Snubber

Semiconductor Components Industries, LLC, 1999

August, 1999 – Rev. 21 Publication Order Number:

AN1048/D

By George TempletonThyristor Applications Engineer

INTRODUCTION

Edited and Updated

RC networks are used to control voltage transients thatcould falsely turn-on a thyristor. These networks are calledsnubbers.

The simple snubber consists of a series resistor andcapacitor placed around the thyristor. These componentsalong with the load inductance form a series CRL circuit.Snubber theory follows from the solution of the circuit’sdifferential equation.

Many RC combinations are capable of providing accept-able performance. However, improperly used snubbers cancause unreliable circuit operation and damage to the semi-conductor device.

Both turn-on and turn-off protection may be necessaryfor reliability. Sometimes the thyristor must function with arange of load values. The type of thyristors used, circuitconfiguration, and load characteristics are influential.

Snubber design involves compromises. They includecost, voltage rate, peak voltage, and turn-on stress. Practi-cal solutions depend on device and circuit physics.

STATIC dVdt

WHAT IS STATIC dVdt

?

Static dVdt

is a measure of the ability of a thyristor to

retain a blocking state under the influence of a voltagetransient.

dVdt

s DEVICE PHYSICS

Static dVdt

turn-on is a consequence of the Miller effect

and regeneration (Figure 1). A change in voltage across thejunction capacitance induces a current through it. This cur-

rent is proportional to the rate of voltage change dVdt . It

triggers the device on when it becomes large enough toraise the sum of the NPN and PNP transistor alphas to unity.

Figure 6.1. ModeldVdt

s

IA CJ

dVdt

1 (N p)

CEFFCJ

1(Np)

IBP

IJ

IJ

IK

IBN

ICN

I1

I2

ICP

IA

TWO TRANSISTOR MODELOF

SCR

CJN

CJP

PNP

A

C

G

CJ

NE

PB

NB

PEV

tGNPN

INTEGRATEDSTRUCTURE

K

A

K

dvdt

http://onsemi.com

APPLICATION NOTE

AN1048/D

http://onsemi.com2

CONDITIONS INFLUENCING dVdt

s

Transients occurring at line crossing or when there is noinitial voltage across the thyristor are worst case. The col-lector junction capacitance is greatest then because thedepletion layer widens at higher voltage.

Small transients are incapable of charging the self-capacitance of the gate layer to its forward biased thresholdvoltage (Figure 2). Capacitance voltage divider actionbetween the collector and gate-cathode junctions and built-in resistors that shunt current away from the cathode emit-ter are responsible for this effect.

PEAK MAIN TERMINAL VOLTAGE (VOLTS)

700 800

80

60

MAC 228A10 TRIACTJ = 110°C

600500400300200

100

120

140

160

201000

180

40

STAT

IC

(

V/

s) µdV dt

Figure 6.2. Exponential versus Peak VoltagedVdt

s

Static dVdt

does not depend strongly on voltage for opera-

tion below the maximum voltage and temperature rating.Avalanche multiplication will increase leakage current and

reduce dVdt

capability if a transient is within roughly 50 volts

of the actual device breakover voltage.

A higher rated voltage device guarantees increased dVdt

at

lower voltage. This is a consequence of the exponential rat-ing method where a 400 V device rated at 50 V/µs has a

higher dVdt

to 200 V than a 200 V device with an identical

rating. However, the same diffusion recipe usually appliesfor all voltages. So actual capabilities of the product are notmuch different.

Heat increases current gain and leakage, lowering

dVdt

s, the gate trigger voltage and noise immunity

(Figure 3).

Figure 6.3. Exponential versus TemperaturedVdt

s

STAT

IC

(

V/

s) µdV dt

170

150

130

110

30

50

70

90

100 115 130 14585705540

MAC 228A10VPK = 800 V

TJ, JUNCTION TEMPERATURE (°C)

2510

dVdt

s

FAILURE MODE

Occasional unwanted turn-on by a transient may beacceptable in a heater circuit but isn’t in a fire preventionsprinkler system or for the control of a large motor. Turn-onis destructive when the follow-on current amplitude or rateis excessive. If the thyristor shorts the power line or acharged capacitor, it will be damaged.

Static dVdt

turn-on is non-destructive when series imped-

ance limits the surge. The thyristor turns off after a half-

cycle of conduction. High dVdt

aids current spreading in the

thyristor, improving its ability to withstand dIdt

. Breakdown

turn-on does not have this benefit and should be prevented.

Figure 6.4. Exponential versusGate to MT 1 Resistance

dVdt

s

STAT

IC

(

V/

s) µdV dt

20

100 10000

10 10,000GATE-MT1 RESISTANCE (OHMS)

MAC 228A10800 V 110°C

40

60

80

100

120

140

RINTERNAL = 600 Ω

AN1048/D

http://onsemi.com3

IMPROVING dVdt

s

Static dVdt

can be improved by adding an external resistor

from the gate to MT1 (Figure 4). The resistor provides a

path for leakage and dVdt

induced currents that originate in

the drive circuit or the thyristor itself.Non-sensitive devices (Figure 5) have internal shorting

resistors dispersed throughout the chip’s cathode area. Thisdesign feature improves noise immunity and high tempera-ture blocking stability at the expense of increased triggerand holding current. External resistors are optional for non-sensitive SCRs and TRIACs. They should be comparable insize to the internal shorting resistance of the device (20 to100 ohms) to provide maximum improvement. The internalresistance of the thyristor should be measured with an ohm-meter that does not forward bias a diode junction.

Figure 6.5. Exponential versusJunction Temperature

dVdt

s

STAT

IC

(

V/

s) µdV dt

800

1000

1200

1400

130120110100

2000

2200

1800

1600

9080706050600

TJ, JUNCTION TEMPERATURE (°C)

MAC 15-8VPK = 600 V

Sensitive gate TRIACs run 100 to 1000 ohms. With an

external resistor, their dVdt

capability remains inferior to

non-sensitive devices because lateral resistance within thegate layer reduces its benefit.

Sensitive gate SCRs (IGT 200 µA) have no built-inresistor. They should be used with an external resistor. Therecommended value of the resistor is 1000 ohms. Higher

values reduce maximum operating temperature and dVdt

s(Figure 6). The capability of these parts varies by more than100 to 1 depending on gate-cathode termination.

GAT

E-C

ATH

OD

E R

ESIS

TAN

CE

(OH

MS)

Figure 6.6. Exponential versusGate-Cathode Resistance

dVdt

s

10MEG

1MEG

100K

0.01 1001010K

0.1 1

MCR22-006TA = 65°C

0.001

KG

A10V

STATIC dVdt

(Vs)

A gate-cathode capacitor (Figure 7) provides a shuntpath for transient currents in the same manner as the resis-tor. It also filters noise currents from the drive circuit andenhances the built-in gate-cathode capacitance voltagedivider effect. The gate drive circuit needs to be able tocharge the capacitor without excessive delay, but it doesnot need to supply continuous current as it would for a

resistor that increases dVdt

the same amount. However, the

capacitor does not enhance static thermal stability.

Figure 6.7. Exponential versus Gateto MT1 Capacitance

dVdt

s

STAT

IC

(

V/

s) µdV dt

GATE TO MT1 CAPACITANCE (µF)

130

120

110

100

90

80

70

60

MAC 228A10800 V 110°C

10.10.010.001

The maximum dVdt

s improvement occurs with a short.

