|
|
|
Figure 1: An unusual application |
What are Transient Limiters?
It should not be surprising
that this question is asked. Transient limiters masquerade as Gas
Discharge Tubes, Overvoltage Gaps, Plasma Arrestors, Spark Gaps,
Surge Arrestors, Surge Voltage Protectors, Transient Protectors,
etc. Their principle purpose is to limit electrical transients
making Transient limiter a reasonable choice for a name.
Transient limiters
produced by Reynolds are hermetically
sealed assemblies constructed from two (sometimes more) electrodes
brazed into the ends
of a ceramic insulator and filled with a rare (or fairly rare) gas.
In the same way that a spark “jumps” across the gap
of an automobile spark plug or between the electrodes of popular
barbeque
lighters, the transient limiter provides a controlled gap, which
arcs over in a similar way when subjected to appropriate electrical
transients.
This action diverts the transient energy and associated voltages
and prevents damaging equipment and components, or causing harm to
personnel. |
Figure 2: Not a Reynolds product |
Reynolds has a broad line of transient limiters with the ability
to provide cost effective protection for modest or severe transient
events. Devices are available for general-purpose application in electronic
equipment, as well as special to type designs developed for particular
applications. The combination of high current carrying capability,
small size and low cost makes these devices unique among transient
protection components
Even though these components are conceptually very simple, it is useful
to have some familiarity with the less obvious aspects of their behavior
in order to gain an insight into application needs. The following notes
may help in this respect in providing some guidelines for consideration
when developing an application. |
Figure 3: An application opportunity |
What are Transients?
Voltage transients are generated by both natural
and man-made sources. Some of the more common sources are:
Atmospheric
The devastating effects of a direct lighting strike are well known
even though property damage is usually quite localized. Most electrical
damage, however, is not caused by the direct strike, but rather by
voltages induced in nearby conductors (power lines, telephone lines,
etc.) as a result of the direct strike. Harmful voltages have been
recorded on a line as far as 15 miles away from a strike. Appliances,
computers, telephone systems and other communications equipment can
all be extensively damaged by such events.
Static Discharge
Another familiar phenomenon, static electricity, can also cause harm.
There are numerous sources of static and the currents and voltages
involved in the discharge process can damage electronic equipment. |
Figure 4: Should have used a Transient Limiter |
Man Made Transients
Back EMF from inductive circuits, caused by the interruption of current
in motors, transformers, solenoids and relays, can produce transients
of several thousand volts and are very common. Voltages induced in
adjacent conductors by power line shorts, especially on 3 phase
systems, can be excessive. Accidental contact between power and signal
lines can result in large AC current flow on signal lines. Quick
acting protection must be provided for such occurrences to prevent
damage during the delay between the initiation of AC current and
the operation of traditional protective devices, such as fuses or
circuit breakers.
Electro-Magnetic Pulse (EMP)
Nuclear detonations produce extremely high electric fields that will
damage unprotected equipment over a radius of several miles. Reynolds
has a range of products with extensive field-testing to protect against
EMP. A Less well known phenomenon is the considerable electric fields
produced by conventional explosives which must also be protected against
in certain conditions. |
Figure 5: An early transient limiter |
How Does a Transient Limiter Work? -
The Basics
If two electrodes, a pair of metal thumbtacks for example, are arranged
such that their faces are close together, say 1mm (0.04”), and
then they are connected across a voltage source, they will serve as
a primitive transient limiter. All gas discharge transient limiters
are a variation of this simple scheme.
If the power supply voltage connected between the thumbtacks is slowly
increased, a number of interesting properties can be observed. These
can be conveniently pictured by considering the relationship between
current flow through, and voltage across the gap. The general form
of such a curve is shown in Figure 6.
Several of these areas to the left of this figure have little relevance
to typical transient limiter applications and are mentioned only for
completeness. Areas of particular importance to typical Reynolds product
applications are the glow region and the arc region at the right of
the curve. |
The curve can be broken into seven regions, each of which have distinct
characteristics. From point A to C (reproduced in Figure 7) the current,
a hundred trillion times less than a 100W light bulb, is proportional
to the applied voltage from zero volts to about 15 VDC. The current
that remains at this level even when the voltage is increased to 60
volts or so. This region has the grand name of Ionization Chamber
Region and the current is some times referred to as the dark current.
In this mode the gap can serve very adequately as a radiation monitor.
Figure 6: Generalized current/voltage curve for gas tube
transient limiter. Key application features are the breakage voltage,
the glow voltage and the arc voltage. |
|
The portion CDE of the curve is known as the Townsend region in which
ionization begins as the breakdown voltage is approached.
Region E to F is the breakdown region. In this region, the current
can change by a factor of one million or so while the terminal voltage
remains fairly steady. Note how different this is to the earlier situation
at B to D, where a fairly large voltage change produced little increase
in current. Very obviously, gas tubes do not obey Ohms Law.
Figure 7 |
Figure 8
|
F to G is the breakdown-to-glow transition and is one of two negative
resistance regions (wherein the current actually increases while the
terminal voltage falls). Negative resistances in excess of minus one
million ohms are frequently seen. Although the current here is still
very small for our purposes, it is worth noting that the humble spark
gap operates very effectively as a Geiger counter in this region. |
Figure 9: The Glow Region |
The G to H region is the glow region, familiar to everyone in the
guise of neon indicators on electric appliances. A neon tube, of course,
is a gas discharge tube designed to operate in the glow region. Every
gas discharge gap has a glow region, although the color and intensity
of the glow varies somewhat with the type and the pressure of the gas.
A notable feature of this region is that the voltage across the device
scarcely varies over a current range of 2000:1 or so. Because current
flow in this region is in a normal range for electronic circuitry at
100 uA to 100 mA, this phenomenon elevates the lowly neon to the status
of a widely used, low cost voltage regulator.
The glow voltage of a tube is relatively independent of the breakdown
voltage and is largely determined by the cathode material, the gas
type and gas pressure. Glow voltages for Reynolds tubes range from
around 70 volts for neon based mixtures up to 400 volts for nitrogen.
For transient limiter applications, a high glow voltage is generally
preferable, as will be described later. |
Figure 10: Beginning of Arc Region |
Next comes the abnormal glow region of H to I. This does not have
any special use and can actually be almost entirely suppressed by selection
of particular electrode geometries. Operating a unit in this region
is inefficient as a light source and, if sustained, will usually damage
the device through overheating due to the high voltage x current product
which, of course, define the dissipated watts.
The final region, I to J, is the arc transition region. Here, the gas
and electrode heating become intense giving the device the ability
to conduct thousands of amperes very effectively through thermionic
emission from the cathode. This is the normal region of operation for
transient limiters. The arc voltage is essentially independent both
of the breakdown voltage and the transient current. It ranges from
approximately 15 volts to 25 volts depending on the electrode materials,
fill gas and geometry of the device. Sustained operation in the arc
mode quickly consumes the life of the device due to excessive electrode
melting and subsequent deposition of the internal walls of the tube. |
Figure 11: DISASTER! |
An Actual Application - The Problem
Figure 12 shows a vacuum tube, which might be a CRT or TWT, with a
large voltage on its anode and a much smaller voltage at its grid,
supplied from a box of Expensive Electronics.
If the vacuum tube suffers an internal anode to grid flashover (a
not uncommon event) the high voltage energy, stored in the anode capacitor,
is connected to the grid momentarily through the arc inside the tube.
When such a flashover occurs, an arc is struck between the anode and
the grid presenting a low resistance path through which large currents
can flow. The voltage at the grid rises rapidly (typically 30,000 volts
per microsecond) towards the anode voltage of 15 kV. When it reaches
a thousand volts or so, 1/30th of a microsecond later, the grid control
electronics are destroyed. Although such a flashover often clears itself,
making the tube fit for immediate reinstatement, the transient at the
grid will destroy the Expensive Electronics rendering the entire assembly
unserviceable. |
The solution
What is needed is a device which can be connected between the grid
and the cathode without affecting normal operation, but which will
turn on very quickly in the presence of abnormal voltages. Such a
device is the transient limiter.
To meet this particular requirement, the limiter must
- withstand the
200 volts or so usually present at the grid without drawing current
- have
extremely low capacitance such as not to degrade the signal
voltages present on the grid
- be able to switch on within 20 nano-seconds
or so after a fault condition occurs
- have the ability to conduct
hundreds or thousands of amperes without difficulty
|
Figure 12: No Problem! |
Reynolds Transient Limiters possesses all of these attributes
Installing a transient limiter at the grid successfully protects the grid circuitry
and permits the tube to be put back on line immediately. Now when a flashover
occurs, the voltage across the transient limiter quickly exceeds the breakdown
voltage and, because there is little to limit the current flow, it transititons
to the arc region. The energy is safely removed from
the capacitor being dissipated in the resistance of the components
making up the circuit. The transient limiter itself does not dissipate
much energy since the terminal voltage remains low, at 15 to 25 volts.
Moreover, hundreds of dollars worth of circuitry have been protected
by the addition of the low cost limiter.
Anatomy of a Gas Tube Transient Limiter
Gas discharge tubes come in mainy sizes. One of the smallest components
made in the industry is the Reynolds MCQ series which is a mere 3/16” in
diameter and less than 1/8” long. At the other extreme, single
tubes occupy entire buildings and filled with exotic gases at the cost
of 3/4 million dollars per discharge, are currently being used for
Star Wars, nuclear fission and other physics studies. Reynolds gaps
are more modest than this and are widely used throughout the defense
and communications industry for equipment protection and personnel
safety. |
Figure 13: Construction of a typical tube assembly showing
all major components and indicating the physical gap. |
Figure 13 shows construction details of a typical Reynolds gas tube
transient limiter. The two metal electrodes are brazed into the ceramic
cylinder using noble metal preforms in a controlled atmosphere furnace
at about 900șC. The controlled atmosphere is captured inside the finished
component and becomes the gas fill. Thus the DC breakdown voltage of
a gas tube is fixed during manufacture, being a function primarily
of the electrode spacing, fill gas type and the fill pressure. External
details of the electrodes can be of almost any form with a common configuration
being wire leads as indicated in the figure. |
Environmentally, the components are extremely rugged and durable.
Some basics attributes are given below.
Typical Gas Tube Environmental Specifications
Operating and Storage Temperature Range –55șC to +130șC
Thermal Shock MIL-STD-202E, Method 107D, Condition B
Humidity MIL-STD-202E, Method 103B, Condition D
High Temperature Storage MIL-STD-202E, Method 108A, Condition D
Explosion MIL-STD-202E, Method 107D, Condition B
Seal MIL-STD-202E, Method 112B, Condition D
Shock MIL-STD-202E, Method 213 B, Condition F
Vibration MIL-STD-202E, Method 204C, Condition F
Terminal Strength MIL-STD-202E, Condition A, 1kG
A variety of fill gases are used in these components, including Argon,
Hydrogen, Neon, Nitrogen and various mixes of these, and fill pressures
range from about 1/30th of an atmosphere to four atmospheres. A small
amount of Tritium, a radioactive nuclide, is added to the fill gas
of many parts to improve performance characteristics – especially
consistency of DC breakdown. Because gap construction is a true hermetic
seal the internal parts are not subject to environmental or time related
deterioration. This makes the devices ideal for long-term protection
of sensitive equipment. |
Gas tubes dissipate a minimal amount of energy during operation,
unlike other device types including avalanche diodes and varistors.
The reason for this is that they function as switches, producing
only a small voltage (15 to 25 volts) across themselves even when conducting
thousands of amperes. The energy of the transient is predominantly
dissipated in other resistive elements of the circuit, including
the
external wiring. Other transient limiter device types work with considerably
higher terminal voltages and can suffer overheating with subsequent
failure when exposed to peak surge currents beyond their ratings.
Reynolds transient limiters provide three distinctive performance characteristics
not found in other component types:
- High insulation resistance – typically greater than 1000
megohms
- Low capacitance – typically less than 2 pF
- High current capability – in excess of 10,000 amperes
The Science of Gas Tubes
A basic gas tube transient limiter is a bi-directional device
in the sense that the voltage-current characteristic is
the same for
either
polarity of voltage. As mentioned previously, the breakdown
voltage of a gap is determined by the gap spacing, the gas
type and the
internal pressure. |
The internal geometry of a given gas tube, including the size of
the gap, is determined by the following:
- required range of breakdown voltages (90 to 10,000 volts covering
the range for Reynolds products)
- achievable mechanical tolerances
- realistic internal pressures
Physical gaps are thus usually between .50mm (0.02”) and 4mm
(0.160”). Gaps smaller than .50mm become increasingly
expensive to build, due to the need for selective assembly
to control tolerances
and individual post-braze adjustment. The larger gap
sizes in the range are desirable to maintain internal pressures
at a reasonable
upper limit.
The relationship between gas type, pressure and gap spacing
is shown in Figure 14. In principle, to make a gap of a particular
breakdown
voltage, it is only necessary to select that voltage on the
left-hand
scale, move horizontally to the right until one of the gas
curves is reached and then drop a vertical to find the gap
spacing-times-pressure
product. As an example, to design a 600 volt hydrogen gap,
select 600
volts on the left-hand scale, move horizontally to the hydrogen
curve and then vertically to the horizontal scale to find
the product of
29 kPa.mm. |
Figure 14 |
The Reynolds MLF style basic assembly has electrode spacing of 0.6mm
making the pressure required 29,000/0.6 or 49,000 Pa, slightly less
than 1/2 an atmosphere.
A gap made in this way would exhibit an average breakdown voltage close
to the design requirement of 600 volts, but would show fairly large
voltage variations from breakdown to breakdown (possibly as much as ± 20%).
Further, the impulse performance, that is the breakdown for a rapidly
rising
voltage, may be close to 2000 volts at a dV/dt of 30 kV per
microsecond. This is more than three times the desired DC breakdown
voltage!
So much for Science… |
The Art of Gas Tubes
Gaseous discharges and related phenomena have been extensively studied
in the western world for the past two centuries. Many aspects have
yielded to investigation, are well documented in the literature
and benefit either from formal scientific understanding or at least
reliable
rules-of-thumb. Many other elements, however, are not well understood
and continue as the subject of experimentation for new designs.
Included among these black elements are the following:
- electrode coating
materials used for extending life
- radio-isotopes to control initial
breakdown behavior
- exotic electrode materials for special purpose
applications
- field control techniques for enhancing impulse performance
- electrode
coating materials for reducing breakdown voltage
- polarizing techniques
for uni-directional applications
- use of wall effect to enhance recovery
time
- electrical conditioning of finished assemblies
In various combinations, these ploys enable breakdown voltages
to be reduced to 90 volts or so, impulse to DC breakdown
ratios to reduced to near unity, breakdown voltage scatter reduced to a few
percent
and so on. Most importantly, they provide a powerful toolbox
for tailoring device performance to meet individual customer
SCDs. |
Figure 15: Impulse or transient response performance of
several different tube types. |
Figure 15 shows how the apparent breakdown voltage of a gas tube
is typically influenced by high rates of voltage change in the transient.
Actually, the observed voltage increase is due to the transition time
required by the tube to reach the arc region.
The two exceptionally flat curves in the Figure are examples of transient
performance tailoring to meet specific requirements. These switches
transition to the arc mode in a few nanoseconds providing protection
to sensitive circuitry before damaging voltages develop. |
Application Caveats
In this short primer no attempt is made to develop detailed applications
for transient limiters. For the present purpose, some key parameters,
which are sometimes overlooked, are discussed with a view to clarifying
their relevance to an application.
Details of peak currents, waveforms shapes and lifetime operations
are clearly primary factors for specific installations. These parameters
are also usually easily resolved – often on a physical size basis
where large transient = large device and vice versa. Frequently, transient
waveforms are defined by some specification or standard in a generalized
form, and the “right size” component is easily established
from the Transient Limiter data book.
What is important for ALL applications however, are a some potential “Gotchas” which
should always be reviewed when specifying a gas discharge transient
limiter. These secondary aspects can easily make or mar an application
and are often not particularly intuitive. |
Impulse Breakdown
For some applications, the rate of voltage rise, dV/dt, across the limiter is
relatively low. For example, a power supply overvoltage crowbar may be expected
to see a dV/dt of a few thousand volts per second. In this case, the effective
protection voltage of the limiter will be very close to its DC breakdown
performance making the application straightforward in this respect.
A very different situation exists in a communication application wherein
normal operating levels are less than 50 volts and protection against
stimuli such as lightning or EMP is required.
Here, the dV/dt can be extremely high – up to many kV per nanosecond
in fact – and energy levels are somewhat unpredictable. The impulse
performance of the transient limiter is crucial in these latter cases
to ensure transition into the arc mode quickly enough before the voltage
can reach a damaging level. Very often, gas tube devices are used in
conjunction with other components to ensure complete protection in
these applications. Figure 15 again illustrates the reason why this
parameter is so critical here.
If a transient with a dV/dt of 10 kV per microsecond appears on a line
protected by a standard DCF-0230, 230-volt gas tube, the voltage would
rise to almost 900 volts before the tube gained control. Although the
time taken for this to happen is short – approximately 80 nanoseconds – the
equipment designer must be aware of this and take steps to deal with
it as necessary.
Reynolds tubes offer industry state-of-the-art tailoring of this parameter
on a case by case basis if required. |
Figure 16: Source recovery trajectories to be avoided |
Extinguishing Characteristics
Perhaps the most common application misunderstanding evolves around
the extinguishing characteristics of a gas tube limiter. Referring
back to the glow region of operation and the TWT grid circuit protector
mentioned earlier, a potential latch-up condition can occur which
will prevent proper operation of the circuit. Suppose that as the
flashover transient decays, the voltage current circumstances across
the tube intersect the glow region on the way back to normal operation.
For example, the circuit energy is reduced below the point where
an arc can be sustained but the grid circuit attempts to maintain
its normal output of, say 250 volts, with a few milliamps current
capability. The gas tube will recover as far as the glow region at
which point it will become a voltage regulator pulling sufficient
current from the circuit to limit the output voltage to the glow
voltage. Disaster. Within a minute or so, either the grid supply,
the gas tube or both will be destroyed. Obviously not desirable. |
What is needed in this situation is a tube with a glow voltage greater
than the grid supply voltage, which absolutely precludes this kind
of latch-up. Fortunately, there is a ready solution in the form of
Reynolds –2S and –3S multi-section gas tubes. In these
components, two or three (or more) tubes are factory assembled in series
with each other such that their individual glow voltage add to each
other. The MLH-500-2S tube, for instance, has a glow voltage of almost
400 volts making it an excellent choice for this application.
A similar latch-up condition could also apply to the power supply output
crowbar discussed above. The design of the supply, including details
of any fold-back current limiting, would need to be known to address
the need properly. On the other hand, in a typical communication circuit,
the extinguishing characteristics are seldom ever of concern. The signal
voltages encountered are usually well below the glow voltage of the
transient limiter precluding any possibility of latching.
In other situations, where intersection of the glow region by the recovery
vector cannot be avoided, it becomes necessary to power down the equipment
following a transient and then put it back on line a second or so later.
The gas tube will recover its full DC hold off capability within 1/10th
of a second and normal operation can resume. |
AC Follow On Current
This situation has some similarity to the extinguishing condition
just as described. In a typical AC limiter application, a transient
can
occur at any point on the voltage waveform. The worst-case situation
is when this happens shortly after the voltage has crossed zero
at the start of a half cycle. The gas tube enters the arc mode,
providing
a low resistance path to ground for the transient and reducing
the voltage at that point to 15 to 25 volts. Referring again
to the gas
tube current/voltage curve, notice that a current of just an amp
or so will maintain the arc and that a typical AC supply line has
no difficulty in supplying this much current (an electric toaster
draws about 10 amps from a 110 wall socket and scarcely reduces
the voltage at all). The “arced” tube however will
attempt to draw hundreds or thousands of amps in its attempt
to maintain
its low terminal voltage. This situation persists until the next
zero crossing occurs at up to 8.33 milliseconds later in a 60 Hz
system (10 milliseconds at 50 Hz). |
Careful consideration of this phenomenon is required to determine
whether to install some other additional components to limit the current,
rely on existing circuit breakers or fuses to disconnect the system,
or just “tough it out” by selecting a device large enough
to survive the event. Such a choice depends on many factors including
the “stiffness of the circuit, the desirability or otherwise
of disconnecting the supply and the cost of the transient limiter
itself. The important point is to be aware of this behavior when
selecting
a component.
End of Life Characteristics
A further consideration when reviewing an application is to properly
specify end of life. Commercial gaps, used in high volume on telephone,
cable television and similar distribution systems, often define this
in the following way:
- Breakdown Voltage – Breakdown voltage
less than 50% of nominal or greater than 150% of nominal shall
be considered
a failure
- IR (Insulation Resistance) – Less than 2 Megohm
- Impulse Voltage – Impulse
breakdown less than 50% of nominal or greater than 150% of
nominal shall be considered failure
These criteria are usually too liberal for the applications discussed
above and will normally require specification on a case-by-case
basis. This is the purpose of SCDs and only specially designed
devices usually
meet such requirements.
Reynolds has an excellent track record of demonstrated
expertise in meeting such needs and welcomes the opportunity
to tackle
new challenges
in this area. |
Glossary of Terms
AC discharge current n. the minimum RMS value
of 11 cycles at 60 Hz AC current that 95% of a given lot of Transient
Limiters can withstand
without entering a failure mode. Peak AC voltage shall exceed DC
breakdown voltage by at least 50%
AC follow current n. The current caused by the
normal AC operating voltage after the Limiter has been ignited
by a transient and which
continues to flow until the device extinguishes at the next zero
crossing of the AC current
Capacitance n. the capacitance between the terminals of a transient
limiter measured at 1 kHz
DC breakdown voltage n. the voltage at which an abrupt transition of
the gap resistance, from a practically infinite value (10,000 megohm)
to a relatively low value (nearly 0 Ohm) occurs, when subject to a
slowly rising voltage (100 Volts per second or less)
DC holdover voltage n. the maximum DC voltage across the terminals
of a Transient Limiter under which it may be expected to extinguish
and recover to the high impedance state within 150 milliseconds or
less after passage of a 500A, 10 x 1000us current surge waveform, when
tested in accordance with REA PE-80, paragraph 4.2.5, under specified
circuit conditions
Delay time n. in a 3-electrode tube is the time between the start of
conduction in one gap and the start of conduction in the other gap,
when tested under conditions specified in IEEE/ANSI C62.31
Failure Modes n.
- short-circuit failure mode – the Transient
Limiter becomes permanently short circuited
- low breakdown voltage
failure mode – the Transient
Limiter exhibits a DC or surge breakdown voltage of less
than 50% of the minimum breakdown voltage
- high breakdown voltage failure
mode – the Transient
Limiter exhibits a DC or surge breakdown voltage of greater
than 150% of the maximum breakdown voltage
- low insulation resistance
failure mode – the Transient
Limiter has an insul;ation resistance of less than 1 megohm
impulse breakdown voltage n. the maximum value of
voltage attained by an impulse of a designated rate of rise and polarity
applied
across a Transient Limiter prior to the device conducting
and clamping the
voltage. Due to the time required to ionize the inert gas
of a tube, the impulse breakdown voltage may substantially exceed
the DC breakdown
voltage. The impulse breakdown voltage is a function of the
risetime of a voltage wave and increases in value with increasing
rate
of rise. See Figure 15 in text.
Impulse Life n. the number of surges from
a 500A peak, 10 x 1000 vS surge waveform, both polarities,
that a lot of Transient
Limiter
can
withstand with 20% or less entering a failure mode
Insulation resistance n. the resistance between the terminals
of a Transient Limiter measured at 100V DC for devices with
DC breakdown
of 230V or more and 50V for devices with DC breakdown below
230V
Maximum single impulse discharge n. the maximum value of
peak current, with an 8 x 20 uS surge waveform, that 95%
of a lot
of transient
limiters shall withstand without entering a failure mode
Surge waveform designations n. the waveforms of a surge is
designated by two numbers, e.g. 8 x 20 uS. The first number
is the rise
time from 10% of the peak value to 90% of peak value. The
second number
is equal
to the rise time plus the decay time from the peak value
to one-half of the peak value
Transition time n. the time required for the voltage across
a transient limiter to fall into the arc region after the
gap initially
begins
to conduct |
|
|
|
