as high voltage as possible (>30kv per stage if possible)
2. Use a wire antenna about 2 feet long ontop.
3. This will give results, especially as you increase stage capacitance.

HIGHLY EFFICIENT EMP (MUCHO difficult)
1. Design a marx generator (4-6 stage maximum) so that the equivalent output impedance of the marx generator is say 100 ohms. To do this is quite difficult as the marx generator is typically modelled as a coaxial line at this point.
2. Design a highly directional antenna (horn, dish, etc...) that can handle the output voltage discharge. The antenna should be designed to match the output impedance (same impedance) of the marx generator
3. Total output voltage should be at least 500kV or close to it.

This got me thinking, so I did some reading into what it takes to turn a Marx into a transmission line with a nominal imedance of Z0=100 Ohm or so. There are a couple of papers out there, e.g.

that describe co-axial Marx generators which can be coupled directly to an antenna to act as a broadband RF source. Seeing that EVR managed to knock out half of his home electronics with a non-optimized generators, there seems to be quite a lot of potential here.

EDIT: Something was very wrong here, now things are more reasonable

The question of the day seems to be: How do I get my generator to look like a 100 Ohm transmission line? Say I have 10 stages of 1.8nF than in the erect state I have 180pF. Plugging this in

Z0 = sqrt (L/C) 
Z0=100 Ohm
C=180pF

gives me L=1.8uH total, which seems a tad high with a co-axial design. This I envision to be the Marx generator sitting in an acrylic tube which is covered in grounded tinfoil or somethings along those lines. The generator will probably be around 7cm diameter and 70cm length (thats 3 x 30"), so rough calculations tell me that I am looking at an inductance in the range of 50 of 200nH. This kinda kills my whole idea, since it means I really have to put all the caps in series or have to settle for a lower impedance...

Now would one of you ham radio folks be able to suggest a suitable broadband antenna to couple this into the Aether? A TEM Horn is probably the best option, correct? For a reasonable size high directivity antanna however I would be looking at low GHz frequencies, so I want a pulse risetime below a nanosecond!!! I think this is the difficult part Dan was talking about, at least to me the short discharge time seems more of a problem than the 100Ohm thing.
From the LC time constant of the above circuit, sqrt(1.8uH*180pF ), I guesstimate the output would be in the 10MHz range. Would this be an EMP at all?

Proud Mary

Fri Apr 04 2008, 04:04PM

Registered Member #543 Joined: Tue Feb 20 2007, 04:26PM
Location: UK
Posts: 4992

Dr. Shark wrote ...
Now would one of you ham radio folks be able to suggest a suitable broadband antenna to couple this into the Aether? A TEM Horn is probably the best option, correct?

Much of the groundwork on UWB 'impulse radiating antennas' driven by Marxes was done by J. R. Mayes, W. J. Carey, Larry Altgilbers, and Everett G. Farr

They started off with a TEM horn, and graduated to a kind of centre-fed dipole centred in a folding umbrella-like parabolic reflector which could be carted about.

The flat sheet metal bow-tie driven elements were driven from the centre, and terminated where they met the rim of the parabolic with hefty resistors, of, as I remember 600 ohms.

The whole thing is described in United States Patent 6340956 here:

I understand that the invention was intended for use in disrupting electronic systems - such as motor vehicles - at a distance.

Searching for "UWB antenna" and "impulse radiating antenna" will get you what you want to know.

Dr. Shark

Fri Apr 04 2008, 05:06PM

Registered Member #75 Joined: Thu Feb 09 2006, 09:30AM
Location: Montana, USA
Posts: 711

Hey, thanks for your reply! I really don't know much about antenna theory, so if you could elaborate a little bit on what the dish does, that would be great. Does it somehow 'focus' the dipole radiation pattern of the antenna to give you more gain?

A related question concerns the impedance of antennas: I understand that free space has a characteristic impedance of about 380 Ohm, so I wonder why most antennas are around 100Ohm. Does an antenna contain some kind of implicit transformer?

One more thing I came up with regarding the design of the Marx generator: We have the frequency omega^-2 =LC and the impedance Z^2=L/C so for a given pulse width and target impedance we can exactly calculate the required capacitance and inductance. E.g. if we want energy at 1GHz, and Z=100Ohm, we get L=20nH and C=2pF. This is completely and utterly ridiculous, so I wonder where I am going wrong?

Proud Mary

Fri Apr 04 2008, 05:51PM

Registered Member #543 Joined: Tue Feb 20 2007, 04:26PM
Location: UK
Posts: 4992

Cold water for a shark, Suomi!

It is a long time since I looked at this material, so I forgot that Everett Farr has a business called Farr Research Inc which has a lot of its research papers published online here:

I understand that a lot of this work has been paid for by the little known US Air Force "Directed Energy Directorate." Uses of this technology are said to include UWB high definition radar, the jamming of missile control systems, and the remote disruption of motor vehicle electronics to create 'exclusion zones.' Whether it can be made to work reliably, or is like the robotic self-propelled Tesla coil that hopelessly failed its field trials as an exploder of 'terror' devices in Afghanistan is another matter...

Very detailed photos and descriptions of the antenna I first told you about, both the folding version and one with a hard non-folding parabolic reflector, are shown in this Farr paper here:

Now for the next bit of your question!

The "impedance of free space" is a physical constant that relates the magnitude of the magnetic and electric components of a wave as it flies through infinite space, thus: Zo = E/H. Happily, it has limited use in practical antenna design, and cannot be measured directly.

The thing to think about is the radiation resistance of the antenna, which is that portion of an antenna's input/feed point impedance caused by the radiation reaction of conduction electrons in the antenna itself.

It is to the radiation resistance of an antenna that a transmission line must be matched. The most famous radiation resistance is that of the half-wave dipole antenna, which has a characteristic impedance of 73 ohms - usually rounded up to 75 ohms for practical purposes - which is why RF co-axial cables are commonly made to have a characteristic impedance of 75 ohms.

And now for the last bit! When Marx generators are used as high power Ultra Wide Band (UWB) generators, the designer tries to make the rising front of the output pulse as nearly vertical as he can. He wants the leading edge to go straight up in zero seconds like a pure square wave, so it will contain all possible frequencies, and thereby be 'ultra wide band.' So there is always a lot of research on how to make the rise-time as short as possible to be as near the perfect vertical rise as possible.

Attempts to shorten output pulse rise times will often include the use of low value charging resistors (50 ohms or so), low inductance capacitors, synchronized simultaneous triggering of all gaps, pre-ionization of gaps, use of pseudo-spark switches, krytrons or sprytrons, and combination of the marx generator with co-axial pulse forming lines and Blumlein circuits.

That's enough for now!

jpsmith123

Fri Apr 04 2008, 11:59PM

Registered Member #1321 Joined: Sat Feb 16 2008, 03:22AM
Location:
Posts: 843

I'm also interested in fast high voltage pulse generation, but for a different application.

One major issue with regard to creating very-fast-rise-time-pulses, whether you're using a blumlein or a Marx generator, is the design and construction of the switches (usually a spark gap).

If you're talking about a few hundred kv into 100 ohms, with a rise time on the order of a nanosecond or so, then you're talking about a switch di/dt on the order of 10^12 A/Sec or so.

According to what I've read, getting this type of fast switching means operating the spark gaps at very high electric fields, which means using short gaps with high pressure gas insulation or alternatively using water switches.

Proud Mary

Sat Apr 05 2008, 12:39AM

Registered Member #543 Joined: Tue Feb 20 2007, 04:26PM
Location: UK
Posts: 4992

jpsmith123 wrote ...
According to what I've read, getting this type of fast switching means operating the spark gaps at very high electric fields, which means using short gaps with high pressure gas insulation or alternatively using water switches.

The references I've cited above in my reply to Dr Shark look at techniques for reducing output pulse rise time in exhaustive practical detail.

HV Enthusiast

Sat Apr 05 2008, 04:12AM

Registered Member #15 Joined: Thu Feb 02 2006, 01:11PM
Location:
Posts: 3068

Hi Dr. Shark. Love the name!!!

I did get your PM and unfortunately, due to obvious reasons (ITAR etc...) i can't discuss the information you requested.
However, there should be numerous public domain IEEE articles on marx generation design which is available, some on coaxial transmission line design.

Dr. Shark

Sat Apr 05 2008, 09:08AM

Registered Member #75 Joined: Thu Feb 09 2006, 09:30AM
Location: Montana, USA
Posts: 711

Hi Dr. Shock, can you guess how I came up with that? I was actually half expecting you to PM me to change my name back to joe since "impersonating authorities" is against the rules

I respect your professional obligations, but obviously I am not asking for detailed plans but rather trying to develop an understanding of the general principles involved.
But I probably got a step closer to that anyway: I found another IEEE paper

(which happens to be available to the general public) where there is a short mention of a "wave erection" Marx.
This seems to be a particular example of the principle Harry mentioned: Get as fast a rising edge as possible. With wave erection, which is maybe somewhat related to the BlÃ¼mlein circuit, a rising edge on the order of 100s of ps can be obtained.
It seems that on the old 4HV someone knew more about this

. Unfortunately the link he gives is dead, but I tracked down his stuff here

. Still I wonder if this is a feasible concept for an amateur construction?

Harry, thanks also for your crash-course in antenna theory, very much appreciated!

Proud Mary

Sat Apr 05 2008, 09:57AM

Registered Member #543 Joined: Tue Feb 20 2007, 04:26PM
Location: UK
Posts: 4992

'Wave erection' is not itself related to the Blumlein circuit, but concerns the sequential order of switch closure in a Marx:

Wave Erection Marx

The general Marx circuit does not force switches to close in succession. Therefore, switches can fire randomly, adding temporal jitter to the systemâ€™s performance.

"The transient wave Marx insures that after closure of the first switch, stray capacitances hold one side of each successive switch near ground until the switch fires. A voltage wave propagates toward the load with increasing intensity, triggering successive switches faster and faster until it reaches the load with sub-ns risetime and low jitter for output voltages of several hundred kV at moderate per pulse energies." See:

jpsmith123

Sat Apr 05 2008, 02:15PM

Registered Member #1321 Joined: Sat Feb 16 2008, 03:22AM
Location:
Posts: 843

To start out with, you might be better off with a stacked blumlein instead of a marx generator, say a stacked coaxial blumlein using coax cable charged to 30 or 40 kv or so, and
only using one switch.

There are several papers available on-line regarding this type of generator.

I have some papers somewhere dealing with design of switches, blumleins and UWB antennas. If I get a chance later I'll try to find them.

1 2 3 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob