Near Field Notes

Note 35EMCRev 1Watch the video · 3:29

MOV

How can an MOV start a fire?

By sitting on a voltage that stays too high. On the 180 V of a broken neutral, a 130 V MOV passes 300 °C in 43 s while drawing 0.14 A, far below what trips a fuse or breaker. A 150 V MOV warms 5 °C on the same 180 V, and it still clamps a 2,000 V surge to 422 V.

A surge lasts microseconds, and the MOV is built for it. A stuck overvoltage lasts until someone fixes the wire, and the MOV conducts near every peak of it.

A thermal link that opens near 180 °C takes the 130 V MOV out at 18 s. A PTC in the line does the same job and resets.

One input, three conditions

What the MOV reaches

A 2,000 V surge150 V MOV422 Vclamped, at 773 AVoltage at the line01,0002,000 VNo MOVWith the MOV01020304050 µs180 V that staysThe same 150 V MOV30 °Clevels off, at 1.7 mATemperature of the disc0150300450 °C0306090120 s180 V that staysA 130 V MOV300 °Cat 43 s, drawing 0.14 ATemperature of the disc0150300450 °C0306090120 s150 V MOV300 °C at 43 sLNURest ofthe supplyFuseMOVA 2,000 V surge150 V MOV422 Vclamped, at 773 AVoltage at the line01,0002,000 VNo MOVWith the MOV01020304050 µs180 V that staysThe same 150 V MOV30 °Clevels off, at 1.7 mATemperature of the disc0150300450 °C0306090120 s180 V that staysA 130 V MOV300 °Cat 43 s, drawing 0.14 ATemperature of the disc0150300450 °C0306090120 s150 V MOV300 °C at 43 sLNURest of the supplyFuseMOVA 2,000 V surge150 V MOV422 Vclamped, at 773 AVoltage at the line01,0002,000 VNo MOVWith the MOV01020304050 µs180 V that staysThe same 150 V MOV30 °Clevels off, at 1.7 mATemperature of the disc0150300450 °C0306090120 s180 V that staysA 130 V MOV300 °Cat 43 s, drawing 0.14 ATemperature of the disc0150300450 °C0306090120 s150 V MOV300 °C at 43 sLNURest of the supplyFuseMOV
The input, with the current’s path wide for the surge and thin on the stuck 180 V, where the MOV glows with its heat. Under it, the line’s voltage in the surge, then the disc’s temperature from 25 °C.

An MOV on a supply’s input

120 V line, 20 mm MOV

An MOV, a metal-oxide varistor, is a disc that sits across the mains to clamp surges. This board is the input of a small supply on a 120 V line.

The line comes in through a fuse. Right after it, the MOV sits across line and neutral, ahead of the X cap, the choke, the bridge and the bulk cap.

When a surge arrives, the MOV conducts, and the current takes the path drawn on the board. Note 11 follows what a charger does after its bulk cap.

The board, and a surge’s path
  • Surge current
Fig. 35aMOVBulk capFuseLine inFig. 35aMOVBulk capFuseBridgeLine inFig. 35aMOVBulk capFuseBridgeLine in
The surge’s current drawn on the board’s copper: in through the fuse, through the MOV and back along neutral, dashed where a part hides it. The board is an example drawn for this page, with generic parts.
The numbers
PartWhat it is
TerminalLine and neutral in. US colors: black is line, white is neutral
FuseA 5 × 20 mm glass fuse in clips, in the line
MOV20 mm, marked 20D241K: 240 V at 1 mA, the 150 V rms class. Right after the fuse, across line and neutral
X capAn X2 film capacitor across the line, for emissions
ChokeA common-mode choke in both lines
BridgeA bridge rectifier
Bulk cap47 µF, 400 V
Drawn pathBlue: the surge current, in through the fuse and the MOV and back along neutral. Dashed where a part hides it

In the circuit, the MOV is a resistor whose value depends on the voltage across it. Its symbol is a resistor crossed by a slanted line and marked U.

At the line’s 170 V peak it is all but open. The 150 V part here draws 50 nA, as much as a 3.4 GΩ resistor would.

The same board, as a circuit
Fig. 35bLN~~+−+FuseChokeBridgeMOVX capBulk capFig. 35bLNU~~+−+FuseChokeBridgeMOVX capBulk capFig. 35bLNU~~+−+FuseChokeBridgeMOVX capBulk cap
The MOV in its IEC symbol. The bulk cap is 47 µF, rated 400 V.

Almost no current, then a flood

Two 20 mm parts

An MOV’s current climbs steeply with voltage. Across Fig. 35c, the 150 V part’s voltage less than triples while its current grows a trillionfold.

Its name is the rms voltage it can sit on for good. Its datasheet gives two more: 216 to 264 V at 1 mA, and at most 395 V at 100 A.

The 130 V part sits about 15% lower. On a stuck 180 V, whose peaks reach 255 V, that gap is the difference between 5.4 mA and 0.43 A.

Voltage against current
  • 150 V part
  • 130 V part
  • Line peaks
Fig. 35cVoltage across the MOV0100200300400500 V50 nA5.0 µA5.4 mA0.43 A1 nA1 µA1 mA1 A1 kACurrent, log scaleFig. 35cVoltage across the MOV0100200300400500 V50 nA5.0 µA5.4 mA0.43 A1 nA1 µA1 mA1 A1 kACurrent, log scaleFig. 35cVoltage across the MOV0100200300400500 V50 nA5.0 µA5.4 mA0.43 A1 nA1 µA1 mA1 A1 kACurrent, log scale
Each curve runs through its part’s datasheet points, in the middle of its tolerance. The dashed lines are the peaks of a 120 V line, 170 V, and of a stuck 180 V, 255 V. The dots are where each part meets them.
The numbers
Middle of each part’s tolerance. The datasheet clamp is a maximum.
PartAt 1 mAαAt 170 VAs a resistorAt 255 VDatasheet, 100 A
V150LA20A240 V28.650 nA3,391 MΩ5.4 mA395 V
V130LA20A205 V28.05.0 µA34 MΩ0.43 A340 V

Inside, a disc of zinc oxide grains

About 2 to 3.5 V per boundary

Under the coating is a ceramic disc of zinc oxide, pressed and fired, with an electrode on each face.

The grains conduct well. The thin layers where they meet do not, until each sees about 2 to 3.5 V. A 150 V part, 240 V at 1 mA, has between about 70 and 120 of those boundaries in a row from face to face.

So the disc’s thickness sets its voltage, and its area sets how much current it can take. TDK gives about 30 to 250 V per mm of ceramic.

An MOV, cut open
Fig. 35dSilver electrodeZinc oxide ceramicEpoxy coatingFig. 35dFired silver electrodeZinc oxide ceramicBlue epoxy coatingFig. 35dFired silver electrodeZinc oxide ceramicBlue epoxy coating
The coating cut back to the silver, and the silver cut back to the ceramic. The grains are 10 to 100 µm across, too small to see here.

A 2,000 V surge, held to 422 V

IEC 61000-4-5

The standard test surge comes from IEC 61000-4-5’s combination-wave generator: 2,000 V into an open line, rising in about a microsecond, and about 1,000 A into a short.

Put the 150 V MOV across it and the line peaks at 422 V, while 773 A runs through the MOV.

The 2,000 V surge on the 150 V MOV
  • With the MOV
  • No MOV
Fig. 35eVoltage at the line−50005001,0001,5002,000 VNo MOV: 2,000 V422 VCurrent through the MOV05001,000 A773 A01020304050 µsFig. 35eVoltage at the line−50005001,0001,5002,000 VNo MOV: 2,000 V422 VCurrent through the MOV05001,000 A773 A01020304050 µsFig. 35eVoltage at the line−50005001,0001,5002,000 VNo MOV: 2,000 V422 VCurrent through the MOV05001,000 A773 A01020304050 µs
The MOV alone across the generator. It lets go at about 25 µs, and the line swings below zero.
The numbers
The MOV alone on the generator. 150 V part unless named. Its rating is 80 J, the 130 V part’s 70 J.
SurgePeak voltsPeak ampsJoules into the MOV
1 kV3872941.7
2 kV, no MOV2,000——
2 kV4227735.8
4 kV4791,74016.9
2 kV, 130 V MOV3658035.5
608

Does every supply need one?

