|
 Part
4: Immunity—Not for Circuitry
by George
Novacek
Start
• Interference Levels
• Let the Lightning Strike
• Ready To Go? • Sources
and PDF
LET THE LIGHTNING
STRIKE
EMC tests are performed
in an anechoic chamber with equipment so expensive
that only a handful of laboratories in the world
can perform them. When it comes to lightning tests,
even fewer labs are available (e.g., perhaps four
in the U.S.). Section 22 describes Lightning Induced
Transient Susceptibility. The indirect lightning
effects tests are divided by the method of injection
and the effects they cause.
The pin-injection test,
also known as the damage tolerance test, represents
the first battery of tests. A lightning generator
is sequentially connected to every connector pin,
and the pin is zapped with a number of defined pulses.
In the end, the system must have suffered no permanent
damage.
The circuit is protected
by spark arresters (spark gaps) and transzorbs.
When sufficient impedance can be placed in series
with the transzorb, the first stage with the arrester
may be omitted. In commercial equipment, MOVs are
used at the first stage, although they have not
gained much popularity in airborne systems because
of their alleged deterioration with age and number
of strikes. Zener diodes are similar to transzorbs,
although they are not suitable for transient protection
because of their slower response and a lower power-handling
capability.
Bulk injected tests
are functional upset tests. The lightning generator
discharges into a toroidal transformer, through
which the interface cable bundle is routed, inducing
common-mode signals. The pass/fail criterion is
the same as with HIRF susceptibility.
Several different waveforms—single
pulses of various shapes and pulse bursts—are used
to zap the system with pin-and-bundle injection.
A series of random pulses check whether software
execution is affected when the pulses coincide with
the clock.
A designer’s main concern
is the pulse burst, which is a series of 10 randomly
spaced bursts of 2-MHz frequency over 2 s. When
such transients reach the protected circuit and
exceed its firing level, the transients may be rectified
and the system can no longer operate.
In essential systems,
the system reverts to safe mode. In critical systems,
if the controller is allowed to freeze for up to
2 s, you can detect an implausible signal and maintain
the system until the signal is validated, which
is sometimes possible on systems with large inertia.
On faster systems where a 2-s freeze is unacceptable,
you must either lower the threat level reaching
the controller I/O by improved shielding or symmetrical
interfaces, or fix the problem by increasing the
upset level through higher common-mode range, higher
component voltage ratings, additional filtering,
or transformer interfacing.
Lightning direct effects,
addressed in Section 23, applies to externally mounted
equipment and the effects of direct lightning current
flowing through this equipment. For instance, a
lightning current discharging through the envelope
of fuel tanks causes them to be heated up. Generally,
you don’t have the same danger with electronic control
systems, with the exception of antennas or sensors,
where the exposed metal parts are connected directly
to electronic circuitry. In most cases, vendors
of such components provide adequate direct-lightning
protection as a mandatory part of their design.
THE ELECTROSTATIC
ZAP
I’ll skip over the
effects of icing, described in Section 24, because
it applies to moving parts that are subjected to
a build up of ice (e.g., aircraft landing gear).
These issues are primarily in the mechanical designer’s
domain.
Instead, let’s focus
on Section 25, the test for Electrostatic Discharge.
The ESD test determines the ability of the equipment
to perform its intended function without degradation
of performance. It does not address the immunity
of our equipment to an ESD-inflicted damage. Although
important, if you did a good job designing your
system to meet the EMC and indirect-lightning effects
requirements, ESD damage tolerance is inherently
provided for. You just have to verify a functional
upset due to discharge current through the cabinet.
In the test, you have
an ESD generator with a probe. Cranking up the generator,
you move the probe towards the operating unit until
the charge is released by touching the cabinet or
a spark jumps over. We repeat this test 10 times
with each polarity of the charge on each wall of
the unit. The unit must exhibit no functional upset,
which should be a fairly safe bet, if you successfully
passed the previous
EMC and lightning tests.
PREVIOUS
• NEXT
Circuit Cellar provides
up-to-date information for engineers. Visit www.circuitcellar.com
for more information and additional articles.
For subscription information, call (860) 875-2199,
subscribe@circuitcellar.com
or subscribe
online. ©Circuit Cellar, the Magazine for Computer
Applications. Posted with permission
|
TESTING
1, 2