Don Tuite

The Use And Misuse Of Circuit Analog/Power Editor

Protection Devices

C
ircuit protection devices aren’t Normal condition Trip condition
the most glamorous part of
an engineer’s design. They’re Temperature rises during
usually an afterthought to the short circuit condition
processors, FPGAs, memory, and other com-
ponents, which most often are considered the
real “meat” of the device. But without circuit
protection, the entire product may be vulner- Temperature drops
able to damage and can pose safety issues after circuit resets
for consumers.
The fuse is the most basic circuit protec-
tion device. When the current drawn by the Polymer has Polymer expands
a crystalline structure; due to I2R heating;
load rises above a certain point, it opens the conductive path is made conductive path breaks down
line carrying that current. A circuit breaker of carbon-black materials due to polymer expansion
is like a resettable fuse. Typically, it takes a
defect in the circuit being protected to “blow” 1. In a polymer fuse device, the conductive particles form low-resistance networks at
(open) a fuse. But if the fuse opens fast normal temperature. If the temperature rises above the device’s switching temperature
enough, it should help protect parts of the (Tsw), either from high current through the part or from an increase in the ambient
circuit that are not defective by sacrificing temperature, the crystallites in the polymer melt and become amorphous. This creates
itself. Alternatively, a fuse may open because an increase in volume that separates the conductive particles, resulting in a large non-
an excessive voltage is applied either by linear increase in resistance. After the fault clears and the polymer material shrinks,
mistake or because of a transient voltage on these devices reset themselves.
the power supply line.
Fuses that have opened must be re- component overheating and reduce the within the ceramic fuse body. An alternative
placed. (Breakers have to be reset.) Often, severity of arc-flash events. is the self-supported “wire in air” element,
equipment needs to be protected during “Fast-acting” means the fuse clears very which provides consistent fusing and cutting
short-duration overvoltage transients without quickly—as fast as a millisecond at high cur- characteristics and is better able to withstand
human intervention afterward. That’s where rents. Note that “high-current” caveat. With high inrush currents.
different types of transient voltage suppres- lower currents, the fuse may take longer to There are also multi-layer designs that
sion (TVS) come in. open. Check the fuse’s data sheet to see expose more fuse element surface area to
what trip times are guaranteed or “typical.” the fuse-body’s glass-ceramic absorption
Types Of Fuses Smaller, fast-acting chip fuses help provide material. With multiple layers, when the fuse
Consider different kinds of fuses and TVS overcurrent protection for systems using dc elements open, there is more material for the
devices. Most designers are familiar with the power sources up to 63 V dc. Conversely, vaporizing fuse metals to absorb into, result-
cylindrical fuses used in external ac-supply slow-blow fuses help minimize the nuisance of ing in a very efficient and effective quenching
lines. Surface-mount devices, which are sol- repeated replacements when a circuit experi- of the fuse arc.
dered directly to circuit boards, come in fast- ences brief but recurring overcurrent spikes. Pulse-tolerant fuses have strong arc
acting and slow-blow (time-lag) varieties. Fast or slow, chip fuse design is far suppression characteristics and withstand
Fast-acting fuses are employed for user from trivial. The most common fuse type, high inrush currents. They are used in high-
safety on critical ac equipment that requires employed for secondary-level overcurrent performance consumer electronics such as
frequent opening for maintenance opera- protection applications, uses a corru- laptops, multimedia players, cell phones, and
tions. They can help prevent conductor and gated (zig-zag) wire element supported other portable devices.

Sponsored by
 Monolithic, multilayer-design, high-
Normal operation
current-rated chip fuses target overcurrent
protection of power supplies, servers,
communications equipment, voltage regula-
tor modules, and other small, high-current
applications. Rc

Resettable PPTC “Fuses”

Beyond fuses, there are resettable circuit Rptc
protection devices that behave like fuses.
They are made from a composite of semi- Fault operation

crystalline polymer and conductive particles

(Fig. 1). These devices can be employed
to protect electric motors and transform-
ers from damage caused by mechanical
overloads, overheating, stalls, and other
potentially harmful conditions. They can also
be used to help protect USB ports, telecom
interfaces, and many other power inputs Leakage current
and outputs.
Known as polymeric positive tem- 2. During normal operation of the MHP-TA device, current passes through the bimetal
perature coefficient (PPTC) devices, they contact due to its low contact resistance. During an abnormal event, the device reacts
help protect against damage caused by to the rise in cell temperature, causing the bimetal contact to open at the specified
overcurrent surges and over-temperature temperature and its contact resistance to increase. The current then shunts to the
faults. Like traditional fuses, they limit lower-resistance PPTC, which acts as a heater and helps keep the bimetal protector
the flow of dangerously high current dur- open and in a latched position until the fault is removed. A typical MHP-TA device has
ing fault conditions. Unlike fuses, they temperature characteristics with a 77°C thermal cutoff (TCO) rating. These resettable
reset after the fault is cleared and power TCOs can handle the higher voltages and battery discharge rates found in high-capacity
to the circuit is removed. The tripping LiP and prismatic cell applications.
point depends on current and ambient
temperature. temperatures, high hold-current ratings, peak voltages can be very high, up to 20 to
and compact size that PPTCs alone can’t 30 kV. Transient-voltage protection devices
Resettable Thermal Cutoff/MHPs always achieve (Fig. 2). include gas-discharge tubes, multi-layered
Another type of circuit protection device, varistors (MLVs), TVS diodes, thyristors, and
specifically for lithium battery cell protec- ESD Devices metal-oxide varistors (MOVs). TVS devices
tion, is the resettable thermal cutoff (TCO) While fuses and similar devices help may either clamp (voltage-limit) or crowbar
device. With battery designers increasingly protect against overcurrent faults, several (short to ground) voltage transients that
using large-surface-area, high-capacity, other types of devices can help protect appear at their inputs.
envelope-shaped lithium-polymer (LiP) cells against electrostatic discharge (ESD).
in tablets and ultra-thin notebooks, they These devices can be zener or TVS diodes, TVS Diodes
need thinner TCOs for battery protection. varistors, or polymer devices. ESD events TVS diodes offer protection up to 30 kV.
Metal Hybrid PPTC devices with thermal are very low-energy, short-duration events They are essentially Zener devices with a
activation (MHP-TA) combine low TCO lasting a few tens of nanoseconds. The cross-section that accommodates large
1.0 1.0 1.0
0.9 0.9 0.9
0.8 0.8 0.8
I OUT/I OUT-peak

0.7 0.7 0.7

V OUT /V OUT-peak

V OUT /V OUT-peak

0.6 0.6 0.6

0.5 0.5 0.5
0.4 0.4 0.4
0.3 ESD 0.3 Burst 0.3 Surge
0.2 0.2 0.2
0.1 0.1 0.1
0 0 0
0 10 20 30 40 50 60 70 80 0 10 20 30 40 50 70 60 80 90 100 0 10 20 30 40 50 60 70 80
(a) (b) (c)
Time (ns) Time (ns) Time (µs)

3. The IEC1000 ESD test simulates the ESD of a human onto electronic equipment (a). The burst test simulates switching transients
due to relay contact bounce or the interruption of inductive loads (b). The surge test simulates transients resulting from lighting
strikes or the switching of power systems including load changes and short circuit switching (c).

Sponsored by
 VCC
VCC
R1 ally exceeds a voltage where avalanche
10k
1
0.1μF multiplication occurs, the impedance of the
RxD R VCC 8
TVS device begins to decrease, and current flow
2 7
MCU RE A begins to flow through it. When that current
XCVR
3 6 falls below a certain value, the thyristor
DIR DE B

TxD
4
D GND
5 becomes a high-impedance device again.
Gas discharge tubes (GDTs) are placed
10k
(a)
R2 on power lines, communication lines, signal
lines, and data transmission lines to help
VCC protect sensitive electronics from transient
VCC
10k R1 TBU1 surge voltages caused by lightning strikes
0.1 μF
RxD
1
R VCC
8 and equipment switching operations. These
TVS MOV1
2 7 devices exhibit high impedance in normal
MCU RE A
XCVR operation. But in the event of an overvolt-
3 6
DIR DE B age surge, such as a lightning strike, the
4 5 MOV2
TxD D GND gas ionizes and diverts the transient energy
10k R2 TBU2
to ground. After the fault condition has
(b) cleared, the GDT then returns to its high-
impedance state.
4. Both of these circuits are designed for 10-kV ESD and 4-kV EFT transient protection.
The circuit in (a) provides surge protection for transients of up to 1-kV. The circuit in (b) Thermal Circuit Protection
can handle surge transients of 5 kV and more. Devices/RTPs
Not all devices help protect against
reverse breakdown current flows. There are MLVS And MOVs overcurrents or overvoltages. Some protect
single-channel and multi-channel silicon TVS devices do not have to be built using against heat, regardless of what causes it.
ESD (SESD) TVS devices that may be unidi- silicon technologies. Probably the most fa- Reflowable Thermal Protection (RTP) devices
rectional or bidirectional. miliar is the MLV. MLVs are stacks of MOVs, help protect systems from thermal runaway
Bidirectional devices exhibit lower ca- which, in turn, are arrays of zinc-oxide balls damage due to extreme environments or
pacitance and insertion loss. They can be in a ceramic matrix. failed power components. They open if
placed on a printed-circuit board (PCB) in In an MOV, each boundary is a diode their internal junction temperature exceeds
any orientation, and they don’t clip signals junction. At each side of the array, there is a certain specified limits.
that swing below ground. Unidirectional metal connection. One of these is connected RTPs are surface-mount devices because
devices exhibit lower negative break- to the line being protected, the other to they need to optimize thermal coupling
over voltages. Multi-channel arrays make it ground. Thus, an MOV is an array of back- with the PCB. After installation, a one-time,
easier to protect multi-line data buses such to-back diodes that normally provide no electronic arming process results in the
as Ethernet. complete path through the array. Numerous device becoming thermally sensitive. Arming
There are also polymer-enhanced, breakdowns occur when a surge occurs, can occur during manufacturing tests or
precision-Zener-diode micro-assemblies, and the line is connected to ground for the in the field. (Before the arming procedure,
consisting of a precision Zener diode and a duration of the surge. the device can go through installation
PPTC that are thermally bonded, that can Stacked into an MLV, the MOVs are temperatures up to 260°C without opening.)
offer resettable protection against multi-watt essentially in series. MLVs have a higher After arming, the device will open when the
fault events without the need for multi-watt in-circuit resistance than a single MOV and junction exceeds the open temperature.
heatsinks. The Zener diodes used in these faster response time. In high-data-rate
micro-assemblies are selected for a flat circuits, they may degrade or distort the Beyond Fuses: Transient
voltage versus current response. This helps signal due to high capacitance. Adding Protection Basics
improve output voltage clamping, even PPTC technology helps reduce current prior Many ICs that interface with the world
when input voltage is high and diode cur- to contact switching events. beyond their own circuit boards are
rents are large. fabricated with internal circuit protection
As assembled, the Zener diode is in Thyristor Devices And Gas devices. This raises the question of whether
series with and thermally coupled to the Discharge Tubes external protection (and proper use of
PPTC layer, which responds to either ex- Thyristor surge protection devices have personal shoe straps) is really necessary.
tended diode heating or overcurrent events four layers of alternating conductivity: PNPN. The short answer is that if the application
by tripping, as described above. Polymer- The four layers are designated the emitter exposes the device to the outside world in
enhanced Zener diodes help protect circuits layer, the upper base layer (cathode), the mid- any way, then precautions against transient
in portable devices from inductive voltage region layer, and the lower base layer (anode). events are necessary.
spikes, voltage transients, incorrect power When a transient voltage is applied to a To get an idea of the energy in even the
supplies, and reverse bias. thyristor, as the voltage increases, it eventu- most modest transient events, consider

