Background
Developing countries often have weak power generation and
distribution systems:
Frequently resulting in mains voltage fluctuations beyond the 184 to
264 vac input range mandated by the European Norm. The underlying IEC
standards assume that power generation and distribution are fairly stable and that voltage
excursions will not be outside the specified input window of 184 to
264 vac. Unfortunately, that is not the case in many developing countries
throughout the world. Even industrialized nations have their share of
problems with ac mains availability. Mains voltages frequently fluctuate
beyond the assumed 184 to 264 vac input range.
In the U.S. State of California for example, a failed deregulation
scheme coupled with the lack of corrective political will led to rolling blackouts that cut
power to millions in early 2001. The desired expansion of generation and
distribution capacity in developing countries is also constrained by a shortage of capital and, in some
cases, political factors as well. Thus, it appears that the quality of
ac mains, for developing countries, is at best uncertain into the
foreseeable future.
The expansion of communications networks, such as cellular radio,
direct radio access telephony (wireless in the local loop), paging
networks, land mobile radio and satellite telephony stations, has led to a rapid increase of installed sites to
support these networks. The proliferation of sites sometimes makes it
difficult to place a particular site in an ideal location. The site
might then be installed in a remote or difficult-to-access location.
The author of this article has worked with infrastructure
manufacturers and operators to address power-related reliability
problems, which can be very expensive if not addressed
properly--especially for remote and hard-to-reach locations. This paper
aims to give readers some helpful hints when addressing their own power-related
reliability concerns.
The problem
The power converters used in modern ac-dc rectifiers, UPS and
ac-dc
switchmode power supplies, are typically based on the IEC input range of
184 to 264 vac. Ac induction motors are sensitive to both high and low mains voltage, where
low voltage can cause rotor lockup following a dip in mains voltage; and to
saturation causing tripping of circuit breakers, as a result of high
mains voltage. Transformers and other 50/60 Hz magnetics can easily get
saturated when exposed to high input voltage, particularly at 50 Hz, when
the magnetization curve of steel becomes non-linear much quicker than
for 60 Hz.
Ac induction motors are sensitive to both high and low mains voltage.
Causes of power supply failure
 |
Sags
and undervoltages can cause
component overheating and destruction of MOS-FETs from over-current. |
|
|
Surges and
overvoltages
can cause component
overheating, destruction of MOS-FETs, or can trigger other electronic
components such as SCRs, MOVs and input capacitors may be destroyed if
they are rated too close to the line voltage. |
|
|
Component overheating
reduces the life and
deteriorates
the real reliability
as opposed to the estimated reliability based on steady-state
conditions of the product. |
|
|
False triggering
of other components can create nuisance
alarm tripping or, worse, can cause overheating or destruction of
other components. |
|
|
Typical EMI filters are not well damped. This has a dramatic
effect on any voltage disturbances, resulting in oscillations inside
the EMI filter under any transitional conditions.
Severe voltage
surges may result from fly-back
from
saturated inductors
looking for a path to release energy.
|
|
|
Boost converters can be destroyed by surges
causing
increased energy storage in the input filter, when the output capacitor is charged to
an unsafe level, depending on capacitance value and the load levels for the
DC/DC converter connected to the output of the boost converter. |
Causes of Semiconductor Failure
|
|
Most
semiconductor devices are intolerant to surge voltages in excess of
their voltage ratings. |
|
|
A fast surge (high dv/dt) of a few kV per microsecond can cause
a semiconductor to fail catastrophically or may degrade it so as to
shorten its useful life. |
|
|
Damage occurs when a high reverse voltage is applied to a
non-conducting PN junction. |
|
|
The PN junction may avalanche at a small point due to the
non-uniformity of the electric field. In this case, thermal runaway
can occur because of localized heat buildup and cause a melt-through
which destroys the junction. |
Mains voltages varying between 160 to 330
vac are common in countries
like India and Pakistan.
Mains voltages varying between 160 to 330
vac are common in countries
like India and Pakistan, but are also found in many other areas. In such cases,
installation of unprotected power systems and equipment can lead to an abnormally high
failure rate and loss of revenue for both suppliers and users of telecom hardware.
Mains voltage anomalies can be caused by the following:
|
|
Fault
clearing: where a short
circuit occurs on the line and a protection device clears the fault,
typically in around 3 cycles. The clearing of a fault results in a
short dip and voltage recovers rapidly if the mains are stable.
However, recovery can take much longer if the mains are weak,
resulting in a protracted voltage sag. |
|
|
Chronic
instability: caused by
interaction of voltage stabilizing devices. A good example would be an
industrial area with an inadequate incoming power line, thus resulting
in unstable mains as loads cycle on and off. Users of power may then
install a voltage stabilizer that will draw more current to provide a
higher voltage for one facility, with the higher current draw causing
a voltage drop for adjacent facilities that in turn draw more power
when regulators attempt to correct the mains voltage. The end result
can be a total loss of power when current draw exceeds the substation’s
capability to deliver the required current. |
|
|
Weak
generation capacity: will
also result in fluctuating mains when utility customers compete for available power. |
|
|
Rolling
brownouts: can result when
the power utility reduces the mains voltage by 10% or more in order to
limit power consumption. This approach may work for fixed-resistance
loads such as light bulbs, but will not for computer and telecom
equipment utilizing switchmode power supplies, as these compensate for
low voltage by increasing current draw. The increased current is
capable of generating increased internal heat that can lead to power
supply failure. |
|
|
Blackouts:
can result when the electric
utility cuts off power to selected areas in an attempt to protect
major parts of the distribution system from interruptions. |
Mitigation devices are often applied as a "bandage," in the
hope that things will get better quickly, to the satisfaction of both
the telecom operator and the equipment provider. Unfortunately, applying mitigation devices without understanding the
basics can actually aggravate the problem. This is especially true for
power-factor-corrected power supplies, as these are more sensitive to
mains disturbances than their simpler siblings with rectifier/ capacitor
inputs. It is the author’s experience that power-factor-corrected
rectifiers can trip internal protection devices when exposed to rapidly
changing mains voltages.
Overvoltage protection (surge protection) against surge voltages,
and from load switching and lightning, is governed by rational
principles, first outlined by IEC-1312 in the early 1990s. The
applicable IEC-Standard: IEC Technical Specification 61312-3
"Requirements of Surge Protective Devices (SPDs)" describes
clearly how to coordinate SPDs to ensure that upstream devices
(Figure 1 in pdf of diagrams) progressively handle more energy than the SPD found on the
front-end of a power supply, typically a metal oxide varistor. These
sensible coordination principles are not always adhered to, which sometimes
results in a lack of coordination between SPDs used in a system. This
can lead to the destruction of downstream SPDs that clamp before the
upstream SPDs, which in turn causes the destruction of the metal oxide
varistors in the power supply itself.
Coordination of surge
protection devices.
Safety earthing systems
Many different earthing systems are in use throughout the world and
need to be understood before primary surge protection is implemented. The most
common earthing systems are shown below:
TN-S
A separate neutral and ground conductor are run throughout. The
ground conductor can be the metallic sheath of the power distribution cable or a separate
conductor. All exposed-conductive-parts of the installation are connected to ground.
TT
One point of the source of energy is grounded and the
exposed-conductive-parts of the installation are connected to independent grounded electrodes.
TN-C
Neutral and ground conductors combine in a single conductor
throughout the system. All exposed-conductive-parts are connected to the combined N-G
conductor.
IT
A system that has no direct connection between live parts and ground.
All exposed-conductive-parts of the installation are connected to independent
grounding electrodes.
TN-C-S
Neutral and ground combine in a single conductor. This system is also
known as multiple earthed neutral (MEN). The protective conductor is
referred to as the combined neutral and ground conductor. The supply N-G
conductor is grounded at a number of points throughout the network and
generally as close to the consumer’s service entrance as possible. All exposed-conductive-parts are connected to the combined N-G
conductor.
Possible solutions
The most obvious solution is to design power supplies to withstand a
much wider input voltage range.
The most obvious solution is to design power supplies to withstand a
much wider input voltage range, to prevent saturation and resulting
energy impulses that cause damage to switching transistors and other circuits. Such changes require more
steel in inductors, capacitors with higher voltage ratings, possibly
larger circuit boards with greater trace separation, as well as transistors with higher voltage and current
ratings. This is not likely, as manufacturers would have to offer
special versions with a wide input voltage range or, as a second option,
make only power systems that can handle all worldwide markets,
regardless of mains quality. This second option is not realistic since
customers in countries with a stable mains infrastructure would be penalized.
Designing a protection device that can handle an entire power plant,
or racks of power systems, is probably a more economical solution that
can be used whenever increased immunity is required. Such a solution can
also be used to address problems that occur in the field after installation.
Considerations for
mitigation devices
Mitigation devices may include a complete solution, or components
thereof, if system analysis reveals that essential elements already
exist. We will present here a complete solution for the readers’
consideration.
Input range of 160 to 330
vac at 50/60Hz without saturation at 50Hz.
Requirements:
|
|
Input range of 160 to 330
vac at 50/60Hz without saturation at 50Hz. |
|
|
Output range of 184 to 264
vac |
|
|
Overvoltage protection (OVP) with cut-off if output exceeds 264
vac |
|
|
Automatic bypass in case of electronics failure |
|
|
Low impedance |
|
|
Coordinated surge protection |
|
|
Conversion to TN safety earthing system |
|
|
Low weight |
|
|
Reasonable cost |
|
|
High power efficiency and low heat loss |
|
