During the Carrington Event solar storm of September 1, 1859, northern lights were reported as far south as Cuba. British astronomer Richard Carrington witnessed the megaflare on the sun through his telescope and was the first to realize the link between activity and geomagnetic disturbances on Earth.
The geomagnetic disturbances were strong enough that year that U.S. telegraph operators reported sparks leaping from their equipment and some utility wires caught fire and were severed.
A less powerful solar event occurred in March 1989, striking Canada and bringing down the Hydro-Quebec power grid. This solar flare, although less extreme, illustrated the vulnerability of the grid.
Solar storms can affect the Earth in three stages. First, high-energy sunlight, mostly x-rays and ultraviolet light, ionizes Earth's upper atmosphere, interfering with radio communications. Next comes a radiation storm, potentially dangerous to unprotected astronauts. And finally comes coronal mass ejection (CME), a slower moving cloud of charged particles that can take several days to reach Earth's atmosphere. When a CME hits, the solar particles will interact with Earth's magnetic field to produce powerful electromagnetic fluctuations. These fluctuations induce a field in any and every circuit, powerline, and antenna.
On March 5, 1989, about 25 years ago, a spectacularly large sunspot came into view. The very next day this sunspot produced a major solar flare. This solar flare, although less extreme than the Carrington Event, illustrated the vulnerability of the grid. Luckily, this flare was on the side of the sun – as we view it – and not pointed in our direction.
At 2 p.m. EST on March 10 1989, another major flare erupted. At 1 a.m. EST on March 13, the first wave of subatomic particles arrived at the Earth's magnetic poles, producing an incredible northern lights display that lasted almost two days. Voltage oscillations from the flare caused tripping of protective equipment, striking Canada and bringing down the Hydro-Quebec power grid, and nearly bringing the Northeast Power Coordinating Council (NPCC) and the Mid-Atlantic Area Council (MAAC) down in a cascading collapse. Circuits also overloaded in Great Britain, New York, and Virginia. A critical transformer melted in New Jersey.
The aurora of March 13, 1989, taken from Sea Cliff, NY. Photo: 
1989 Ken Spencer
On April 2, 2001, the sun unleashed the biggest solar flare ever recorded, as observed by the Solar and Heliospheric Observatory (SOHO) satellite. The flare was definitely more powerful than the one of 1989. It hurled a CME into space at a speed of roughly 7.2 million kilometers-per-hour. Luckily, the flare was not aimed directly towards Earth. The flare and solar ejection generated a storm of high-velocity particles NOAA rated as a moderate S2 to S3, on a scale that goes to S5. The solar flare produced an R4 radio blackout on the sunlit side of the Earth. An R4 blackout is second to the most severe classification with a flux of more than 0.001 W/m2.
Why Your Design May Be Vulnerable
There are two facets you need to consider because of solar flare voltage surges. One, you may very well lose power. Those long electric transmission lines are perfect antennas for this type of assault. This power loss may be over an extended time if many HV transformers are damaged – a likely situation.
Two, you need to protect your design from voltage surges. With any flare, its light (photons) takes about eight minutes to reach earth. This has little effect on an electrical system. It's the stuff that comes afterwards (from 6 hours to 3 days) that may deliver a crippling blow. Also, keep in mind that solar flares are not five-second sun bursts. A flare unleashed Feb. 24, 2011, that just missed us, lasted 90 minutes. If you get a warning about a flare, it may be best to turn off the main circuit breaker.
The design engineer, on the other hand, must be concerned about protecting the power input, the data/analog inputs, and the data/analog outputs. He/she may also want to have ESD protection on pushbuttons and other operator interfaces.
Surge protection devices may be categorized as voltage-limiting devices such as gas discharge arrestors, metal oxide varistors, suppressor diodes, triacs, diacs, and switches or current-limiting devices like fuses, circuit breakers, and thermal cutouts. Let’s take a look at voltage protection, as this is more pertinent to flare protection.
Wikipedia offers a good overview of the basic transient voltage suppressor devices:
| Type | Surge Capability (typical) | Lifetime - Number of Surges | Response Time | Shunt Capacitance | Leakage Current (approx.) |
|---|---|---|---|---|---|
| TVS-diode | 1A (small surface-mount device) to 15kA (large through-hole device) | ? | ≈ 1ps (limited by pin lengths) | < 1pF (small surface-mount device) to > 10nF (large through-hole device) | 1μA |
| Metal-oxide varistor (MOV) | Up to 70kA | @ 100A, 8x20µs pulse shape: 1,000 surges | ≈ 1ns | Typically 100–1,000pF +++ | 10μA |
| Avalanche diode, Zener diode | 50A | @ 50A, 8x20µs pulse shape: infinite | < 1μs | 50pF | 10μA |
| Gas discharge tube | > 20kA | @ 20kA, 8x20µs pulse width: > 20 surges | < 5μs | < 1pF | < 1nA |
Specialized ICs Offer Advantages
An important resource, in addition to the devices in the table above, are specialized ICs such as the ones offered by Maxim and a number of other suppliers. These chips improve on the basic devices characteristics – all of the basic devices, that is, except the gas discharge tube (GDT), which may always be needed for high voltage/current impulse suppression.
One good example IC is the MAX17525. This adjustable power limiter provides overvoltage, undervoltage, and over-current protection, and reverse-polarity safeguard. It handles up to 40V and 4.2A. Undervoltage lockout adjustment range is 5.5V to 24V - so you can’t use this device with lower voltage designs. Overvoltage trip range is 6V to 40V.
Forward current limit is settable from 0.6A to 6A with ±15% accuracy. The current limit setpoint can be raised by 1.5x or 2x during startup to charge downstream input capacitors. When used with an optional external pMOS FET, the chip will protect downstream circuitry from voltage faults up to 60V and add reverse current-blocking protection. The device features integrated FETs with 31mΩ on-resistance and is available in a 20-pin (5mm x 5mm) TQFN package. It operates over -40°C to 125°C. The MAX17525EVKIT is the evaluation kit for this IC.
Fig. 2: The MAX17525 high-accuracy, adjustable power limiter
The MAX14670–MAX14673 overvoltage protection ICs are similar to the MAX17525, but come in a variety of lower trip voltages. The MAX14670 has an overvoltage lockout of 6.8V, the MAX14671 is 15.5V, the MAX14672 is 5.825V, and the MAX14673 is 22V. All setpoints are ±3%, and all are adjustable with external resistors.
The chips provide soft-start to minimize inrush current, feature 65mΩ internal N-FETs, and offer 4.5A continuous current capability. In their very small 15-bump, 1.6mm x 2.1mm WLP package they are great for smartphones and tablets. The ICs are specified over -40° to 85℃. The MAX14670 and MAX14671 come in a 15-bump WLP package and the MAX14672 and MAX14673 in a 10-pin TDFN package.
Fig. 3 MAX14670–MAX14673 package types
A good choice for protecting signal lines is the MAX9940. This IC provides protection for digital communication lines such as 1-Wire and I2C protocols. Signal lines can't afford big series resistance or high capacitance to filter transients. This chip's input capacitance is just 38pf when the switch is on, and when it’s off, 27pf. It works from a 2.2V to 5.5V supply, drawing just 21µa maximum, 13µa typical.
Fig. 4: The MAX9940 signal-line overvoltage protector
The chip has a series switch with an on resistance of 77Ω maximum (Typ 38Ω). Insertion loss is near zero until about 3MHz and reaches 3db at about 25MHz. The internal switch is turned off when disabled or if a fault condition exists, isolating the connected device—often a microcontroller—from any possible damage. The output is protected up to 28V and has up to ±4kV IEC 61000-4 contact ESD protection. The device has a fault reaction time of 60ns and operates over the -40° to 125°C automotive temperature range.
Many Choices to Protect Your Circuitry
These days, designers have a large number of excellent choices for protecting their circuitry. Advanced ICs are available as alternatives to, or to use in conjunction with, basic protection devices and are definitely worth considering. One difficulty is in finding the right IC on manufacturer or distributor websites as they may be classified as "IC/special" or "IC/Analog" or "Power/Protection" or "Circuit Protection" or many other types – so be sure to check thoroughly.