Sponsored by
 Volts
transients off the PCB to the greatest
extent possible.
• Design for multiple layers with full VCC
Power and ground planes to provide low-induc-
tance paths for the transient currents to
Ground return to their source.
Time • Because high-frequency currents follow
the path of least inductance, not neces-
sarily the path of least resistance, be sure
to position the protection components in
line with the signal path. This prevents
5. Injecting a large surge current into the circuit ground through the TVS limiter will the transient currents on the signal path
cause more or less ground bounce, depending on the connection inductance. Good to the protection device.
board design—that is, a low-impedance ground plane with good high-frequency • Place 100- to 220-nF bypass capacitors
decoupling on the power bus—will maintain an essentially constant potential difference as close as possible to the VCC pins of
during the ringing. In that case, the transient should not affect the behavior of the the transceiver, UART, and controller
system. ICs. Use at least two vias for VCC and
ground connections of bypass capaci-
the test standards used to evaluate circuit direct strikes or voltages and currents tors and protection devices to minimize
protection from various types of events, induced by indirect strikes), or simply by the effective via inductance. Use 1k to 10k
from a simple static discharge to power- switching of power systems with resulting pull-up/down resistors to enable lines to
line transients. The International Electro- load changes and short-circuit switching. limit noise currents in these lines during
technical Commission (IEC) recognizes At lower levels (500 V to 1 kV), these transient events.
three major types of transients: ESD, transients occur in industrial automation. At • If the TVS clamping voltage is higher
electrical fast transients (EFTs), and surge higher levels (5 to 6 kV), they occur in power than the specified maximum voltage
transients (Fig. 3). The IEC61000-4 family grid systems. In either case, due to the high of the transceiver bus terminals, insert
of electromagnetic compatibility (EMC) pulse energy, surge testing is usually limited pulse-proof resistors into the A and B bus
standards specifies IEC61000-4-2 for ESD to five positive pulses and five negative pulses lines. They will limit the residual clamping
immunity, IEC61000-4-4 for EFT immunity, with a one-minute pause between pulses. current into the transceiver and prevent it
and IEC61000-4-5 for surge immunity. Those test requirements illustrate the from latching up.
The first of these, the ESD test, simu- different levels of protection needed. • TVS protection alone is sufficient for
lates the ESD of a human onto electronic Generally, the on-chip IEC-ESD protection surge transients up to 1 kV, but higher
equipment. The standard describes an ESD provided by manufacturers of various kinds transients require MOVs, which reduce
generator that generates pulses of less of bus receivers can absorb the energy of the transients to a few hundred volts of
than 100-ns duration and 1-ns rise time. single ESD and EFT pulses, but it breaks clamping voltage, along with transient
A sequence of 20 discharges of positive down in the face of EFT pulse trains and blocking units (TBUs) that limit transient
and negative polarity is to be applied with a surge transients, even if the transient volt- currents to less than 1 mA.
one-second pause between pulses. age is not very high.
The second, the EFT or burst test, The energy that’s in pulse trains does Non-destructive ESDs And
simulates switching transients that often are not give chipmakers’ internal protec- Signal Integrity
encountered in industrial environments, such tion circuits time to recover. Instead, the What does the TVS protection device do,
as factory automation and process control. successively applied packages of energy beyond shunting spikes to ground to actu-
They may be caused by relay contact bounce are converted into thermal energy—heat ally maintain circuit functionality? That is, in
or the interruption of inductive loads. that breaks down the protection cells, a digital circuit, how is it possible to avoid
Testing requires a generator that destroying the internal transceiver input or false signal transitions caused by ESDs?
produces a burst of test pulses. Each burst output circuit. Practical protection requires Due to circuit inductance, the real effect
provides roughly 15,000 transients with external circuitry (Fig. 4). of the current transient being bypassed to
rise times of around 5 ns. A test sequence ground is that the discharge pulse will ring
comprises six bursts of 10 seconds each Circuit Implementation (Fig 5). That is less severe than the tran-
with a 10-second pause between bursts. Both ESD and EFT transients have a wide sient step, but it could still cause a missed
This adds up to several million pulses ap- frequency bandwidth. Generally, frequency or false transition on the signal line. How-
plied to the device under test (DUT) during components start at around 3 MHz and ever, the ringing will be reflected on both
one minute. reach as high a 3 GHz. Therefore, high- the ground and power busses, and external
Both ESD and EFT events involve relative- frequency layout techniques are essential: signals input to the protected circuit will
ly low energy levels. In contrast, the surge reflect that ringing. So, signal transitions
test simulates much larger transients that • Place protection circuitry as close to the referenced to the bouncing ground should
might be caused by lightning strikes (either bus connector as possible to keep noise be masked. n

Sponsored by