Actual improvement stops before this because of spreadingresistance in the thyristor. An external capacitor of about0.1 µF allows the maximum enhancement at a higher valueof RGK.

AN1048/D

http://onsemi.com4

One should keep the thyristor cool for the highest dVdt

s.

Also devices should be tested in the application circuit atthe highest possible temperature using thyristors with thelowest measured trigger current.

TRIAC COMMUTATING dVdt

WHAT IS COMMUTATING dVdt

?

The commutating dVdt

rating applies when a TRIAC has

been conducting and attempts to turn-off with an inductiveload. The current and voltage are out of phase (Figure 8).The TRIAC attempts to turn-off as the current drops belowthe holding value. Now the line voltage is high and in theopposite polarity to the direction of conduction. Successfulturn-off requires the voltage across the TRIAC to rise to theinstantaneous line voltage at a rate slow enough to preventretriggering of the device.

TIMEΦ

i

PHASEANGLE

iVLINE G

1

2R L

TIME

VLINE

MT2

-1V

VOLT

AGE/

CU

RR

ENT

Figure 6.8. TRIAC Inductive Load Turn-Off dVdt

c

dIdt

c

dVdt

c

VMT2-1

dVdt

c DEVICE PHYSICS

A TRIAC functions like two SCRs connected in inverse-parallel. So, a transient of either polarity turns it on.

There is charge within the crystal’s volume because ofprior conduction (Figure 9). The charge at the boundariesof the collector junction depletion layer responsible for

dVdt

s is also present. TRIACs have lower dV

dt

c than

dVdt

s because of this additional charge.

The volume charge storage within the TRIAC dependson the peak current before turn-off and its rate of zero

crossing dIdt

c. In the classic circuit, the load impedance

and line frequency determine dIdt

c. The rate of crossing

for sinusoidal currents is given by the slope of the secantline between the 50% and 0% levels as:

dIdt

c

6 f ITM1000

Ams

where f = line frequency and ITM = maximum on-state cur-rent in the TRIAC.

Turn-off depends on both the Miller effect displacement

current generated by dVdt

across the collector capacitance

and the currents resulting from internal charge storagewithin the volume of the device (Figure 10). If the reverserecovery current resulting from both these components ishigh, the lateral IR drop within the TRIAC base layer willforward bias the emitter and turn the TRIAC on. Commu-

tating dVdt

capability is lower when turning off from the pos-

itive direction of current conduction because of devicegeometry. The gate is on the top of the die and obstructscurrent flow.

Recombination takes place throughout the conductionperiod and along the back side of the current wave as itdeclines to zero. Turn-off capability depends on its shape. If

the current amplitude is small and its zero crossing dIdt

c is

low, there is little volume charge storage and turn-off

becomes limited by dVdt

s. At moderate current amplitudes,

the volume charge begins to influence turn-off, requiring alarger snubber. When the current is large or has rapid zero

crossing, dVdt

c has little influence. Commutating dI

dt and

delay time to voltage reapplication determine whether turn-off will be successful or not (Figures 11, 12).

STORED CHARGEFROM POSITIVECONDUCTION

PreviouslyConducting Side

NP

+ –

LATERAL VOLTAGEDROP

REVERSE RECOVERYCURRENT PATH

G MT1

TOP

MT2

N N N

N

N N N

Figure 6.9. TRIAC Structure and Current Flowat Commutation

AN1048/D

http://onsemi.com5

CHARGEDUE TO

dV/dt

Figure 6.10. TRIAC Current and Voltageat Commutation

didt

c

dVdt

c

TIME

IRRMVOLUME

STORAGECHARGE

VMT2-1

0

VOLT

AGE/

CU

RR

ENT

MAI

N T

ERM

INAL

VO

LTAG

E (V

)

Figure 6.11. Snubber Delay Time

E

td0

VT

TIME

EV

NO

RM

ALIZ

ED D

ELAY

TIM

E

Figure 6.12. Delay Time To Normalized Voltage

0.2

DAMPING FACTOR

0.02

0.03

0.05

0.1

0.2

0.5

10.50.30.001 0.002 0.005 0.01 0.02 0.2

0.1

0.05

0.02

0.01

0.05 0.1

(t *

= W

d0

0.005VTE

RL = 0M = 1IRRM = 0

t d)CONDITIONS INFLUENCING dV

dt

c

Commutating dVdt

depends on charge storage and recov-

ery dynamics in addition to the variables influencing staticdVdt

. High temperatures increase minority carrier life-time

and the size of recovery currents, making turn-off more dif-ficult. Loads that slow the rate of current zero-crossing aidturn-off. Those with harmonic content hinder turn-off.

Circuit Examples

Figure 13 shows a TRIAC controlling an inductive loadin a bridge. The inductive load has a time constant longerthan the line period. This causes the load current to remainconstant and the TRIAC current to switch rapidly as the linevoltage reverses. This application is notorious for causing

TRIAC turn-off difficulty because of high dIdt

c.

Figure 6.13. Phase Controlling a Motor in a Bridge

LR8.3 s

+

t

i

CRS

60 Hz

LS

–R L

DC MOTOR

i

dIdtc

High currents lead to high junction temperatures andrates of current crossing. Motors can have 5 to 6 times thenormal current amplitude at start-up. This increases bothjunction temperature and the rate of current crossing, lead-ing to turn-off problems.

The line frequency causes high rates of current crossingin 400 Hz applications. Resonant transformer circuits aredoubly periodic and have current harmonics at both the pri-mary and secondary resonance. Non-sinusoidal currentscan lead to turn-off difficulty even if the current amplitudeis low before zero-crossing.

dVdt

c FAILURE MODE

dVdt

c failure causes a loss of phase control. Temporary

turn-on or total turn-off failure is possible. This can bedestructive if the TRIAC conducts asymmetrically causing adc current component and magnetic saturation. The windingresistance limits the current. Failure results because ofexcessive surge current and junction temperature.

AN1048/D

http://onsemi.com6

IMPROVING dVdt

c

The same steps that improve dVdt

s aid dV

dt

c except

when stored charge dominates turn-off. Steps that reducethe stored charge or soften the commutation are necessarythen.

Larger TRIACs have better turn-off capability thansmaller ones with a given load. The current density is lowerin the larger device allowing recombination to claim agreater proportion of the internal charge. Also junctiontemperatures are lower.

TRIACs with high gate trigger currents have greaterturn-off ability because of lower spreading resistance in thegate layer, reduced Miller effect, or shorter lifetime.

The rate of current crossing can be adjusted by adding acommutation softening inductor in series with the load.Small high permeability “square loop” inductors saturatecausing no significant disturbance to the load current. Theinductor resets as the current crosses zero introducing alarge inductance into the snubber circuit at that time. Thisslows the current crossing and delays the reapplication ofblocking voltage aiding turn-off.

The commutation inductor is a circuit element thatintroduces time delay, as opposed to inductance, into the

circuit. It will have little influence on observed dVdt

at the

device. The following example illustrates the improvementresulting from the addition of an inductor constructed bywinding 33 turns of number 18 wire on a tape wound core(52000-1A). This core is very small having an outsidediameter of 3/4 inch and a thickness of 1/8 inch. The delaytime can be calculated from:

ts (N A B 108)

Ewhere:

ts = time delay to saturation in seconds.B = saturating flux density in GaussA = effective core cross sectional area in cm2

N = number of turns.

For the described inductor:

ts (33 turns) (0.076 cm2) (28000 Gauss)

(1 10–8) (175 V) 4.0 s.

The saturation current of the inductor does not need to bemuch larger than the TRIAC trigger current. Turn-off fail-ure will result before recovery currents become greater thanthis value. This criterion allows sizing the inductor with thefollowing equation:

Is Hs ML0.4 N

where :

Hs = MMF to saturate = 0.5 OerstedML = mean magnetic path length = 4.99 cm.

Is (.5) (4.99)

.4 33 60 mA.

SNUBBER PHYSICS

UNDAMPED NATURAL RESONANCE

0 ILC Radianssecond

Resonance determines dVdt

and boosts the peak capacitor

voltage when the snubber resistor is small. C and L are

related to one another by ω02. dVdt

scales linearly with ω0when the damping factor is held constant. A ten to one

reduction in dVdt

requires a 100 to 1 increase in either

component.

DAMPING FACTOR

ρ R2

CL

The damping factor is proportional to the ratio of thecircuit loss and its surge impedance. It determines the trade

off between dVdt

and peak voltage. Damping factors between

0.01 and 1.0 are recommended.

The Snubber Resistor

Damping and dVdt

When ρ 0.5, the snubber resistor is small, and dVdt

depends mostly on resonance. There is little improvement

in dVdt

for damping factors less than 0.3, but peak voltage

and snubber discharge current increase. The voltage wavehas a 1-COS (θ) shape with overshoot and ringing. Maxi-

mum dVdt

occurs at a time later than t = 0. There is a time

delay before the voltage rise, and the peak voltage almostdoubles.

When ρ 0.5, the voltage wave is nearly exponential in

shape. The maximum instantaneous dVdt

occurs at t = 0.

There is little time delay and moderate voltage overshoot.

When ρ 1.0, the snubber resistor is large and dVdt

depends mostly on its value. There is some overshoot eventhrough the circuit is overdamped.

High load inductance requires large snubber resistors andsmall snubber capacitors. Low inductances imply smallresistors and large capacitors.

AN1048/D

http://onsemi.com7

Damping and Transient Voltages

Figure 14 shows a series inductor and filter capacitorconnected across the ac main line. The peak to peak voltageof a transient disturbance increases by nearly four times.Also the duration of the disturbance spreads because ofringing, increasing the chance of malfunction or damage tothe voltage sensitive circuit. Closing a switch causes thisbehavior. The problem can be reduced by adding a dampingresistor in series with the capacitor.

V (V

OLT

S)

Figure 6.14. Undamped LC Filter Magnifies andLengthens a Transient

V0.1µF

100 µH 0.05

0 10 µs

340 VVOLTAGESENSITIVECIRCUIT

0

+ 700

– 700

TIME (µs)0 2010

dIdt

Non-Inductive Resistor

The snubber resistor limits the capacitor discharge

current and reduces dIdt

stress. High dIdt

destroys the thyristor

even though the pulse duration is very short.The rate of current rise is directly proportional to circuit

voltage and inversely proportional to series inductance.The snubber is often the major offender because of its lowinductance and close proximity to the thyristor.

With no transient suppressor, breakdown of the thyristorsets the maximum voltage on the capacitor. It is possible toexceed the highest rated voltage in the device seriesbecause high voltage devices are often used to supply lowvoltage specifications.

The minimum value of the snubber resistor depends onthe type of thyristor, triggering quadrants, gate currentamplitude, voltage, repetitive or non-repetitive operation,and required life expectancy. There is no simple way to pre-dict the rate of current rise because it depends on turn-onspeed of the thyristor, circuit layout, type and size of snub-ber capacitor, and inductance in the snubber resistor. Theequations in Appendix D describe the circuit. However, thevalues required for the model are not easily obtained exceptby testing. Therefore, reliability should be verified in theactual application circuit.

Table 1 shows suggested minimum resistor values esti-mated (Appendix A) by testing a 20 piece sample from thefour different TRIAC die sizes.

Table 1. Minimum Non-inductive Snubber Resistorfor Four Quadrant Triggering.

TRIAC TypePeak VC

VoltsRs

Ohms

dIdt

A/µs

Non-Sensitive Gate(IGT 10 mA)8 to 40 A(RMS)

200300400600800

3.36.8113951

170250308400400

Reducing dIdt

TRIAC dIdt

can be improved by avoiding quadrant 4

triggering. Most optocoupler circuits operate the TRIAC inquadrants 1 and 3. Integrated circuit drivers use quadrants 2and 3. Zero crossing trigger devices are helpful becausethey prohibit triggering when the voltage is high.

Driving the gate with a high amplitude fast rise pulse

increases dIdt

capability. The gate ratings section defines the

maximum allowed current.Inductance in series with the snubber capacitor reduces

dIdt

. It should not be more than five percent of the load

inductance to prevent degradation of the snubber’s dVdt

suppression capability. Wirewound snubber resistorssometimes serve this purpose. Alternatively, a separateinductor can be added in series with the snubber capacitor.It can be small because it does not need to carry the loadcurrent. For example, 18 turns of AWG No. 20 wire on aT50-3 (1/2 inch) powdered iron core creates a non-saturat-ing 6.0 µH inductor.

A 10 ohm, 0.33 µF snubber charged to 650 volts resulted

in a 1000 A/µs dIdt

. Replacement of the non-inductive snub-

ber resistor with a 20 watt wirewound unit lowered the rateof rise to a non-destructive 170 A/µs at 800 V. The inductorgave an 80 A/µs rise at 800 V with the non–inductiveresistor.

The Snubber Capacitor

A damping factor of 0.3 minimizes the size of the snub-

ber capacitor for a given value of dVdt

. This reduces the cost

and physical dimensions of the capacitor. However, it raisesvoltage causing a counter balancing cost increase.

Snubber operation relies on the charging of the snubbercapacitor. Turn-off snubbers need a minimum conductionangle long enough to discharge the capacitor. It should be atleast several time constants (RS CS).

AN1048/D

http://onsemi.com8

STORED ENERGY

Inductive Switching Transients

E 12

L I02 Watt-seconds or Joules

I0 = current in Amperes flowing in theinductor at t = 0.

Resonant charging cannot boost the supply voltage atturn-off by more than 2. If there is an initial current flowingin the load inductance at turn-off, much higher voltages arepossible. Energy storage is negligible when a TRIAC turnsoff because of its low holding or recovery current.

The presence of an additional switch such as a relay, ther-mostat or breaker allows the interruption of load current andthe generation of high spike voltages at switch opening. Theenergy in the inductance transfers into the circuit capacitanceand determines the peak voltage (Figure 15).

dVdt

IC

VPK I LC

Figure 6.15. Interrupting Inductive Load Current

VPK

C

L

I

FAST

SLOW

(b.) Unprotected Circuit(a.) Protected Circuit

OPTIONALR

Capacitor Discharge

The energy s tored in the snubber capac i tor

Ec 12

CV2 transfers to the snubber resistor and

thyristor every time it turns on. The power loss is propor-tional to frequency (PAV = 120 Ec @ 60 Hz).

CURRENT DIVERSION

The current flowing in the load inductor cannot changeinstantly. This current diverts through the snubber resistor

causing a spike of theoretically infinite dVdt

with magnitude

equal to (IRRM R) or (IH R).

LOAD PHASE ANGLE

Highly inductive loads cause increased voltage and

dVdt

c at turn-off. However, they help to protect the

thyristor from transients and dVdt

s. The load serves as the

snubber inductor and limits the rate of inrush current if the

device does turn on. Resistance in the load lowers dVdt

and

VPK (Figure 16).

dVdt

M = 0.25

M = 0.5

M = 0.75

VPK

1.3

E

0.4

0.2

00 0.2 0.4 0.6 0.8 1

0.9

1

1.1

1.2

1.4

1.5

1.6

1.7

1.4

1.2

1

0.8

0.6

2.2

2.1

2

1.9

1.8

DAMPING FACTOR

M = 0

M = RS / (RL + RS)

M = 1

dV dt( )

0

NO

RM

ALIZ

EDdV dt

M RESISTIVE DIVISION RATIO RS

RL RS

IRRM 0

Figure 6.16. 0 To 63% dVdt

/ (E

W )

NO

RM

ALIZ

ED P

EAK

VOLT

AGE

V

/EPK

CHARACTERISTIC VOLTAGE WAVES

Damping factor and reverse recovery current determinethe shape of the voltage wave. It is not exponential whenthe snubber damping factor is less than 0.5 (Figure 17) orwhen significant recovery currents are present.

V

(VO

LTS)

MT 2-

1

063% dVdt

s 100 Vs, E 250 V,

RL 0, IRRM 0

Figure 6.17. Voltage Waves For DifferentDamping Factors

ρ = 0

1

TIME (µs)

0.30.1

0

ρ = 0.1

ρ = 0.3 ρ = 1

3.52.82.1 4.2 4.9 5.6 6.30.70

100

200

300

400500

1.4 70

AN1048/D

http://onsemi.com9

NO

RM

ALIZ

ED P

EAK

VOLT

AGE

AND

dV dt

(RL 0, M 1, IRRM 0)

NORMALIZED dVdt

dVdtE 0

NORMALIZED VPK VPK

E

Figure 6.18. Trade-Off Between V PK and dVdt

DAMPING FACTOR (ρ)0

0–63%E

1

10–63%

0.80.60.40.2 1.2 1.81.4 1.6 2

2.8

2.6

2.4

2.2

1.8

1.6

1.4

2

1

1.2

0.80.6

0.4

0.2

0

dVdt

dVdt

o

VPKdVdt

10–63%

dVdtMAX

A variety of wave parameters (Figure 18) describe dVdt

Some are easy to solve for and assist understanding. These

include the initial dVdt

, the maximum instantaneous dVdt

, and

the average dVdt

to the peak reapplied voltage. The 0 to 63%

dVdt

s and 10 to 63% dV

dt

c definitions on device data

sheets are easy to measure but difficult to compute.

NON-IDEAL BEHAVIORSCORE LOSSES

The magnetic core materials in typical 60 Hz loadsintroduce losses at the snubber natural frequency. Theyappear as a resistance in series with the load inductance and

winding dc resistance (Figure 19). This causes actual dVdt

to

be less than the theoretical value.

Figure 6.19. Inductor Model

L R

C

L DEPENDS ON CURRENT AMPLITUDE, CORESATURATION

R INCLUDES CORE LOSS, WINDING R. INCREASESWITH FREQUENCY

C WINDING CAPACITANCE. DEPENDS ONINSULATION, WIRE SIZE, GEOMETRY

COMPLEX LOADS

Many real-world inductances are non-linear. Their corematerials are not gapped causing inductance to vary withcurrent amplitude. Small signal measurements poorly char-acterize them. For modeling purposes, it is best to measurethem in the actual application.

Complex load circuits should be checked for transientvoltages and currents at turn-on and off. With a capacitiveload, turn-on at peak input voltage causes the maximumsurge current. Motor starting current runs 4 to 6 times thesteady state value. Generator action can boost voltagesabove the line value. Incandescent lamps have cold startcurrents 10 to 20 times the steady state value. Transformersgenerate voltage spikes when they are energized. Powerfactor correction circuits and switching devices createcomplex loads. In most cases, the simple CRL modelallows an approximate snubber design. However, there isno substitute for testing and measuring the worst case loadconditions.

SURGE CURRENTS IN INDUCTIVE CIRCUITS

Inductive loads with long L/R time constants causeasymmetric multi-cycle surges at start up (Figure 20). Trig-gering at zero voltage crossing is the worst case condition.The surge can be eliminated by triggering at the zero cur-rent crossing angle.

i (AM

PER

ES)

Figure 6.20. Start-Up Surge For Inductive Circuit

240VAC

20 MHY

i 0.1Ω

TIME (MILLISECONDS)

40

ZERO VOLTAGE TRIGGERING, IRMS = 30 A

0

90

80 160120 200

Core remanence and saturation cause surge currents.They depend on trigger angle, line impedance, core charac-teristics, and direction of the residual magnetization. Forexample, a 2.8 kVA 120 V 1:1 transformer with a 1.0ampere load produced 160 ampere currents at start-up. Softstarting the circuit at a small conduction angle reduces thiscurrent.

Transformer cores are usually not gapped and saturateeasily. A small asymmetry in the conduction angle causesmagnetic saturation and multi-cycle current surges.

AN1048/D

http://onsemi.com10

Steps to achieve reliable operation include: 1. Supply sufficient trigger current amplitude. TRIACs

have different trigger currents depending on theirquadrant of operation. Marginal gate current oroptocoupler LED current causes halfwave operation.

2. Supply sufficient gate current duration to achievelatching. Inductive loads slow down the main terminalcurrent rise. The gate current must remain above thespecified IGT until the main terminal current exceedsthe latching value. Both a resistive bleeder around theload and the snubber discharge current help latching.

3. Use a snubber to prevent TRIAC dVdt

c failure.

4. Minimize designed-in trigger asymmetry. Triggeringmust be correct every half-cycle including the first. Usea storage scope to investigate circuit behavior during thefirst few cycles of turn-on. Alternatively, get the gatecircuit up and running before energizing the load.

5. Derive the trigger synchronization from the line insteadof the TRIAC main terminal voltage. This avoidsregenerative interaction between the core hysteresisand the triggering angle preventing trigger runaway,halfwave operation, and core saturation.

6. Avoid high surge currents at start-up. Use a currentprobe to determine surge amplitude. Use a soft startcircuit to reduce inrush current.

DISTRIBUTED WINDING CAPACITANCE

There are small capacitances between the turns and lay-ers of a coil. Lumped together, they model as a single shuntcapacitance. The load inductor behaves like a capacitor atfrequencies above its self-resonance. It becomes ineffective

in controlling dVdt

and VPK when a fast transient such as that

resulting from the closing of a switch occurs. This problemcan be solved by adding a small snubber across the line.

SELF-CAPACITANCE

A thyristor has self-capacitance which limits dVdt

when the

load inductance is large. Large load inductances, high powerfactors, and low voltages may allow snubberless operation.

SNUBBER EXAMPLES

WITHOUT INDUCTANCE

Power TRIAC Example

Figure 21 shows a transient voltage applied to a TRIACcontrolling a resistive load. Theoretically there will be aninstantaneous step of voltage across the TRIAC. The onlyelements slowing this rate are the inductance of the wiringand the self-capacitance of the thyristor. There is an expo-nential capacitor charging component added along with adecaying component because of the IR drop in the snubber

resistor. The non-inductive snubber circuit is useful whenthe load resistance is much larger than the snubber resistor.

e(t o) E RSRS RL

et (1 et)

Figure 6.21. Non-Inductive Snubber Circuit

Vstep ERS

RS RL

RESISTORCOMPONENT

TIMEt = 0

eE τ = (RL + RS) CS

e

CS

RS

RL

E

CAPACITORCOMPONENT

Opto-TRIAC Examples

Single Snubber, Time Constant Design

Figure 22 illustrates the use of the RC time constantdesign method. The optocoupler sees only the voltageacross the snubber capacitor. The resistor R1 supplies thetrigger current of the power TRIAC. A worst case designprocedure assumes that the voltage across the powerTRIAC changes instantly. The capacitor voltage rises to63% of the maximum in one time constant. Then:

R1 CS 0.63 E

dVdt

s

where dVdt

sis the rated static dV

dt

for the optocoupler.

DESIGN dVdt

(0.63) (170)

(2400) (0.1 F) 0.45 Vs

φ CNTL

MOC3021

10 V/µs

TIME240 µs

0.63 (170)

L = 318 MHY

1 A, 60 Hz

RinVCC

C10.1 µF

170 V6 2.4 k18012N6073A

1 V/µs4

2

dVdt

(Vs)

Power TRIAC Optocoupler

0.99 0.35

Figure 6.22. Single Snubber For Sensitive Gate TRIACand Phase Controllable Optocoupler ( ρ = 0.67)

AN1048/D

http://onsemi.com11

The optocoupler conducts current only long enough totrigger the power device. When it turns on, the voltagebetween MT2 and the gate drops below the forward thresh-old voltage of the opto-TRIAC causing turn-off. The opto-

coupler sees dVdt

s when the power TRIAC turns off later

in the conduction cycle at zero current crossing. Therefore,it is not necessary to design for the lower optocoupler

dVdt

c rating. In this example, a single snubber designed

for the optocoupler protects both devices.

Figure 6.23. Anti-Parallel SCR Driver

(50 V/µs SNUBBER, ρ = 1.0)

120 V400 Hz

1 MHY

MCR265–4

430

100

MCR265–4

1N4001

0.022µF

1

51

1N4001

100VCC

6

5

4

3

2

MO

C30

31

Optocouplers with SCRs

Anti-parallel SCR circuits result in the same dVdt

across

the optocoupler and SCR (Figure 23). Phase controllableopto-couplers require the SCRs to be snubbed to their lowerdVdt

rating. Anti-parallel SCR circuits are free from the

charge storage behaviors that reduce the turn-off capabilityof TRIACs. Each SCR conducts for a half-cycle and has thenext half cycle of the ac line in which to recover. The turn-

off dVdt

of the conducting SCR becomes a static forward

blocking dVdt

for the other device. Use the SCR data sheet

dVdt

s rating in the snubber design.

A SCR used inside a rectifier bridge to control an ac loadwill not have a half cycle in which to recover. The availabletime decreases with increasing line voltage. This makes thecircuit less attractive. Inductive transients can be sup-pressed by a snubber at the input to the bridge or across theSCR. However, the time limitation still applies.

OPTO dVdt

cZero-crossing optocouplers can be used to switch

inductive loads at currents less than 100 mA (Figure 24).

However a power TRIAC along with the optocouplershould be used for higher load currents.

LOAD

CU

RR

ENT

(mA

RM

S)

Figure 6.24. MOC 3062 Inductive Load Current versus T A

CS = 0.01

CS = 0.001

NO SNUBBER

(RS = 100 Ω, VRMS = 220 V, POWER FACTOR = 0.5)

TA, AMBIENT TEMPERATURE (°C)

020

80

70

60

50

40

30

20

10

100908030 40 50 60 70

A phase controllable optocoupler is recommended with apower device. When the load current is small, a MAC97ATRIAC is suitable.

Unusual circuit conditions sometimes lead to unwanted

operation of an optocoupler in dVdt

c mode. Very large cur-

rents in the power device cause increased voltages betweenMT2 and the gate that hold the optocoupler on. Use of alarger TRIAC or other measures that limit inrush currentsolve this problem.

Very short conduction times leave residual charge in theoptocoupler. A minimum conduction angle allows recoverybefore voltage reapplication.

THE SNUBBER WITH INDUCTANCE

Consider an overdamped snubber using a large capacitorwhose voltage changes insignificantly during the timeunder consideration. The circuit reduces to an equivalentL/R series charging circuit.

The current through the snubber resistor is:

i VR

1 et ,

and the voltage across the TRIAC is:

e i RS.

The voltage wave across the TRIAC has an exponentialrise with maximum rate at t = 0. Taking its derivative givesits value as:

dVdt

0

V RSL

.

AN1048/D

http://onsemi.com12

Highly overdamped snubber circuits are not practicaldesigns. The example illustrates several properties: 1. The initial voltage appears completely across the circuit

inductance. Thus, it determines the rate of change of

current through the snubber resistor and the initial dVdt

.

This result does not change when there is resistance inthe load and holds true for all damping factors.

2. The snubber works because the inductor controls therate of current change through the resistor and the rateof capacitor charging. Snubber design cannot ignorethe inductance. This approach suggests that the snubbercapacitance is not important but that is only true forthis hypothetical condition. The snubber resistor shuntsthe thyristor causing unacceptable leakage when thecapacitor is not present. If the power loss is tolerable,dVdt

can be controlled without the capacitor. An

example is the soft-start circuit used to limit inrushcurrent in switching power supplies (Figure 25).

Figure 6.25. Surge Current Limiting Fora Switching Power Supply

dVdt

ERSL

SNUBBERL G

RS

E

Snubber With No C

E

C1AC LINERECTIFIER

BRIDGE

SNUBBERL

AC LINERECTIFIER

BRIDGE C1G

RS

TRIAC DESIGN PROCEDURE dVdt

c1. Refer to Figure 18 and select a particular damping

factor (ρ) giving a suitable trade-off between VPK and dVdt

.

Determine the normalized dVdt

corresponding to the chosen

damping factor.The voltage E depends on the load phase angle:

E 2 VRMS Sin () where tan1XLRL where

φ = measured phase angle between line V and load IRL = measured dc resistance of the load.Then

Z VRMSIRMS

RL2 XL

2 XL Z2 RL2 and

L XL

2 fLine.

If only the load current is known, assume a pure inductance.This gives a conservative design. Then:

L VRMS

2 fLine IRMSwhere E 2 VRMS.

For example:

E 2 120 170 V; L 120(8 A) (377 rps)

39.8 mH.

Read from the graph at ρ = 0.6, VPK = (1.25) 170 = 213 V.

Use 400 V TRIAC. Read dVdt (ρ0.6)

1.0.

2. Apply the resonance criterion:

0 spec dVdt dV

dt(P)E.

0 5 106 VS(1) (170 V)

29.4 103 r ps.

C 102 L

0.029 F

3. Apply the damping criterion:

RS 2ρ LC 2(0.6) 39.8 103

0.029 106 1400ohms.

dVdt

c SAFE AREA CURVE

Figure 26 shows a MAC15 TRIAC turn-off safeoperating area curve. Turn-off occurs without problem

under the curve. The region is bounded by static dVdt

at low

values of dIdt

c and delay time at high currents. Reduction

of the peak current permits operation at higher linefrequency. This TRIAC operated at f = 400 Hz, TJ = 125°C,and ITM = 6.0 amperes using a 30 ohm and 0.068 µFsnubber. Low damping factors extend operation to higher

dIdt

c, but capacitor sizes increase. The addition of a small,

saturable commutation inductor extends the allowedcurrent rate by introducing recovery delay time.

AN1048/D

http://onsemi.com13

dV dt( )

(V/

s)

µc

dIdt

cAMPERESMILLISECOND

Figure 6.26. versus T J = 125°CdVdt

cdIdt

c

100

10

0.15010

1

WITH COMMUTATION L

3014 18 22 26 34 38 42 46

– ITM = 15 A

dIdt

c 6 f ITM 103 Ams

MAC 16-8, COMMUTATIONAL L 33 TURNS # 18,52000-1A TAPE WOUND CORE 34 INCH OD

STATIC dVdt

DESIGN

There is usually some inductance in the ac main andpower wiring. The inductance may be more than 100 µH ifthere is a transformer in the circuit or nearly zero when ashunt power factor correction capacitor is present. Usuallythe line inductance is roughly several µH. The minimuminductance must be known or defined by adding a seriesinductor to insure reliable operation (Figure 27).

Figure 6.27. Snubbing For a Resistive Load

50 V/µs

0.33 µF10

LS1

100 µH20 A

340V 12 Ω

HEATER

One hundred µH is a suggested value for starting thedesign. Plug the assumed inductance into the equation forC. Larger values of inductance result in higher snubber

resistance and reduced dIdt

. For example:

Given E = 240 2 340 V.

Pick ρ = 0.3.

Then from Figure 18, VPK = 1.42 (340) = 483 V.Thus, it will be necessary to use a 600 V device. Using thepreviously stated formulas for ω0, C and R we find:

0 50 106 VS(0.73) (340 V)

201450 rps

C 1(201450)2 (100 106)

0.2464 F

R 2 (0.3) 100 106

0.2464 106 12 ohms

VARIABLE LOADS

The snubber should be designed for the smallest load

inductance because dVdt

will then be highest because of its

dependence on ω0. This requires a higher voltage device foroperation with the largest inductance because of the corre-sponding low damping factor.

Figure 28 describes dVdt

for an 8.0 ampere load at various

power factors. The minimum inductance is a component

added to prevent static dVdt

firing with a resistive load.

LRMAC 218A6FP

8 A LOAD

68 Ω 120 V60 Hz

0.033 µF

dVdt

s 100 Vs dV

dt

c 5 Vs

ρR L Vstep VPK

dvdtρ

Ω MHY V V V/µs

0.75 15 0.1 170 191 86

0.03 0 39.8 170 325 4.0

0.04 10.6 28.1 120 225 3.3

0.06 13.5 17.3 74 136 2.6

Figure 6.28. Snubber For a Variable Load

AN1048/D

http://onsemi.com14

EXAMPLES OF SNUBBER DESIGNS

Table 2 describes snubber RC values for dVdt

s.

Figures 31 and 32 show possible R and C values for a 5.0

V/µs dVdt

c assuming a pure inductive load.

Table 2. Static Designs

(E = 340 V, Vpeak = 500 V, ρ = 0.3)

dVdt

5.0 V/µs 50 V/µs 100 V/µs

LµH

CµF

ROhm

CµF

ROhm

CµF

ROhm

47 0.15 10100 0.33 10 0.1 20220 0.15 22 0.033 47500 0.068 51 0.015 1101000 3.0 11 0.033 100

TRANSIENT AND NOISE SUPPRESSION

Transients arise internally from normal circuit operationor externally from the environment. The latter is partic-ularly frustrating because the transient characteristics areundefined. A statistical description applies. Greater orsmaller stresses are possible. Long duration high voltagetransients are much less probable than those of loweramplitude and higher frequency. Environments with infre-quent lightning and load switching see transient voltagesbelow 3.0 kV.

Figure 6.29. Snubber Resistor For = 5.0 V/ µsdVdtc

R

(OH

MS)

S

80 A

40 A

20 A

0

0.6 A RMS 2.5 A

5 A

DAMPING FACTOR

10.90.80.70.60.50.40.30.20.1

10K

100

1000

10

10 A

PURE INDUCTIVE LOAD, V 120 VRMS,IRRM 0

Figure 6.30. Snubber Capacitor For = 5.0 V/ µsdVdtc

C

( F

)S

µ

0.001

2.5 A0.01

0.6 A

5 A

10 A

20 A

1

0.1

80 A RMS

40 A

0

DAMPING FACTOR

10.90.80.70.60.50.40.30.20.1

PURE INDUCTIVE LOAD, V 120 VRMS,IRRM 0

The natural frequencies and impedances of indoor acwiring result in damped oscillatory surges with typical fre-quencies ranging from 30 kHz to 1.5 MHz. Surge ampli-tude depends on both the wiring and the source of surgeenergy. Disturbances tend to die out at locations far awayfrom the source. Spark-over (6.0 kV in indoor ac wiring)sets the maximum voltage when transient suppressors arenot present. Transients closer to the service entrance or inheavy wiring have higher amplitudes, longer durations, andmore damping because of the lower inductance at thoselocations.

The simple CRL snubber is a low pass filter attenuatingfrequencies above its natural resonance. A steady statesinusoidal input voltage results in a sine wave output at thesame frequency. With no snubber resistor, the rate of rolloff approaches 12 dB per octave. The corner frequency is atthe snubber’s natural resonance. If the damping factor islow, the response peaks at this frequency. The snubberresistor degrades filter characteristics introducing anup-turn at ω = 1 / (RC). The roll-off approaches 6.0dB/octave at frequencies above this. Inductance in thesnubber resistor further reduces the roll-off rate.

Figure 32 describes the frequency response of the circuitin Figure 27. Figure 31 gives the theoretical response to a3.0 kV 100 kHz ring-wave. The snubber reduces the peakvoltage across the thyristor. However, the fast rise input

causes a high dVdt

step when series inductance is added to the

snubber resistor. Limiting the input voltage with a transientsuppressor reduces the step.

AN1048/D

http://onsemi.com15

Figure 6.31. Theoretical Response of Figure 33 Circuitto 3.0 kV IEEE 587 Ring Wave (R SC = 27.5 Ω)

V

(VO

LTS)

MT 2-

1

400

WITH 5 µHY

450 V MOVAT AC INPUT

WITH 5 µHY ANDWITHOUT 5 µHY

654210

TIME (µs)

3– 400

0

Figure 6.32. Snubber Frequency Response VoutVin

VOLT

AGE

GAI

N (d

B)

FREQUENCY (Hz)10K

WITH 5 µHY

WITHOUT 5µHY

– 40

0.33 µF

10

5 µH

12

Vin

100 µH

1M

+ 10

0

100K

– 30

– 20

– 10

Vout

The noise induced into a circuit is proportional to dVdt

when coupling is by stray capacitance, and dIdt

when the

coupling is by mutual inductance. Best suppressionrequires the use of a voltage limiting device along with arate limiting CRL snubber.

The thyristor is best protected by preventing turn-on

from dVdt

or breakover. The circuit should be designed for

what can happen instead of what normally occurs.In Figure 30, a MOV connected across the line protects

many parallel circuit branches and their loads. The MOV

defines the maximum input voltage and dIdt

through the load.

With the snubber, it sets the maximum dVdt

and peak voltage

across the thyristor. The MOV must be large because thereis little surge limiting impedance to prevent its burn-out.

In Figure 32, there is a separate suppressor across eachthyristor. The load impedance limits the surge energy deliv-ered from the line. This allows the use of a smaller devicebut omits load protection. This arrangement protects eachthyristor when its load is a possible transient source.

Figure 6.33. Limiting Line Voltage

VMAX

Figure 6.34. Limiting Thyristor Voltage

It is desirable to place the suppression device directlyacross the source of transient energy to prevent the induc-tion of energy into other circuits. However, there is noprotection for energy injected between the load and its con-trolling thyristor. Placing the suppressor directly acrosseach thyristor positively limits maximum voltage and snub-

ber discharge dIdt

.

EXAMPLES OF SNUBBER APPLICATIONS

In Figure 35, TRIACs switch a 3 phase motor on and offand reverse its rotation. Each TRIAC pair functions as aSPDT switch. The turn-on of one TRIAC applies the differ-ential voltage between line phases across the blockingdevice without the benefit of the motor impedance toconstrain the rate of voltage rise. The inductors are added to

prevent static dVdt

firing and a line-to-line short.

AN1048/D

http://onsemi.com16

Figure 6.35. 3 Phase Reversing Motor

MOC3081

SNUBBERALL MOV’S ARE 275VRMSALL TRIACS AREMAC218A10FP

0.15µF

22 Ω2 W

WIREWOUND

43

4

6

G1

2

91

1/3 HP208 V

3 PHASE

100 µH

REV

φ1

91

G

12

300

6

FWD

4

SNUBBER

N

φ2

φ3

MOC3081

91

G

12

300

64

SNUBBER

MOC3081

91

G

12

300

64

SNUBBER

MOC3081

91

G

12

300

46

SNUBBER

REV

FWD

100 µHMOC3081

SNU

BBER

AN1048/D

http://onsemi.com17

Figure 36 shows a split phase capacitor-run motor withreversing accomplished by switching the capacitor in serieswith one or the other winding. The forward and reverseTRIACs function as a SPDT switch. Reversing the motorapplies the voltage on the capacitor abruptly across theblocking thyristor. Again, the inductor L is added to prevent

dVdt

s firing of the blocking TRIAC. If turn-on occurs, the

forward and reverse TRIACs short the capacitors (Cs)resulting in damage to them. It is wise to add the resistor RSto limit the discharge current.

Figure 6.36. Split Phase Reversing Motor

5.6

3.75330 V

MOTOR1/70 HP0.26 A

2N6073

115

500 µH

LS

REV

46 V/µsMAX

FWD

0.191

CS

0.1

RS

91

Figure 37 shows a “ tap changer.” This circuit allows theoperation of switching power supplies from a 120 or 240vac line. When the TRIAC is on, the circuit functions as aconventional voltage doubler with diodes D1 and D2 con-ducting on alternate half-cycles. In this mode of operation,

inrush current and dIdt

are hazards to TRIAC reliability.

Series impedance is necessary to prevent damage to theTRIAC.

The TRIAC is off when the circuit is not doubling. In thisstate, the TRIAC sees the difference between the line volt-age and the voltage at the intersection of C1 and C2. Tran-

sients on the line cause dVdt

s firing of the TRIAC. High

inrush current, dIdt

, and overvoltage damage to the filter

capacitor are possibilities. Prevention requires the additionof a RC snubber across the TRIAC and an inductor in serieswith the line.

THYRISTOR TYPES

Sensitive gate thyristors are easy to turn-on because oftheir low trigger current requirements. However, they have

less dVdt

capability than similar non-sensitive devices. A

non-sensitive thyristor should be used for high dVdt

.

TRIAC commutating dVdt

ratings are 5 to 20 times less

than static dVdt

ratings.

Figure 6.37. Tap Changer For Dual VoltageSwitching Power Supply

–C1

+

RL

0 G

CSRS

120 V

240 V

SNUBBER INDUCTOR

120 VACOR

240 VAC

D1D2

D4D3

–C2

+

Phase controllable optocouplers have lower dVdt

ratings

than zero crossing optocouplers and power TRIACs. Theseshould be used when a dc voltage component is present, orto prevent turn-on delay.

Zero crossing optocouplers have more dVdt

capability than

power thyristors; and they should be used in place of phasecontrollable devices in static switching applications.

APPENDIX A

MEASURING dVdt

s

Figure 38 shows a test circuit for measuring the static dVdt

of power thyristors. A 1000 volt FET switch insures that thevoltage across the device under test (D.U.T.) rises rapidlyfrom zero. A differential preamp allows the use of aN-channel device while keeping the storage scope chassisat ground for safety purposes. The rate of voltage rise isadjusted by a variable RC time constant. The chargingresistance is low to avoid waveform distortion because ofthe thyristor’s self-capacitance but is large enough to pre-

vent damage to the D.U.T. from turn-on dIdt

. Mounting the

miniature range switches, capacitors, and G-K networkclose to the device under test reduces stray inductance andallows testing at more than 10 kV/µs.

AN1048/D

http://onsemi.com18

Figure 6.38. Circuit For Static Measurement of Power ThyristorsdVdt

X100 PROBE

X100 PROBE

DIFFERENTIALPREAMP

MOUNT DUT ONTEMPERATURE CONTROLLEDCµ PLATE

DUT

2

1G

RGK

2 W20 k

VDRM/VRRM SELECT 2 W

27

0.33 1000 V 0.0471000 V

100010 WATT

WIREWOUND

10001/4 W

562 W

VERNIER

822 W

1002 W

470 pF

0.001

0.005

0.01

0.047

0.1

0.47

1.2 MEG2 W EACH1 MEG

POWER2 W

TEST

MTP1N100

ALL COMPONENTS ARE NON-INDUCTIVE UNLESS OTHERWISE SHOWN

0–1000 V10 mA

1N967A18 V

1N914

f = 10 HzPW = 100 µs50 Ω PULSEGENERATOR

20 V

dVdt

APPENDIX B

MEASURING dVdt

c

A test fixture to measure commutating dVdt

is shown in

Figure 39. It is a capacitor discharge circuit with the loadseries resonant. The single pulse test aids temperature con-trol and allows the use of lower power components. Thelimited energy in the load capacitor reduces burn and shockhazards. The conventional load and snubber circuit pro-vides recovery and damping behaviors like those in theapplication.

The voltage across the load capacitor triggers the D.U.T.It terminates the gate current when the load capacitor volt-age crosses zero and the TRIAC current is at its peak.

Each VDRM, ITM combination requires different compo-nents. Calculate their values using the equations given inFigure 39.

Commercial chokes simplify the construction of the nec-essary inductors. Their inductance should be adjusted byincreasing the air gap in the core. Removal of the magneticpole piece reduces inductance by 4 to 6 but extends the cur-rent without saturation.