1, 2 and 4 kV

No. The bulk cap behind the bridge soaks up a surge too. With no MOV, a 1 kV surge lifts it from 170 V to 298 V, well under its 400 V rating.

Power Integrations’ DER-849, a 30 W reference supply, passes ±1 kV line to neutral that way, with 54 µF of bulk cap and no MOV.

At 2 kV the same cap reaches 429 V, past its rating, and at 4 kV 713 V. With the MOV it stays at 303 V and 350 V.

Road and street lights are tested at 2 kV line to neutral under IEC 61547. A survey of 250 US homes, published in 1970, found repeated surges over 1,200 V in 2.4% of them.

The bulk cap after a surge
  • No MOV
  • 150 V MOV
Fig. 35fBulk cap’s peak0200400600800 V2982611 kV4293032 kV7133504 kVCap rated 400 VSurge, line to neutralFig. 35fBulk cap’s peak0200400600800 V2982611 kV4293032 kV7133504 kVCap rated 400 VSurge, line to neutralFig. 35fBulk cap’s peak0200400600800 V2982611 kV4293032 kV7133504 kVCap rated 400 VSurge, line to neutral
On the input of Fig. 35b, with the surge at the line’s peak. The cap starts at 170 V.
The numbers
Peak values, volts and amps. Surge at the 90° peak of the 120 V line, onto a 47 µF cap charged to 170 V.
SurgeBulk cap, no MOVWith MOVBridge amps, no MOVWith MOVMOV amps
1 kV298261340241182
2 kV429303680285619
4 kV7133501,3603471,532

Two ways to fail

Short, or slowly hot

Each big surge uses up some of the disc. In the pulse rating NIST quotes for 20 mm parts, a disc takes 120,000 surges of 100 A, but only 20,000 of 180 A.

A surge far past its rating cracks the disc or punches through it, and it fails as a short. That failure is the easy one: the fuse ahead of it blows, and the product goes dark.

The other way is slow. A voltage that stays above the MOV’s rating makes it conduct on every cycle, at a current a fuse sized to pass surges will not notice.

A disc after a surge too big for it
Fig. 35gCracked at the edgePunched throughFig. 35gCracked at the edgePunched throughFig. 35gCracked at the edgePunched through
The same 20 mm MOV, failed: a crack in from the rim and a hole through the face, each scorched. Either one leaves a short.

A broken neutral can leave 180 V on an outlet

NFPA 921’s example

A US house gets 240 V from a center-tapped transformer: two 120 V legs and a shared neutral.

If the neutral breaks, each leg’s loads end up in series across the 240 V and share it by their size. NFPA 921, the fire investigators’ guide, gives 60 V on one leg and 180 V on the other as an example.

Anything plugged into the lightly loaded leg then sits on 180 V, MOV and all, until the neutral is fixed.

A split-phase supply, neutral broken
Fig. 35hBrokenUL1NL2Light load180 VHeavy load60 VMOVTransformerFig. 35hBrokenUL1NL2Light load180 VHeavy load60 VMOVTransformerFig. 35hBrokenUL1NL2Light load180 VHeavy load60 VMOVTransformer
Drawn for NFPA 921’s example. The light load has three times the resistance of the heavy one, so it takes three quarters of the 240 V.

On 180 V, the 130 V MOV keeps heating

300 °C at 43 s

180 V rms peaks at 255 V. The 130 V part draws 1 mA at 205 V, so near every peak it conducts: pulses of 0.43 A, 0.14 A rms and 16.2 W of heat.

The heat leaves through the two leads, which need 24.4 °C/W of rise to carry each watt away. So the disc climbs toward the air’s temperature plus P × Rθ, slowed by its own heat capacity:

T=Ta+P×Rθ×(1 − e−t/τ)

=25 °C+16.2 W×24.4 °C/W×(1 − e−43 s/36 s)=300 °C

For the 130 V part that level is 419 °C. It passes 300 °C on the way, at 43 s.

Two cycles at 180 V
  • Line
  • 130 V part
  • ±205 V
Fig. 35iVoltage, 180 V rms−2000200 V1 mA at 205 V255 VCurrent through the 130 V MOV−0.400.4 A0.43 A0102030 msFig. 35iVoltage, 180 V rms−2000200 V1 mA at 205 V255 VCurrent through the 130 V MOV−0.400.4 A0.43 A0102030 msFig. 35iVoltage, 180 V rms−2000200 V1 mA at 205 V255 VCurrent through the 130 V MOV−0.400.4 A0.43 A0102030 ms
The current flows only near the peaks, where the line is past the part’s 1 mA voltage.
The numbers
The middle of each part’s tolerance unless named. The low end is its lowest voltage at 1 mA.
At 180 V rmsPeakRmsHeat
130 V part0.43 A0.14 A16.2 W
130 V part, low end8.0 A2.6 A303 W
150 V part5.4 mA1.7 mA0.20 W
150 V part, low end0.11 A35 mA4.0 W
At 180 V, two MOVs heat up
  • 150 V part
  • 130 V part
Fig. 35jTemperature of the disc0150300450 °C300 °C at 43 s30 °C405 °C0306090120 sFig. 35jTemperature of the disc0150300450 °C300 °C at 43 s30 °C405 °C0306090120 sFig. 35jTemperature of the disc0150300450 °C300 °C at 43 s30 °C405 °C0306090120 s
From 25 °C, both parts in the middle of their tolerance.
The numbers
From 25 °C. The model has no melting or cracking, so a level-off far above 300 °C only says the disc keeps heating.
At 180 V rmsRms currentHeatLevels off at180 °C at300 °C at
150 V part1.7 mA0.20 W30 °CNeverNever
150 V part, low end35.4 mA4.0 W124 °CNeverNever
130 V part0.14 A16.2 W419 °C18 s43 s
130 V part, low end2.6 A303 WThousands0.8 s1.4 s
610

A thermal link pulls the MOV off the line

Opens near 180 °C

A thermally protected MOV has a thermal fuse in series with the disc, pressed against it. In Littelfuse’s patent for its TMOV, the fuse is a solder wire that melts at about 180 °C, balls up and pulls away, and a hot-melt fills the gap.

It works once. The MOV is gone, and the product keeps running with no surge protection. Littelfuse’s iTMOV adds a third lead to report it.

A thermally protected MOV
Fig. 35kDisc and link insideThird leadreports the tripFig. 35kDisc and link insideThird leadreports the tripFig. 35kDisc and link insideThird leadreports the trip
A generic block-style part, after the 20 mm thermally protected types. The link is inside, against the disc.

Modeled as a link that opens at 180 °C, it takes the 130 V part off the 180 V at 18 s. The disc then cools through its leads.

Without the link, the same part would pass 300 °C at 43 s.

The link opens
  • With a link
  • Without
Fig. 35lTemperature of the 130 V disc0150300450 °COpens at 18 sNo link0306090120 sFig. 35lTemperature of the 130 V disc0150300450 °COpens at 18 sNo link0306090120 sFig. 35lTemperature of the 130 V disc0150300450 °COpens at 18 sNo link0306090120 s
The 130 V part at 180 V, from 25 °C.
The numbers
130 V part at 180 V rms, from 25 °C. The link opens at 180 °C, at 18.1 s.
TimeWith the linkWithout
0 s25 °C25 °C
10 s120 °C120 °C
18 s180 °C180 °C
30 s136 °C247 °C
60 s74 °C344 °C
90 s46 °C386 °C
120 s34 °C405 °C

Or a PTC that cuts the power and resets

Up to 75 VA on 120 V

A PTC is a resistor that jumps to a high value when it gets hot. The 2Pro, made by TE and now sold by Littelfuse, puts one in the line and an MOV right after it, in one body. The MOV’s heat trips the PTC, and the load loses power.

TE tests it at 400 V, a lost neutral on a 230 V line, and it holds the MOV under 150 °C. TE warns that a plain 275 V MOV there may overheat even with a fuse upstream.

It resets once it cools. All the load’s current runs through the PTC, so it suits small loads, up to 75 VA on 120 V.

An MOV with a PTC in the line
Fig. 35mLN+t°ULoadPTCMOVOne part, three leadsFig. 35mLN+t°ULoadPTCMOVOne part, three leadsFig. 35mLN+t°ULoadPTCMOVOne part, three leads
The PTC in its IEC symbol, marked +t°. Both sit in one body, so the MOV’s heat reaches the PTC.

What the standards ask

Three of them

No product safety standard makes a product fit an MOV. Once one is fitted, the standards test how it fails.

What each standard asks of an MOV

Fig. 35n

StandardCoversWhat it does to an MOVA pass
UL 1449Surge protectors, a power strip’s includedHolds it at full line-to-line voltage, the current limited in steps down to 0.125 A for a plug-in strip, for up to 7 hNo flame, molten metal or holes in the case. Disconnecting counts
IEC 62368-1, Annex G.8Audio, video and IT equipmentTreats it as a possible ignition source, rated at least 1.25 times the line. Rated 1.25 to 2 times: twice the line voltage through a resistor, from 0.125 A and doubled each run. Over 2 times, line to neutral: no testA fuse or thermal fuse that opens before fire
IEC 60335-1Household appliancesShorts it as a fault, even one with a thermal link, per VDE’s 2021 ruling. Its 2 kV surge test runs with the MOV taken outSomething ahead of it, like a fuse, clears the short safely
Paraphrased from documents about the standards, all in the sources below, not from the standards’ own text: DeGregoria’s slides and Mersen’s note on UL 1449, Maytum’s slides on IEC 62368-1, and VDE’s ruling and an SGS test report on IEC 60335-1. A thermal link or a fuse that opens before anything burns is a pass in the first two.

Hold a voltage yourself

130 and 150 V parts, 120 to 240 V

Pick a part, where it sits in its tolerance, and a voltage held across it from 25 °C. The curve is the disc’s temperature.

In the middle of its tolerance the 150 V part stays under 180 °C up to 202 V, and the 130 V part up to 174 V. At the low end those fall to 183 V and 157 V.

Part
Tolerance
180 V
The disc’s temperature, against time

130 V part, 180 V: 300 °C at 43 s

  • This part
  • 180 °C
Temperature of the disc150300450 °C0060120180 sTemperature of the disc150300450 °C0060120180 sTemperature of the disc150300450 °C0060120180 s
Current
0.14 A
Heat
16.2 W
Levels off at
419 °C
Passes 180 °C
18 s
The current is rms. The dashed line is 180 °C, about where a thermal link opens. A curve that passes 450 °C stops at the top.
The numbers
Five of the 25 voltages for each part and tolerance. From 25 °C.
PartHeldRms currentHeatLevels off at180 °C at
130 V, middle150 V849.6 µA0.08 W27 °CNever
130 V, middle165 V12 mA1.3 W57 °CNever
130 V, middle180 V0.14 A16.2 W419 °C18 s
130 V, middle200 V2.6 A337 WOver 450 °C0.7 s
130 V, middle240 V145 A24,580 WOver 450 °CUnder 0.1 s
130 V, low end150 V16 mA1.6 W63 °CNever
130 V, low end165 V0.24 A24.8 WOver 450 °C11 s
130 V, low end180 V2.6 A302 WOver 450 °C0.8 s
130 V, low end200 V36.4 A4,824 WOver 450 °C0.1 s
130 V, low end240 V414 A76,980 WOver 450 °CUnder 0.1 s
150 V, middle150 V9.5 µAUnder 0.01 W25 °CNever
150 V, middle165 V145.2 µA0.02 W25 °CNever
150 V, middle180 V1.7 mA0.20 W30 °CNever
150 V, middle200 V35 mA4.5 W134 °CNever
150 V, middle240 V6.1 A935 WOver 450 °C0.2 s
150 V, low end150 V193.4 µA0.02 W25 °CNever
150 V, low end165 V2.9 mA0.31 W33 °CNever
150 V, low end180 V35 mA4.0 W124 °CNever
150 V, low end200 V0.71 A90.5 WOver 450 °C2.6 s
150 V, low end240 V68.4 A11,120 WOver 450 °CUnder 0.1 s

Fit one, and guard it how?

Three questions

Which way fits depends on three things: what a surge does to the supply without an MOV, how high the MOV is rated, and how much the load draws.

Key to an MOV on a mains input

Does it need one, and what stops it burning?

  1. 1aAt the product’s test surge, the bulk cap stays under its ratingNo MOV
  2. 1bIt goes overGo to 2
  3. 2aThe MOV is rated over twice the line voltageA plain MOV
  4. 2bIt is rated 1.25 to 2 times the line voltageGo to 3
  5. 3aThe load is small, up to 75 VA on 120 VAn MOV with a PTC
  6. 3bThe load is biggerA thermally protected MOV

IEC 62368-1 wants an MOV rated at least 1.25 times the line voltage, so on 120 V the 130 V part is out before the key starts. Over twice the line voltage, it lets a line-to-neutral MOV skip its overload test. The board in Fig. 35a has a 150 V MOV, exactly 1.25 times its line, and its 47 µF cap goes past 400 V at 2 kV, so it reaches 3.

Conditions

Simulated
Every voltage, current and temperature on this page comes from a simulation. None was measured.
MOV
NIST’s model: a power law, I = kVα, in series with a bulk resistance, fitted to Littelfuse’s datasheet points at 1 mA and 100 A. V150LA20A: 216 to 264 V at 1 mA, 395 V at 100 A, α 28.6, 0.047 Ω, 1.6 nF. V130LA20A: 184.5 to 225.5 V, 340 V, α 28.0, 0.040 Ω, 1.9 nF. The middle of tolerance is the middle of the 1 mA range, and the low end its bottom.
Surge
ngspice. IEC 61000-4-5’s combination wave from Carobbi and Bonci’s generator circuit with its 18 µF coupling capacitor: 1.2/50 µs into an open line, 8/20 µs into a short. The MOV alone across it, and the input of Fig. 35b: 1.5 mH to the mains, a 0.1 µF X cap, the MOV, 0.5 Ω and 5 µH for the fuse, choke and traces, a bridge and a 47 µF bulk cap at 170 V, with the surge at the line’s peak.
Heating
NIST’s model: a 3 g disc of zinc oxide, 1.48 J/°C, losing heat only through two copper leads 0.81 mm across and 10 mm long, 24.4 °C/W, with no air cooling. Its time constant is 36 s. The heat is the average over one 60 Hz cycle of 20,000 points. k and α stay fixed as the disc heats, as in NIST’s paper.
Check
For a low-end 130 V part, NIST’s paper gives about +130 °C at 156 V and nearly +300 °C at 160 V. This rebuild gives +119 °C and +247 °C.
Thermal link
Opens at 180 °C, the link wire’s melting point in Littelfuse’s patent. The disc then cools through the same leads.
Runs
11 surge runs in ngspice. 113 heating cases, 100 of them for the explorer.
Video
The video says about 80 grain boundaries. This page gives the range, 70 to 120, from 2 to 3.5 V per boundary.
Renders
Blender, the video’s models rendered again on white. Generic parts.

Simulated only. Nothing on this page has been measured on a board yet.

Limits

  • The model’s curve stays put as the disc heats. NIST gives a real part’s 1 mA voltage as falling 0.05% per °C, with the effect fading above 1 mA, so a hot disc conducts somewhat more than modeled.
  • Heat leaves only by the leads, NIST’s worst case. Air and the board would carry some of it away.
  • Nothing melts, cracks or burns in the model. Far past 300 °C it only says the disc keeps heating.
  • The clamp has no upturn past the bulk resistance, so at kiloamps it reads a little low.
  • Wear from earlier surges is not modeled. A worn MOV leaks more and runs away at a lower voltage.
  • 120 V only. On 230 V the same story runs with a 275 V part and about 400 V from a broken neutral, TE’s test voltage.

Files

  • sim/mov.pyThe MOV model and the two parts’ datasheet numbers
  • sim/surge.pyThe surge generator, the input circuit and its surge runs
  • sim/runaway.pyThe heating model, its check against NIST and the thermal link
  • sim/site.pyThe runs added for this page: the V-I curves, the explorer’s states, the limits and the 130 V part’s surge
  • render/parts.pyThe board and the MOVs in Blender

Sources

Like this note?

Rev 1 ·
  1. Rev 1First published

Each fix gets a line here, with the reader's name if they want credit.

Did it answer your question?

Want more on this?

Report a problem

What should the next note be? Ask for one