1 000 000 VOLTS


MARX GENERATORS, MARX BANKS, IMPULSE VOLTAGES GENERATORS


Le générateur de Marx

Le générateur de choc multiétages de Marx, du nom de son inventeur Erwin Marx (1893-1980), fonctionne selon le principe d'un système de condensateurs haute-tension chargés en parallele par une tension continue et déchargés en série.

La multiplication du voltage est directement liée au nombre d'étages. La superstructure forme un échafaudage, une sorte d'échelle, où les condensateurs seraient les barreaux et les résistances de charge les montants.

A Marx Generator is a clever way of charging a number of capacitors in parallel, then discharging them in series. Originally described by Erwin Marx in 1924, Marx generators are probably the most common way of generating high voltage impulses for testing when the voltage level required is higher than available charging supply voltages.

Hipotronics Impulse Generator
Spark gap - 800 kV Marx Generator
Colored resistors ans spark gaps

Le choc de tension est la génération d'une haute tension impulsionnelle de 10 kV à 10 MV avec un temps de montée de 1 nS à 1 mS et un temps de descente , ou jusqu'à la coupure, de 100 nS à 1 s. Les chocs sont uniques ou répétitifs à des intervalles de 10 mS à quelques minutes.
Choc de tension impulsionnelle O1 = origine conventionnelle du front - T1= durée conventionnelle du front - T2= durée conventionnelle de la queue.

Typical Impulse Waveform  O1=Nominal origin T1=Nominal duration of wave front  T2=Nominal duration of wave tail

La fonction uc(t )correspond à une différence de deux exponentielles dont les constantes de temps sont différentes et conduisent à la forme du choc.

12 stages Marx Generator
12 stages  x  [ C = 0.6 µF @ 100 kV ] =  1.2 MV - 36 kJ -
Haefely ™ Marx Generator


  Principe de fonctionnement

1 - Dans une premiere étape, on charge en parallèle les condensateurs Cge de chaque étage par l'intermédiaire de la source de tension continue é travers les résistances de charge Rc, les résistances série Rsi et paralleles Rpe à la tension d'amorcage Ug0 prévue ( la tension d'éclatement des éclateurs à sphère étant ajustée légérement au-dessus de cette valeur).

2 - Dans une deuxième étape, on décharge en série les condensateurs Cge, par l'intermediaire des éclateurs de couplage E, à la suite de l'impulsion de trigger (6) provenant du bloc d'allumage (trigatron ) et ionisant le premier éclateur E1.Ce phénomene déclenche par déséquilibrage des tensions et par effet de champ électromagnétique rayonnant les (n-1) autres éclateurs de couplage du générateur.

  Il est juste de mentionner que le procédé avait été déjà pratique, sous le nom de machine rhéostatique, par Gaston Plante.

6 Trigatron

Rsi and Rpe = wave shaping resistors

Rsi = front resistors ( for the control of the wave front duration )

Rpe = tail resistors : control the time to half value for the impulse waveshape. The standard lightning impulse tail time is 50 microseconds+/- 20 %

Rc =charging resistors (high ohmic value )

The charging voltage is applied to the system. The stage capacitors charge through the charging resistors (Rc). When fully charged, either the lowest gap is allowed to breakdown from overvoltage or it is triggered by an external source (if the gap spacing is set greater than the charging voltage breakdown spacing). This effectively puts the bottom two capacitors in series, overvoltaging the next gap up, which then puts the bottom three capacitors in series, which overvoltages the next gap, and so forth. This process is referred to as "erecting". A common specification is the erected capacitance of the bank, equal to the stage capacitance divided by the number of stages.

The charging resistors are chosen to provide a typical charging time constant of several seconds. A typical charging current would be in the 50-100 mA range. The charging resistors also provide a current path to keep the arc in the spark gaps alive, and so, should be chosen to provide a current of 5-10 amps through the gap.

For example, with a stage voltage of 100 kV, a desired output voltage of 1 MV (i.e. 10 stages), the charging resistors should be about 20-40 kOhms (corresponding to an arc current of 5 to 10 Amps). If the capacitors were 1 uF, then the discharge time constant would be 20 milliseconds, much, much longer than the 50 microsecond time constant of the standard test impulse. This example generator would have a stored energy of 5 kJ/stage or 50 kJ for the total system. At a charging current of 50 mA, it would take at least 20 seconds to charge the entire stack.


Applications

Selon la forme de l'onde impulsionnelle engendrée par le dispositif ( forme qui dépend des valeurs choisies pour les résistances de front et de queue) on distingue :

  Le choc de manoeuvre (switching impulse): T1/T2 = 250/2500 µs - temps de montee : 250 microsecondes , temps de queue 2500 microsecondes. Simulation des surtensions de manoeuvre . Essais diélectriques de choc.

  Le choc de foudre ( lightning impulse) :T1/T2 = 1.2/50 µs- temps de montée : 1,2 microseconde , temps de queue 50 microsecondes. Simulation des surtensions de foudre.Essais diélectriques de choc

  Le choc à front raide (Steep front chopped wave) : temps de montée : 10 nanosecondes, temps de descente : 250 nanosecondes. Simulation des surtensions engendrées par des éclateurs de reseaux électriques lors de leur amorcage.

  Simulation des perturbations électromagnétiques , en particulier les ondes NEMP  ( Nuclear Electromagnetic Pulse ) engendrées par des explosions nucléaires de haute altitude.

>What are the real world uses of Marx Generator?

Production of artificial lightning, mostly. The 3 MV unit at Mississippi State University is about four stories tall and has twenty stages, each with its own set of spark gaps and a couple of suitcase-size capacitors. The capacitors are charged in parallel through a particularly troublesome resistor network and then discharged in series through the spark gaps.

The waveform is shaped by an RC network at the output of the generator to conform to whichever IEEE or CIGRE standards for lightning or switching surge is being reproduced at the time. The actual waveform is recorded ona digital storage oscilloscope for documentation purposes. Depending on the voltage, the device takes something on the order of thirty seconds from the start of the charging process to the discharge. The discharge is initiated automatically when the voltage reaches the desired level. This is done by initiating an arc across the bottom sphere gap pair with a 4 kV impulse.

If a large (several meter) gap is being used at a high voltage, the report of the machine can be heard at some distance across the campus. Ear protection is now mandatory, but it's an unfortunate fact that many high voltage engineers suffer severe hearing loss during their careers.(M Kinsler)

They are used to generate high voltage to simulate lightning, EMP (byproduct of Nuclear explosion) and to drive particle beam accelerators. They are also sometimed used in big laser's, nuclear effect machines (Gamma and x-ray sources).

Exemples de realisation

1-Le generateur de choc du laboratoire Ampère ( Ivry - France ) , 1935

Un des deux premiers accelerateurs de particules francais ( Frederic Joliot Curie , 1937) : photo 65 k

Dans un vaste hall, haut de 18 mètres, entièrement metallise, de façon a constituer une cage de Faraday, se dresse la colonne, haute de 13 mètres, du generateur de choc de 3 000 000 V. Poids  : plus de 20 tonnes.

Le courant alternatif a 30 000 V , redresse par kenotron, charge en parallèle des groupes de condensateurs etablis en series de 5, en sorte que la decharge des eclateurs intermediaires se produit a 150 000 V .

La totalisation a 3 000 000 V s'effectue donc en 20 etages ( 100 condensateurs) .

C'est dans le Laboratoire de synthese Atomique d'Ivry , qui est la transformation des Laboratoires Ampere de la Compagnie Generale d'Electro-Ceramique, ( 1936), que fut signe l'acte de naissance de l'energie nucleaire ( CNRS). Une partie de son instrumentation historique ( spheres de decharge et colonnes de condensateurs) est transferee, en reserve, au musee Electropolis de Mulhouse.

2-Realisations personnelles  (hvlist)

Design equations are given in Craggs & Meek, "HV Laboratory Technique", 1954 (unfortunately out of print, but most libraries have it).

[Marx voltage multipliers] One of my Marx generators consists of six stages, the capacitors all hand made from long strips of mylar and aluminum foil rolled into cylinders around PVC pipe. These form the rungs affixed to two PVC uprights. The charging resistors are all 1 megohm and the spark gaps are metal balls running diagonally from rung to rung. The trigger is nothing more than a motorcycle spark coil connected to a 12 volt supply via momentary contact switch ( from Lawrence M. MacCary, 1996).

I've done this also. I've built two using the polyeth. caps. of 20 stages each. The high energy one creates a real bang when it fires.I created a new design which eliminates the resistors (and their losses and consequent slow charging, leakage etc.) The design is based on an idea I once saw in the Amateur Scientist in Sci. Am.which was for a pulsed U.V. laser utilizing a Blumlein switch.Basically, the resistors are replaced by chokes. I've found that inductance approaching 1 mH is sufficiently large for the 7000pF caps. to allow O/P voltage to reach about 95% of ideal. The rule is that the lower the capacitances, the larger must be the chokes to compensate (they effectively do a D.C. supply disconnect job). The unit I built stands about 2 feet high, and generates sparks about 7-8" long (about 270kV). The tower also emits a MASSIVE RF pulse when it fires - if you build one, DON'T operate it anywhere near any electronic gear. The spark gaps are formed from thick copper wire with the end bent into loop, the end of the loop being equivalent to your spheres.My Marx is self triggering (careful spark gap adjustment) ( from Malcolm Watts, 1996).

Basically, I constructed the towers from perspex (acrylic) with a stage height to suit the home rolled capacitors. As a result, the towers are quite short and hold 22 capacitors of around 5nF each which can be charged up to around 18kV each.Tower height is around 600 - 700mm. Max discharge length with this storage is around the 250mm mark (a bit less than half the height of the tower). With higher voltage rated caps, the discharge length could be improved substantially. I used single-layer chokes wound on ferrite rods instead of resistors to facilitate fast charging (this was the "novel" bit) and thick copper wire with one end curled as a spark gap electrode and the other end as a pressure contact to the capacitor (these are extended foil types).The general construction method for these caps was published in the March issue of the British magazine "Electronics World and Wireless World".I should note that this is a dangerous device. Total energy storage prior to firing is about 17 odd Joules with the figures given. It not only goes off with a real detonation (ear protection a must), but can give a whale of a shock if you accidentally trigger a portion of it while reaching out with a screwdriver to fire a gap part way up. Additionally, it emits a HUGE RF pulse and potentially damaging ground current. I lost count of the gearI had to repair after delightedly testing the first one in a work shop full of test equipment. The spark is a real lightning bolt. It is also oscillatory in nature (you can sometimes see this on a photograph and it could be easily shown with a panned camera). I have also charged the towers up with opposite polarities and fired them at each other. No ball lightning yet but I'm betting on getting there if it can be done. Using the choke setup, you can vary the resonant frequency of each tower. I think it would also be capable of running directly from AC if the choke inductances are tailored right and the spark gaps were more robust ( from Malcolm Watts, 1996).

From Jochen Kronjaeger, 1998

My specimen has four caps (same as I use with the tesla coil) connected by 20mOhm resistors (2x10MOhm). The three spark gaps are just bent wire. The caps are charged to nearly 30kV through another 500MOhm resistor, giving a theoretical output of 120kV; the spark length is about 10cm (4"). The fact that the resulting spark length is much shorter than e.g. with the 100kV cascade has (at least) two reasons: first, pulse voltages always have shorter spark lengths than DC (the spark has not so much time to develop). Second, the caps can never be charged to full 30kV within finite time

Date: Fri, 1 May 1998 From: "D.C. Cox"  To: hvlist Subject: Marx Bank

Basically they are HV caps charged in parallel and discharged in series. Stage parallel charge separating resistors are typically 40-50K ohms and series wave-shaping resistors are usually 100-400 ohm values. A good protection scheme is necessary between power supply and the actual cap bank as the reflected standing wave as the spark gap fires can damage the supply. In our commercial systems we use inductors, resistors, and then actually use a HV relay to disconnect the charging supply from the cap bank at 0.5 seconds before firing. We use a 558 timer with optoisolated triacs to do the timing functions after the firing button is pressed. DR.RESONANCE@next-wave.net

3- Generateur de Marx du Dr G. Hublart ( g.hub@free.fr )
Generateur de Marx realise par empilement de 9 condensateurs THT ( aluminium / polyethylene / huile minerale dans boite Tupperware )
This Marx' is powered by a 12 V induction coil
Eclateurs regles a 10 mm sauf l'eclateur d'amorcage qui est regle a 9 mm. Decharge de 300 mm ( 270 kV - 300 kV )
5 Mohm resistors are embedded in hot glue. Spark-gaps = 10 mm

4 - Refurbishment of the Ferranti 800kV Impulse Generator :

The Ferranti Impulse Generator comprises of two vertical columns each carrying four high voltage capacitors. The capacitors are charged in parallel to 100 kV and then discharged in series to give a peak output voltage of 800 kV. The performance of the generator had deteriorated over a number of years due to the failure of the high-voltage switches or "polytron gaps" .The traditional method of firing impulse generators is to use a system known as a "trigatron". In this system, the main high voltage sphere gap has a third triggering electrode, hence the name "trigatron". The impulse generator is now fully operational , capable of producing 800 kV electrical surges.( Dr Ken S. Smith )


QuestionS

> Can anyone tell me how quickly the spark gaps (triggered or not) in a Marx generator will break down ? What is the largest dV/dt I can expect on the output ?

Well, the spark gap breakdown is widely dependent on gap length, gas,overvoltage, supply of initial electrons etc, but for a 1 mm-gap, you can have it of the order of nanoseconds.The lagerst dV/dt? I don't know the largest, but in standard high-voltage testing you use a rise-time of 1,2 microseconds and apply MV, which gives a dV/dt of 10E12 V/s.

Question From: Gary Weaver ( Hvlist, 1997)

I don't know the correct answer to your question. But if I were building a Marx generator I would use something like a switch to discharge the 1st spark gap. The switch could be a jumper wire. I would short out the 1st spark gap with the wire and it would cause all the other spark gaps to discharge and produce the high voltage output spark.

Answer From: Malcolm Watts ( Hvlist, 1997)

You don't need a triggered gap at all. I have built two Marx banks that work just fine with careful gap setting. The first one to fire  should be at one end (usually the bottom) so its setting defines total energy in the discharge. The models I built used chokes instead of resistors and can be run on AC with robust gaps. I would not recommend using jumper wires or getting anywhere near a Marx bank to fire it. They are extremely dangerous. You can end up with the full  output voltage tracking along your jumper wire back to your hand. There is virtually no inductance in the discharge path and the currents are enormous. This is real lightning.

Answer From : TimRaney , TCBOR ( Hvlist , 1997)

On using a jumper wire to trigger the first gap of a Marx generator is probably not a really good idea. A three electrode trigger gap for the first gap is preferable or a hydrogen thyratron would do the trick. I tried a Marx circuit design with all gaps being the same and didn't have much luck, but I hadn't done my homework on the circuit. Frungel's book, "High Speed Pulse Technology," 1965, is a great reference and includes Marx generator design, along with some of the older physics texts that cover using a Marx generator as a HV source for a particle accelerator.

Answer From : Richard Hull , TCBOR ( Hvlist , 1997)

I agree, simple gaps, carefully setup, are definitely the way to go here for amateur work.

Sites - Liens  ( Websites - links)

Liens vers les autres sites du generateur de Marx
Web Page Remarques


Haefely impulse voltage generators Impulse generators in Switzerland
Hipotronics High Voltage Testing Impulse generators in USA
Marx Towers in Russia ( drone video )





Science & industrie Pulse power supplies (France )
Science and Technology http://home.earthlink.net/~gweaver/
Literature
Haefely literature pdf.files , html. files
Serie 100 impulse voltage generators Hipotronics , Serie 100
Instituts
Institut für Hochspannungstechnik Munich
Institut für Elektrische Energieversorgung Munich
Museums




Amateurs of HVlist (Marx' generators)
Jim Lux Jim Lux'Marx page


Daniel McCaulay 400 kV Marx Generator
Alan Yates 150 kV Marx Generator
Mike's Electric Stuff Quick and dirty Marx Generator
Jonathon Caywood 100 kV Marx Generator
Mike's Electric Stuff 1000 kV Marx Generator
Jochen's High Voltage Page 300 kV Marx Generator

4-Bibliographie / Literature

W. James Sarjeant and R.E. Dollinger -"High Power Electronics"- 1989 , Tab Prof and Ref Books, 392 pp -ISBN 0-8306-9094-8

Frungel's book -"High Speed Pulse Technology " - 1965

Craggs & Meek, -"High Voltage Laboratory Technique" - 1954

Bazelyan & Raizer -"Spark Discharge"-(CRC press)

Meek, J. M. and Craggs, J. D., "High Voltage Laboratory Technique", Butterworths, 1954, 404 pp

Kuffel, E., and Abdullah, M., "High Voltage Engineering", Pergamon Press, 1970, 377 pp

Heilbronner K.W - Firing and Voltage Shape of Multistage Impulse Generator- IEEE Trans,Power App.Syst., vol.PAS-90,num 5,1971, pp 2233-2238.

Pai, S. T, & Zhang, Q., "Introduction to High Power Pulse Technology", World Scientific, 1995, 307 pp

Aguet M. , Ianoz M. - Generateurs de hautes tensions transitoires -in Traite d'Electricite,Volume XXII, Presses polytechniques et universitaires romandes, 1990,Lausanne

Ensemble d'experimentation haute tension, Publ. Messwandler-Bau, GMBH, Bamberg.

Michel Ellenberger- Quel destin pour le laboratoire des Joliot-Curie ? - La Recherche , sept. 1994, vol. 25, n° 268 , p 948  

Pinault Michel - Frederic Joliot -Curie - 2000 , Editions Odile Jacob


ACCUEIL (RUHMKORFF)

VAN DE GRAAFF  |   RESONATEUR TESLA  |   CASCADE GREINACHER  |

PAGE GENERATEUR DE FELICI  

1999, 2016 © Lyonel Baum