The load capacitor consists of a parallel bank of 1500Vdc non-polar units, with individual bleeders mounted ateach capacitor for safety purposes.

An optional adjustable voltage clamp prevents TRIACbreakdown.

To measure dVdt

c, synchronize the storage scope on the

current waveform and verify the proper current amplitudeand period. Increase the initial voltage on the capacitor tocompensate for losses within the coil if necessary. Adjustthe snubber until the device fails to turn off after the firsthalf-cycle. Inspect the rate of voltage rise at the fastestpassing condition.

AN1048/D

http://onsemi.com19

–

+

–

+

2 W51+ 5CS

2N3904

0.1

2.2 k1/2SYNC

TRIACUNDERTEST

562 WATT

CASECONTROLLED

HEATSINK

0.1

2N3906

– 5 G1

2

PEARSON301 X

2N3904

1/2 W120

2N3906

1201/2 W

2N3906

2N3904RS

1/2 W

– +

0.22270 k 1N5343

7.5 V

0.22Q1Q3

2N3906

2N3904

1/2 W360

1 k

– 5 + 5

NON-INDUCTIVERESISTOR DECADE

0–10 k, 1 Ω STEP

2 W51 k

51 k2 W

HG = W AT LOW LD10-1000-1000

+ CLAMP

MR760

– CLAMP

62 µF1 kV

LL

6.2 MEG 2W

150 k

910 k2 W

2 W

910 k

RL

Q3 Q1

70 mA

1.5 kV

TRIAD C30X50 H, 3500 Ω

2 W51

6.2 MEG 2W

360

1 k

270 k

2N39042N3906

Q3 Q1

0-1

kV 2

0 m

A

C

(NO

N-P

OLA

R)

L

MR

760

MR

760

2.2

M2.

2 M

2.2

M, 2

W

2.2

M, 2

W2.

2 M

2.2

M

0.01

Figure 6.39. Test Circuit For Power TRIACsdVdt

c

µ

CL IPK

W0 VCi

Ip T

2 VCiLL

VCiW0 IPK

T2

4 2 CLW0

I

LLdIdtc 6f IPK 106

As

dVdt

CAP

ACIT

OR

DEC

ADE

1–10

F

, 0.0

1–1

F,

100

pF–

0.0

1

Fµ

µ

0.01

AN1048/D

http://onsemi.com20

APPENDIX CdVdt

DERIVATIONS

DEFINITIONS

1.0 RT RL RS Total Resistance

1.1 M RSRT

Snubber Divider Ratio

1.2 0 1L CS Undamped Natural Frequency

Damped Natural Frequency

1.3 RT2 L

Wave Decrement Factor

1.4 χ2 12 LI2

12 CV2

Initial Energy In InductorFinal Energy In Capacitor

1.5 χ IE

LC Initial Current Factor

1.6 ρ RT2

CL

0 Damping Factor

1.7 V0L E RS I Initial Voltage drop at t 0

across the load

1.8 ICS

–E RL

L

dVdt

0 Initial instantaneous dV

dtat t 0, ignoring

any initial instantaneous voltage step att 0 because of IRRM

1.9 dVdt

0 VOL

RTL . For all damping conditions

2.0 When I 0, dVdt

0

E RSL

dVdt

max Maximum instantaneous dV

dt

tmax Time of maximum instantaneous dVdt

tpeak Time of maximum instantaneous peakvoltage across thyristor

Average dVdt

VPK tPK Slope of the secant line

from t 0 through VPKVPK Maximum instantaneous voltage across the

thyristor.

CONSTANTS (depending on the damping factor):

2.1 No Damping (ρ 0) 0RT ρ 0

2.2 Underdamped (0 ρ 1)

02 2 0 1 ρ2

2.3 Critical Damped (ρ 1)

0, 0, R 2 LC , C 2

RT

2.4 Overdamped (ρ 1)

2 02 0 ρ2 1

Laplace transforms for the current and voltage in Figure 40are:

3.0 i(S)ELSI

S2SRTL 1

LC

; e ES

S V0L

S2RTL

S 1LC

INITIAL CONDITIONSI IRRMVCS 0

Figure 6.40. Equivalent Circuit for Load and Snubber

t = 0I

RL L

CS

RSe

The inverse laplace transform for each of the conditionsgives:

UNDERDAMPED (Typical Snubber Design)

4.0 e E V0LCos (t)

sin (t)e t

sin (t) et

4.1 dedt

V0L2 Cos (t)

(2–2)

sin (t)e–t

Cos (t)–

sin (t) e–t

4.2 tPK 1 tan1

2 V0L

V0L22

When M 0, RS 0, I 0 : tPK

AN1048/D

http://onsemi.com21

4.3 VPK E

0 tPK 0

2 V0L2 2 V0L

When I 0, RL 0, M 1:

4.4VPK

E (1 e tPK)

Average dVdt

VPKtPK

4.5 tmax 1ATN (2 V0L (2 32))

V0L(3 32) (2 2)

4.6 dVdt

max V0L

2022 V0L 2 e– tmax

NO DAMPING

5.0 e E (1 Cos (0t)) IC 0

sin (0t)

5.1 dedt

E 0 sin (0t) IC

Cos (0t)

5.2 dVdt

0 I

C 0 when I 0

5.3 tPK tan1 I

CE 0

0

5.4 VPK E E2 I2

02C2

5.5 dVdtAVG

VPKtPK

5.6 tmax 10tan10 EC

I 1

02

when I 0

5.7 dVdt

max I

CE202 C2I2 0E when I0

CRITICAL DAMPING

6.0 e E V0L(1 t)et tet

6.1 dedt

VOL (2 t) (1 t) et

6.2 tPK

2

2 V0L

V0L

6.3 VPK E – V0L (1– tPK)– tPK e– tPK

6.4 Average dVdt

VPKtPK

When I 0, RS 0, M 0

e(t) rises asymptotically to E. tPK and average dVdt

do not exist.

6.5 tmax 3V0L 2

2V0L

When I 0, tmax 0

ForRSRT

34,

then dVdt max

dVdt

0

6.6 dVdt

max

V0L (2– tmax) (1– tmax) e– tmax

APPENDIX D

SNUBBER DISCHARGE dIdt

DERIVATIONS

OVERDAMPED

1.0 i VCS LS

–t sinh (t)

1.1 iPK VCS

CSLS e– tPK

1.2 tPK 1 tanh –1

CRITICAL DAMPED

2.0 i VCSLS

te–t

2.1 iPK 0.736VCSRS

2.2 tPK 1

AN1048/D

http://onsemi.com22

UNDERDAMPED

3.0 i VCS LS

e–t sin (t)

3.1 iPK VCS

CSLS

e– tPK

3.2 tPK 1

tan –1

Figure 6.41. Equivalent Circuit for Snubber Discharge

LSRS

CSVCS i

t = 0

INITIAL CONDITIONS :i 0, VCS INITIAL VOLTAGE

NO DAMPING

4.0 i VCS LS

sin (t)

4.1 iPK VCS

CSLS

4.2 tPK

2

BIBLIOGRAPHY

Bird, B. M. and K. G. King. An Introduction To PowerElectronics. John Wiley & Sons, 1983, pp. 250–281.

Blicher, Adolph. Thyristor Physics. Springer-Verlag, 1976.

Gempe, