A non-arcing lightning arrester is a surge protection device that diverts lightning-induced overvoltages to ground without sustaining a power-frequency arc after the surge passes. It solves the follow-current problem — early spark-gap arresters would conduct line current after firing and burn themselves out. Modern non-arcing designs use metal oxide varistor (MOV) blocks or sealed gap-MOV combinations that reset to a near-infinite resistance once voltage falls back to normal. The result is millions of operations over a 25-30 year service life on transmission lines up to 800 kV.
Non-arcing Lightning Arrester Interactive Calculator
Vary surge voltage, line voltage, current, pulse duration, and stroke count to see MOV arrester energy absorption during a lightning event.
Equation Used
This calculator estimates the energy a non-arcing MOV lightning arrester must absorb during a surge. The voltage term is the clamped residual voltage above the normal line-to-ground operating voltage; multiplying by surge current, pulse duration, and number of strokes gives total event energy.
- Surge pulse is approximated as a triangular waveform.
- Ures and Uline are in kV, Isurge is in kA, and tau is in us, so kV*kA*us gives joules.
- Energy is zero if residual voltage is not above line-to-ground voltage.
- No sustained follow current or power-frequency arc is included.
How the Non-arcing Lightning Arrester Actually Works
A lightning strike on a transmission line drives a voltage spike of hundreds of kilovolts down the conductor in microseconds. Without protection, that surge punches through transformer insulation or flashes over insulator strings. The arrester sits between the line and ground, and stays effectively open-circuit at normal operating voltage. When the surge arrives, the arrester turns conductive within nanoseconds, dumps the energy to ground, and shuts off cleanly when voltage drops back to normal — that clean shut-off is what 'non-arcing' actually means.
The trick is the metal oxide varistor block. It's a sintered ceramic disc, mostly zinc oxide doped with bismuth, cobalt, and antimony. Below its MCOV (Maximum Continuous Operating Voltage) rating, leakage current sits in the milliamp range. Push voltage above the knee — typically 1.6 to 2.0 times MCOV — and resistance collapses by 6 to 8 orders of magnitude. The disc passes 10 kA or more to ground, then recovers automatically when the surge ends. No moving parts, no follow current, no arc to extinguish.
Tolerances matter. The MOV stack must be voltage-graded so each disc shares the surge equally — a 10% mismatch in V-I characteristic between discs concentrates energy on the weakest one and you get thermal runaway. Sealing is the other failure mode. If moisture penetrates the porcelain or polymer housing, partial discharge tracks across the disc surface, leakage current creeps up over months, and the arrester eventually fails short-circuit and trips the line. Field surveys at utilities like BPA show roughly 70% of arrester failures trace back to seal degradation, not MOV degradation.
Key Components
- Metal Oxide Varistor (MOV) Block: Sintered ZnO ceramic disc, typically 40-100 mm diameter and 20-45 mm thick per block. Provides the non-linear V-I characteristic that switches from megohms to milliohms above the knee voltage. Stack height scales with system voltage — a 230 kV class arrester uses 18-24 stacked discs.
- Housing (Porcelain or Polymer): Seals the MOV stack against moisture and contamination. Polymer (silicone rubber over fiberglass) housings have largely replaced porcelain since the 1990s because they fail safely without shrapnel. Creepage distance must meet 25-31 mm/kV for polluted coastal environments per IEC 60815.
- Pressure Relief Vent: Activates when an internal fault generates gas pressure above roughly 0.5 MPa. Vents the arc externally instead of letting the housing explode. Must coordinate with the system fault current rating → typically 20-65 kA for transmission class units.
- Grading Ring: Aluminum corona ring at the line-end terminal on units above 145 kV. Equalizes voltage distribution along the MOV stack — without it, the top disc sees 30-40% of the total stress and ages prematurely.
- Disconnector (on distribution arresters): Small explosive cartridge that physically drops the ground lead if the arrester fails short-circuit. Gives line crews a visual indicator and prevents the failed arrester from holding the feeder out of service.
Industries That Rely on the Non-arcing Lightning Arrester
Non-arcing arresters appear anywhere a transient overvoltage can damage equipment that costs more than the arrester. That covers transmission substations, distribution feeders, wind turbine generators, HVDC converter stations, and the surge-protection module inside a 240 V residential service panel. The physics scales — same MOV chemistry, different stack height.
- Transmission Utility: Hydro One uses Hitachi Energy EXLIM Q polymer-housed station class arresters at 500 kV switchyards to protect autotransformers, with MCOV of 318 kV and 10 kA discharge rating.
- Wind Power: GE Cypress 5.5 MW turbines mount Siemens 3EL2 arresters at the nacelle to clamp lightning surges entering through the blade lightning receptors before they reach the converter.
- Rail Traction: Deutsche Bahn 15 kV 16.7 Hz catenary systems use Tridelta SBK polymer arresters every 1.5 km along the contact wire to protect locomotive transformers from atmospheric surges.
- HVDC Transmission: The Rio Madeira ±600 kV HVDC link in Brazil uses ABB PEXLIM arresters at the converter valve halls, each rated 408 kV MCOV and 20 kA, sized to coordinate with the thyristor valve BIL.
- Distribution: Eaton VariSTAR AZE 10 kV polymer arresters mounted on every pole-top transformer along rural feeders for cooperatives like Tri-State G&T.
- Industrial Plant: Dofasco Hamilton steel mill uses Cooper UltraSIL arresters at the 13.8 kV switchgear ahead of arc-furnace transformers to suppress switching surges during furnace breaker operations.
The Formula Behind the Non-arcing Lightning Arrester
The core sizing question for an arrester is energy absorption — how many kilojoules per kV of MCOV the MOV stack can dump before thermal runaway. Push the arrester near the low end of the typical 2.5-10 kJ/kV range, like a distribution class unit, and you cover lightning strokes but a single switching surge on a long cable run will cook it. At the nominal 5-7 kJ/kV (intermediate class) you handle most utility duties. Push to the high end at 10+ kJ/kV (station class) and you cover capacitor bank switching and HVDC duty, but unit cost roughly doubles for each step up. The sweet spot for most 230 kV substations sits at 7.5 kJ/kV.
Variables
| Symbol | Meaning | Unit (SI) | Unit (Imperial) |
|---|---|---|---|
| W | Energy absorbed by the arrester per event | J | ft·lb |
| Ures | Residual (discharge) voltage of the arrester at the surge current | kV | kV |
| Uline | Line-to-ground operating voltage | kV | kV |
| Isurge | Peak surge current through the arrester | kA | kA |
| τ | Effective surge duration (time constant) | µs | µs |
| N | Number of successive strokes in a multi-stroke flash | dimensionless | dimensionless |
Worked Example: Non-arcing Lightning Arrester in a 138 kV substation transformer protection scheme
A municipal utility in Tampa Florida is sizing arresters for a 138 kV / 13.8 kV step-down transformer at a new substation feeding a downtown commercial district. Florida sees 12-15 lightning flashes per km² per year — among the highest in North America — so multi-stroke duty is the design driver. The engineer specifies a station class arrester with 108 kV MCOV and 350 kV residual at 10 kA, 8/20 µs waveshape, and needs to verify the energy absorption against a credible direct-stroke event.
Given
- Ures = 350 kV
- Uline = 80 kV (138 kV / √3)
- Isurge,nom = 10 kA
- τ = 20 µs
- N = 3 strokes
Solution
Step 1 — compute the voltage across the MOV stack during the surge. The arrester clamps at Ures while the line sits at Uline, so the net driving voltage is the difference:
Step 2 — at nominal 10 kA peak surge current with 20 µs effective duration and 3 successive strokes (a typical Florida multi-stroke flash):
Per kV of MCOV that's 81 / 108 = 0.75 kJ/kV per flash. A station class unit rated 7.5 kJ/kV has 10× margin against this single event — comfortable headroom for the design life.
Step 3 — at the low end of typical surge current (5 kA, a moderate induced surge from a nearby strike rather than direct hit), energy drops linearly:
This is the bread-and-butter case — the arrester barely warms up and recovers in milliseconds. You'd see thousands of these over a 30-year service life with no measurable degradation.
Step 4 — at the high end, a severe direct stroke at 40 kA with 6 successive strokes (worst-case Florida summer storm, 1-in-100-year event):
(Note Ures rises to roughly 380 kV at 40 kA because of the MOV's slight non-linearity above 10 kA.) That's 8.4 kJ/kV — right at the edge of a 7.5 kJ/kV unit's capability. For Florida coastal duty you'd step up to a 10 kJ/kV class arrester like the Hitachi EXLIM R, not the standard EXLIM Q.
Result
Nominal energy absorption is 81 kJ per multi-stroke flash, or 0. 75 kJ/kV of MCOV. That's well inside the 7.5 kJ/kV station-class rating — the MOV stack temperature rises maybe 3-4°C and bleeds off in seconds. Across the operating range, a moderate 5 kA induced surge dumps 40 kJ (trivial), while a 40 kA direct stroke with 6 restrikes pushes 912 kJ and uses 84% of the unit's energy budget — meaning Florida utilities should specify the next class up. If your arrester fails earlier than the 30-year design life, the most common diagnostic causes are: (1) MCOV undersized for actual TOV during single-line-to-ground faults, driving steady-state leakage above 1 mA and slow thermal aging, (2) corona ring missing or damaged on units above 145 kV, concentrating stress on the top disc, or (3) ground lead inductance above 5 µH per metre creating an L·di/dt voltage that effectively raises Ures at the protected equipment terminals.
Choosing the Non-arcing Lightning Arrester: Pros and Cons
The non-arcing MOV arrester replaced two earlier technologies — the simple spark gap and the silicon carbide gapped arrester. Each has different speed, lifespan, and cost behaviour, and the choice still matters for legacy retrofits and specialty applications.
| Property | Non-arcing MOV Arrester | Silicon Carbide Gapped Arrester | Simple Spark Gap |
|---|---|---|---|
| Response time | < 25 ns | 0.5-1 µs (gap firing delay) | 1-5 µs |
| Follow current after surge | None — self-resetting | Interrupted at next current zero | Sustained until breaker trips |
| Service life (operations) | 10⁶+ at rated energy | 10⁴-10⁵ | 10²-10³ before erosion |
| Energy capability (kJ/kV MCOV) | 2.5-13 depending on class | 1-3 | Limited by gap erosion |
| Residual voltage at 10 kA | 2.5-3.0 × MCOV | 3.0-3.5 × MCOV | Highly variable, 4-6 × nominal |
| Cost (138 kV class, USD) | $3,000-6,000 | Obsolete, $8,000+ refurb | $200-500 (rod gap only) |
| Typical application | All modern transmission and distribution | Pre-1985 substations still in service | Backup protection on wood-pole distribution |
Frequently Asked Questions About Non-arcing Lightning Arrester
Leakage current on a healthy ZnO arrester has a capacitive component (roughly 0.5-1 mA at MCOV from the disc geometry) and a resistive component (should be <0.1 mA new). A field meter reading the total RMS current sees both, so 1-1.5 mA right out of the box is normal.
If you're reading 2 mA or more, check whether your meter measures total current or third-harmonic resistive current — only the resistive component indicates real degradation. A LCM-500 or similar third-harmonic analyzer is the correct tool. If resistive leakage really is climbing above 0.3 mA, suspect surface contamination on the housing or moisture ingress through a damaged seal, not MOV aging.
Drive the decision from energy duty, not voltage class. Station class (per IEEE C62.11) gives you 7-13 kJ/kV and 65 kA fault withstand. Intermediate class gives 4-7 kJ/kV and 40 kA. For a 69 kV substation feeding a transformer with no long cable runs and no capacitor banks, intermediate is fine. Add a capacitor bank for power factor correction and you must go station class — switching a 10 MVAR bank generates restrike surges that cumulatively dump 200-400 kJ into the arrester over its life.
Lightning ground flash density also pushes the decision. Above 8 flashes/km²/year (most of the US Southeast), step up one class regardless of equipment.
Disconnectors only fire on a sustained fault current — typically when the arrester has failed short-circuit and is drawing several hundred amps from the line. A normal lightning operation passes 5-20 kA for 20-100 µs and the disconnector cartridge sees that pulse as a non-event. So no operation of the disconnector is the expected outcome of a successful surge clamping.
To confirm health, do a leakage current measurement after the storm and compare to the baseline reading from commissioning. A 50%+ rise in resistive leakage means the surge exceeded the energy rating and you should replace, even though the arrester still appears intact.
Two reasons, both about lead inductance. First, the ground lead and line lead together add roughly 1 µH per metre of straight conductor. A 30 kA/µs lightning di/dt through 2 metres of lead produces 60 kV of inductive voltage drop on top of the arrester's residual voltage. The transformer terminal sees Ures + L·di/dt, not Ures.
Second, separation distance between arrester and transformer matters because of travelling-wave doubling. Beyond about 5-10 metres of cable between them, the surge reflects off the transformer's capacitive termination and the voltage at the transformer can reach 1.8× the arrester clamping voltage. Rule of thumb: keep total lead length plus separation under 3 metres for transmission class, mount the arrester directly on the transformer tank if possible.
In theory yes, in practice no — and the failure mode is brutal. MOV blocks have a slight negative temperature coefficient above the knee voltage. Whichever arrester has the marginally lower V-I curve hogs the surge current, heats up faster, drops its curve further, and takes more current. Within a few cycles one arrester does 80% of the work and fails, then the other follows.
Manufacturers who do offer parallel arresters (for HVDC duty mainly) factory-match the V-I curves to within 1% and bond them with a common heatsink. Field-paralleling units from different production batches is asking for a double failure. Buy the correctly sized single unit.
Polymer housings (silicone rubber over fiberglass rod) fail through a different mechanism than porcelain. UV exposure plus surface pollution creates dry-band arcing on the silicone weather sheds, which over years carbonizes a tracking path. Once a track forms, leakage current along the housing surface heats the MOV stack from outside, raising internal temperature into thermal runaway.
Coastal salt fog environments and agricultural areas with conductive dust shorten polymer life dramatically — IEC 60815 Class IV pollution sites need 31 mm/kV creepage minimum, and many older polymer arresters were specified for Class II at 20 mm/kV. If you're in a coastal or industrial zone and seeing premature failures, the housing creepage was undersized for the site, not a manufacturing defect.
The protective margin is what matters. IEEE C62.22 calls for at least 20% margin between the transformer's BIL (Basic Insulation Level) and the arrester's lightning impulse protective level (LPL, equal to Ures at 10 kA for station class).
For a 138 kV transformer with 650 kV BIL and an arrester with 350 kV LPL, margin = (650 − 350) / 350 = 86%. Plenty. But push to a 230 kV transformer with reduced BIL of 825 kV (modern dry-type) and an arrester at 565 kV LPL, margin = 46% — still fine. The trap is older equipment where decades of moisture absorption have effectively lowered the in-service BIL by 15-20%, eroding the margin without anyone noticing until a strike causes a winding failure.
References & Further Reading
- Wikipedia contributors. Surge arrester. Wikipedia
Building or designing a mechanism like this?
Explore the precision-engineered motion control hardware used by mechanical engineers, makers, and product designers.
