Pulse Power Switching Devices - An Overview

By John Pasley

5/8/96

Copyright John Pasley 1996. This document may be freely distributed via. any
means in part or in whole, however the authors name must be included and
correctly attributed. The References Listed and The Disclaimer must also be
included.

An essential component of modern nuclear weapon technology is the ability 
to rapidly switch high voltage/high current electrical circuits at very high 
speeds.
The detonators that fire high explosive implosion systems (exploding wire or 
exploding foil detonators) require voltages in the range of (roughly) 2-20 
kilovolts, a complete detonating system may draw currents ranging from 10 
to 100 kiloamps. Pulse neutron tubes, used to precisely control the 
initiation of fission chain reactions, require voltages of 100 to 200 
kilovolts, and currents in the ampere range. These currents must be turned 
on rapidly and precisely, timing accuracies of tens to hundreds of 
nanoseconds are required.

Switching devices that meet these stringent requirements often require 
specialized technologies or skills to manufacture. They are also dual use - 
in addition to weapons applications they have many civilian uses too. 
Examples include controlling flash lamps used in high speed photography or 
industrial photochemistry, generating high power radar pulses, in high 
energy physics laboratory equipment, to name but a few. Consequently 
commercial sale and international trade in these devices is permitted, but 
it is also regulated. Attempts to circumvent these regulations have gained 
considerable public attention in a number of technology smuggling cases in 
the 1980s and 90s involving one particular type of device - the Krytron. 
These devices have even appeared prominently in popular entertainment as 
the "McGuffin" used to drive espionage thrillers - like Roman Polanski's 
"Frantic".

This article summarizes the basic technologies and devices, and their 
principle properties.

Section 1.0: Introduction.

Before entering into a consideration of the individual devices that concern
us, it would be as well to explain some of the associated technology/
terminology. 

1.1 Switching basics and terminology.

The switch is possibly the most elementary device in the field of
electronics. A switch controls the flow of current in a circuit in a manner
such that either the current flows at a value determined by the other
components in series with it, or does not flow at all, as the case may be.
However this ideal behavior is actually never exactly what is seen in real
life. A switch has it's own parameters that determine how fast it can switch
from open to closed, or how rapidly it can interrupt the flow of current
once it is has been opened. Also of course there are more elementary
considerations such as the current handling capacity of the switch and the
peak voltage it can cope with before damage or other unwanted effects occur.

Mechanical switches such as are common in the home are in actuality far
from ideal in their behavior. The time taken to switch from off to on ( the
commutation time) is typically in the millisecond range. Also spurious
effects such as bouncing may occur as the switch fluctuates rapidly from
open to closed in the process of being physically manipulated by the operator.

Electromagnetic relays and reed switches experience similar problems to
those seen in the humble light switch. Long commutation and switch bounce
are standard features of virtually all mechanical switching devices. 

With the advent of transistors and similar devices such as thyristors one
would have thought that these slow switching problems would be things of the
past. This is in fact largely true. But semiconductors are limited in other
ways, it is very hard to find semiconductors capable of switching many
kiloamperes especially at potentials in the kilovolt region, and those
devices that can manage high currents such as the larger thyristors are
troubled by overly high commutation times. Whilst there are now
semiconductors coming onto the market capable of performing at these
extremes of current and voltage there are some requirements which put even
these devices to shame. If you want to switch 50 kilo Amperes with a sub 20
nanosecond commutation time at 20kV you are going to be in trouble if you
are relying on semiconductor technology. However there is an alternative
class of devices that have been around long before the humble transistor
came on the scene. You might think that vacuum tubes and similar are a
thing of the past. But for problems of this magnitude they are the only
things on the market that will do the job.

1.2 Vacuum and Gas filled switching tubes, introduction and terminology.

There are a great many different types of vacuum tube in existence, however
it is possible to group tubes according to some fairly basic criteria. There
are two primary distinguishing features, the source of free electrons
within the device and the gaseous filling (or lack of it) within the tube
envelope. The later of these two concepts we have already introduced by
implication. A vacuum tube is a device with a vacuum (very low pressure gas)
filling. And a gas filled device is, as the name would suggest, filled with
gas that might be at a pressure somewhat above or below atmospheric. The
type of gas used is also an important feature, particularly in switching
tubes where a wide variety of fillings are encountered. 

The source of the free conduction electrons in the device may be either
thermal such as a heated filament physically associated with the cathode of
the device - a hot cathode, or alternatively a simple consequence of a high
voltage gradient across the device, resulting in autoemission from the
cathode. A device employing this latter method is known as a cold cathode
device. In high voltage switching the presence of high voltages, and hence
the possibility of large voltage gradients within devices means that the
cold cathode system, quite a rarity in most other types of tubes, is the
norm rather than the exception. 

Other important terms encountered in gaseous state switching tubes:

Delay time. 
The delay time is the time taken between the application of a trigger pulse
and the commencement of conduction between the primary electrodes.

Jitter. 
Jitter is the variation of time delay from shot to shot given similar
electrical stimulus.

Commutation time.
The commutation time is the time taken for the conduction to reach maximum
once it has commenced. (i.e. From the time from the end of the delay time to
the time at which the maximum level of conduction occurs.)

It should be pointed out that none of the switching tubes we are about to
consider look very much like the things in the back of an old radio set.
Many are large, some exceptionally so. Also glass has largely given way to
ceramic in the higher powered devices. Before you go down your local
electronics shop or radio shack it should also be pointed out that many of
these devices besides costing $100's (often $1000's) a piece, and are also
largely unavailable to the general public due to their application in
advanced missile and nuclear weapon technologies. Of these devices the most
'everyday' is the ignitron which finds much application in industrial
welding situations.

The following devices are considered herein:

            2.0  Vacuum and Gaseous State Switching devices
            2.1  Introduction to Cold Cathode Trigger Tubes
	        2.2  The Krytron.
	        2.3  The Sprytron.
            2.4  The Thyratron.
            2.5  The Over Voltage Spark Gap.
            2.6  The Triggered Spark Gap.
            2.7  The Ignitron

In addition I will include a short section on some of the solid state
devices that are finally beginning to fill the shoes of the above gaseous
state device (to a very marginal extent in most cases).

Section four will detail the ways in which these devices might be employed
in nuclear weapons.

2.0 Vacuum and gaseous state switching devices

Most of the devices in this section switch by inducing an arcing process in
a gaseous medium. I have included in the triggered spark gap section some
mention of devices that actually use a liquid or solid substitute for the
gaseous material that is the norm in triggered spark gaps. 

The process of arc formation is actually quite complex physically, and it
will not be gone into in any depth. Anyone who wishes to look more deeply
into this aspect of device operation may contact the author for some
suggestions as to suitable text books for use in such study.

2.1 An Introduction to Cold Cathode Switching Tubes.
Cold cathode trigger tubes are physically small devices designed to switch
impulse currents and voltages of relatively small amplitude. Usually they
are intended, as their name suggests, to trigger other larger devices. 

Typically cold cathode trigger tubes are designed to switch pulses of a few
hundred volts and a few hundred milliamperes. Most trigger tubes have three
or four electrodes, anode, cathode (+ve and -ve terminals respectively), a
trigger/control electrode and sometimes a priming electrode. 

A trigger tube performs in a very simple manner akin to that of a triggered
spark gap, excepting that usually the conduction is not by an arcing but
glow discharge. The glow discharge is initiated when all of the following
factors are present:
     i) A sufficiently high voltage is present across the device 
       (between anode and cathode)
    ii) A trigger pulse of sufficient amplitude is present at the trigger
        electrode.
   iii) The gas in the tube is primed.

Cold cathode trigger tubes rely upon some external or internal source to
ionize the gas suitably for conduction to commence (This is called priming).
This means that in theory some of these tubes will only switch a minute or
so after the application of a suitable triggering voltage to the appropriate
terminal of the device when some natural source of ionizing radiation
ionizes the gas (forming a plasma) and hence causes conduction to commence.
The triggering is basically random- it is subject to huge statistical
variation even in apparently similar environments. Some devices incorporate
a suitably ionizing source to reduce the maximum possible time delay after
trigger application considerably. This source may be an electronic,
radioactive or photon source of some form or other. However even the
standard commercial devices often display a large variation (up to and above
an order of magnitude different) between devices fired in sunlight and
darkness, a standard commercial tube Z900T for instance displays a 20us
delay in day light and a 250us delay in darkness.

2.2 The Krytron:
Krytrons are completely different to the familiar Klystrons often encountered.

Krytrons are a highly specialized variety of cold cathode trigger tube. They
were one of the first products developed by the US based company EG&G. The
Krytron has 4 electrodes, and is filled with a gas at low pressure. A
Krytron is distinguished among cold cathode trigger tubes for a variety of
reasons. 

The Krytron is designed to switch moderately high impulse currents (up to
around 3kA) and voltages (Up to around 5kV) in an arc discharge mode,
compare this with the usual glow discharge of the standard trigger tube.
Also, and perhaps more importantly, the Krytron is able to turn on this arc
discharge very rapidly, the reason being that it relies on an already
present plasma to support the conduction, rather than waiting for the plasma
to be formed as a result of priming etc. This plasma is created and
sustained by a keep-alive current between the keep-alive electrode and the
cathode of the device. When the trigger is applied under the conditions of a
high anode to cathode voltage, this plasma forms an easy path for the main
conduction between anode and cathode. 

The fact that a conduction path is already established prior to triggering
makes a huge difference in the commutation time of these devices compared to
standard cold cathode trigger tubes. Commutation times below 1 nanosecond
are achievable with Krytrons and the time lag between application of trigger
and the commencement of switching may be less than 30 ns with an optimized
driver circuit. (Note this delay is largely due to the fact that the ionized
path will need to spread from the keep alive terminal to the anode of the
device). 

Compare this delay time to that seen in the standard trigger tube
which is dependent upon many environmental factors and typically 3 or 4
orders of magnitude greater. Note that the variation in time delay exhibited
by the krytron is almost totally independent of environment, however the
time delay may be reduced up to a point with increasing trigger voltage.
Likewise the commutation time is generally decreased if the rise time of
the trigger pulse is also decreased. Given identical trigger pulses however
a krytron will have a very similar time delay from one shot to the next.
This variation is known as jitter and may be less than 5ns in optimal
circumstances.

This short commutation time and inherent environmental insensitivity of the 
krytron is achieved by including a radioactive priming source, the weak beta 
emitter Nickel-63. The source maintains weak the gas filling the tube in a weak 
state of ionization, which  aides the formation of the initial glow discharge 
between the keep alive and the cathode. Ni-63 has a half-life of 92 years and 
produces beta particles with an energy of only 65.9 KeV and no gamma radiation 
at all. The quantity in each device is less than 5 microcuries and presents no 
significant hazard. Usually the source is pulse welded to a piece of Nickel wire that 
is in turn welded to one of the electrode supports. 

Priming sources  (which is also occasionally a radioactive source) are also be 
employed in standard trigger tubes to reduce their environmental sensitivity.

Krytrons typically come in a small glass envelope somewhat similar to a neon
indicator bulb with more leads. 

Krytrons require a high voltage pulse (500V to 2kV) to be applied to the
trigger electrode to fire successfully. This pulse is almost always
generated by a pulse transformer fired by a capacitor discharge in the
primary (rather like a simple strobe tube firing circuit). 

The krytron often has only a short life expectancy if used regularly (often
as few as a couple of hundred shots) However when used within the
appropriate parameters and well within the expected life time they are
extremely reliable, requiring no warm up and being immune to many
environmental factors to a large extent (e.g. vibration, temperature,
acceleration).

These properties, combined with the small size make the krytron ideal for
use in the detonating circuitry of certain types of missiles and smart bombs.
The krytron may be used directly to fire a high precision exploding wire, or
alternatively as part of the triggering circuitry for a triggered spark gap
or similar ultra high current triggering device as used in exploding foil
slapper type detonators and larger EBW circuits.

Krytrons are used in firing circuits for certain lasers and flash tubes and
also in some pulse welding applications, often as triggering devices for
other larger devices such as Thyratrons and spark gaps. 

The export/sale of krytrons is controlled under Dual Use Guidelines. 

2.3 The Sprytron.
The Sprytron, otherwise known as the Vacuum Krytron, is a device of very
similar performance to the Krytron. Though it generally exhibits a somewhat
lower time delay after triggering. The Sprytron is designed for use in
environments were high levels of radiation are present. The sprytron is a
hard vacuum 'filled' device unlike the krytron which, as noted above
contains a low pressure gas. 

The Sprytron has only three leads, (no keep alive), but is otherwise very
similar in outward construction to the Krytron. The reason for the use of a
vacuum filling is almost certainly that there is no medium present for
radiation from the external environment to ionize (such ionization could
promote spurious triggering effects.)

The Sprytron requires a more powerful trigger pulse than the Krytron, as the
device works by forming an arc directly between the trigger and the cathode,
which causes the tube to breakdown (go into conduction) by disrupting the
field between the anode and cathode.

Although it is usually stated that Krytrons were the devices sought in smuggling
involving Israel, Iraq and Pakistan, it may well be that Sprytrons were involved with 
at least some of these cases, rather than the gas filled krytron, due to the very similar 
size and performance coupled with very high radiation resistance. 

The reason that radiation sensitivity is important in nuclear warheads is not actually 
that the radiation levels inside a bomb are liable to be especially high. It is rather that a nearby nuclear explosion may in the case of a non radiation hard device, cause an
undesirable condition in the warhead. This defect may prevent the weapon
from being able to explode either at the correct target or at all. Also it
is conceivable that such a scenario could result in the detonation of the
second device. (nuclear dominos if you like).  It is not considered
particularly helpful to have nuclear weapons explode at the wrong moment. 
Even a few tenths of a second can be vital if the target is such that
it requires a very precisely timed detonation. (e.g. if an attempt is being
made to damage a  hardened installation such as a missile silo or command
center)

A Sprytron is triggered in a similar fashion to Krytron, but as mentioned
requires a higher energy trigger pulse and therefore a more powerful trigger
transformer etc. EG&G makes trigger transformers optimized for use with
their various tubes, and also make devices named Krytron-Pacs which
incorporate a gas filled krytron and trigger transformer in a single housing.

One final point. It is interesting to note that in application circuits
(references 1 and 4) the sprytron is always shown directly switching a load
(an Exploding bridge Wire.) and a Krytron is always shown triggering a
secondary device such as a triggered spark gap. This may just be
coincidence, but as there is no apparent reason why a sprytron couldn't be
used in either role, it also occurs to me that possibly such a design might
be significant in the design of nuclear warheads, certainly this would
appear to be a useful combination (see section 4).

2.4 Thyratrons: 
Thyratrons come in several varieties. All work similarly to the
semiconductor Thyristor, one difference being that in many designs (Hydrogen
Thyratrons are a common exception) the gate must be biased highly negative
in the off state and then biased positive to achieve switching. Like
Thyristors, Thyratrons operate like a latching switch, ie. once you have
turned them on you can only turn off by cutting the supply to the main
circuit. Mercury filled Thyratrons are the slowest, least useful type and
are much more restricted environmentally than other types due chiefly to
problems with the mercury condensing . They are rarely used as they have few
advantages of the thyristor. Hydrogen Thyratrons are *much* faster
switching than Thyristors. Some can achieve commutation in under 20ns. Inert
gas fillings tend to offer superior performance compared to mercury filled
devices, without matching the speed of the Hydrogen filled devices.

Note that Hydrogen Filled Devices employ a hot cathode.

The actual Physical construction/ operation of the thyratron is quite
complicated compared to the other devices we have looked at and no attempt
will be made to explain it's operation. The reader is advised  to consult a
wide range of books as devices employing different fillings, or electrode
heating methods operate differently. It is not considered to be especially
important to consider all these variations here as this is merely an
overview of these devices and is not intended to be the final word on the
subject. However, in order to differentiate the thyratron form other similar
devices and to define it in at least some physical manner here follows
Frungel's (Ref.4) definition of the device: 

'By the term 'thyratron' there is meant a discharge chamber in which are
arranged a cathode, one or several grids, and an anode, and which is filled
with an inert gas or metal vapor.'

Some Thyratrons can  handle up to 50kV(double gap types)  switch thousands
of Amperes and handle very high power outputs( e.g. CX 1154 can handle peak
powers of 40MW). Typical applications are Radar pulse modulators, Particle
accelerators, Lasers and high voltage medical equipment. Another variety of
thyratron is filled with Deuterium. These Deuterium filled devices are
similar to their Hydrogen filled counterparts  but the sparking potential
for Deuterium is higher thus allowing even higher voltages to be handled.
E.g. E3213 can switch 70kV (double gap type). Specialist Thyratrons with
ceramic and metal bodies are encountered. These are  designed to be used in
extreme environmental conditions. There is a wide variety of grid
configurations seen in Thyratrons, it would be impractical to consider them
all here. Manufacturers of Thyratrons Include EG&G, GEC, English Electric
Valve Co.Ltd, M-O Valve co.Ltd. Big Thyratrons often require you to get a
big box full of driver/control circuitry. Prices vary from a couple of
dollars to thousands. Hot and cold cathode type devices are encountered. 

Note these ratings are the exception rather than the rule in Thyratron
devices, devices designed for sub kilovolt voltages and only capable of
handling a few tens of amps pulsed are common enough.

Thyratrons typically come in either small multi pin base type packages such
as are common in other vacuum tubes or in the case of the higher current
devices large tubular packages with hefty end connectors. 

2.5 The Over Voltage Spark Gap
The Over voltage spark gap is essentially just two electrodes with a gap
between. When the voltage between the two electrodes exceeds the breakdown
voltage of the gas, the device arcs over and a current is very rapidly
established. The voltage at which arcing occurs in these devices is given by
the Dynamic Breakdown Voltage, which is the voltage at which the device will
breakdown for a fast rising impulse voltage. Note that this voltage may be
as much as 1.5 times greater than the static breakdown voltage (breakdown
voltage for a slowly rising voltage.) how much greater than the static
breakdown voltage the actual breakdown voltage is will be depends almost
entirely on how rapidly the voltage rise, a shorter rise time means a higher
breakdown voltage. Commutation times for these devices are exceptionally low
(sometimes less than 1nanosecond).

Overvoltage gaps are primarily used for protection. But in combination with
the other devices mentioned here they are commonly used to sharpen the
output pulses (decrease the rise times) of very high current pulses form
triggered switching devices e.g. Thyratrons.

The size of these devices is almost entirely dependent upon how much
current/voltage they are intended to switch, There is really no limit as to
the size of these devices they can be as small as krytrons, however they can
also be very big, and devices intended to switch MA will be just that.

2.6 Triggered spark gaps
The triggered spark gap is a simple device, a high voltage trigger pulse
applied to a trigger electrode initiates an arc between anode and cathode.
This trigger pulse may be utilized within the device in a variety of ways to
initiate the main discharge. Different spark gaps are so designed to employ
one particular method to create the main anode to cathode discharge. The
different methods areas follows-

Triggered spark gap electrode configurations:

i)   Field distortion: three electrodes; employs the point discharge (actually
     sharp edge) effect in the creation a conducting path

ii)  Irradiated: three electrodes; spark source creates an illuminating
     plasma that excites electrons between the anode and cathode.

iii) Swinging cascade: three electrodes; trigger electrode nearer to one of
     the main electrodes than the other.

iv)  Mid plane: three electrodes; basic triggered spark gap with trigger
     electrode centrally positioned.

v)   Trigatron: trigger to one electrode current forms plasma that spreads to
     encompass a path between anode and cathode.

The triggered Spark gap may be filled with a wide variety of materials, the
most common are-
1) Air
2) SF6
3) Argon
4) Oxygen

Often a mixture of the above materials is employed. However a few spark gaps
actually employ liquid or even solid media fillings. Solid filled devices
are often designed for single shot use (they are only used once- then they
are destroyed) Some solid filled devices are designed to switch powers of
10TW (10 000 000 000 000 Watts) such as are encountered in extremely
powerful capacitor bank discharges. Except (obviously) in the case of solid
filled devices, the media is usually pumped through the spark gap. Some
smaller gaps do not use this system though.

Usually Gas filled spark gasp operate in the 20-100kV / 20 to 100kA range
though much higher power devices are available. I have one spec for a
Maxwell gas filled device that can handle 3 MA - that's 3 Million Amperes!
But then it is the size of a small car!! More commonly gas filled devices
have dimensions of a few inches. Packages are often shaped like large ice
pucks though biconical, tubular and box like structures are also seen.

Sparkgaps are often designed for use in a certain external environment(eg.
they might be immersed in oil). A system for transmitting the media to the
appropriate part of the device may sometimes be included. Common
environments used are:
a)Air
b)SF6
c)Oil

Typical spark gap device no.'s are: TG7, TG113, TG 114 etc. etc. 

Spark gaps are damaged by repeated heavy discharge. This is an inevitable
consequence of such high discharge currents. Electrode pitting being the
most common form of damage. Between 1 and 10 thousand shots per device is
usually about what is permissible before damage begins to severely degrade
performance.

EG&G make miniature triggered spark gaps specially designed for defense
applications. these devices are physically much smaller than normal spark
gaps (few cm typical dimensions) and designed for use with exploding foil
slapper type detonators. 

Laser switching of spark gaps. The fastest way to switch a triggered spark
gap is with an intense pulse of Laser light which creates a plasma between
the electrodes with extreme rapidity. There have been quite a few designs
employing this method, chiefly in the plasma research area. 

Triggered spark gaps tend to have long delay times than Thyratrons (their
chief competitor, at least at lower energies) However once conduction has
started it reaches a peak value exceptionally rapidly (couple of nanoseconds
commutation.)

2.7 Ignitrons
The ignitron is mercury vapor rectifier in which an arc is switched between
a (usually graphite) anode and a mercury pool cathode. The discharge is
initiated by an ignitor electrode which dips into the mercury pool cathode.
On application of a suitable impulse current/voltage to this ignitor an
electron emitting source is formed at the point at which the ignitor
contacts the pool. This initiates the arcing between the anode and cathode. 

It is important that the ignitor should be triggered correctly. The ignitor
requires a certain energy for successful ignition and also an 'ignitor
characteristic' application of this energy in terms of current and voltage
with respect to time. Misfiring or ignitor damage will otherwise occur. It
is also vital that no significant negative voltage should appear at the
ignitor with respect to the cathode else ignitor destruction will be the
inevitable result.

There are two main ways by which the trigger can be biased:

Anode excitation: common in resistance welding applications here the anode
bias is connected to the ignitor by means of a switch (thyristor, thyratron
etc.) and a resistor/fuse network. The ignitor current drops rapidly on
ignition as the anode-cathode voltage drops very low during conduction. 

Separate excitation: as the name suggests, here the ignitor circuit is
largely independent of the main circuit. 

Ignitrons are often used in parallel for AC power control applications.

Ignitrons must often be cooled when used continuously (ie. Not single shot
as in capacitor discharge) Water cooling is commonly employed. It is vital
that Ignitrons must be used in the correct temperature range to hot or to
cold can be very bad news for these devices- (cold leads to mercury vapor
condensing on the anode.) 

Ignitrons are very limited with regards their physical orientation. This
reason being simple that they rely upon a pool of liquid at one end of the
device that must be correctly positioned for the ignitor to function
correctly. Positioning the device so that it leans over at an angle of more
than 2 or 3 degrees from the vertical is fatal.

Most ignitrons operate at most currents between 5 Amps and 100kA and may be
suitable for voltages from a couple of hundred to 20 000 Volts.

Thyratrons and Krytrons are sometimes used in ignitron triggering circuits
along with the familiar thyristor. 

Ignitrons are suited to applications were power control of high voltages or
currents is required. Welding is probably the most common application.

3.0 Solid State Devices.

(Note this section may well be considerably
expanded following further research by the author.)

There are now a few commercially available transistors on the Market which
can switch many tens of kV. There are also a few transistors about that can
handle pulsed currents above 5kA. These devices may match for example
Krytrons and Sprytrons in terms of electrical performance, but not in terms
of size and (in the case of the Sprytron) radiation hardness.

Thyristors are widely available in designs that can handle upwards of 10kA
pulsed at several kV. They are however very slow switching devices and are
not capable of achieving even low microsecond switching speeds. 

A new class of devices is at present showing great promise in the R&D
sector. These devices are optically (usually LASER) switched devices
employing GaAs or Diamond film technologies. The reader is advised to
consult the appropriate reference below for more information relating to
these devices.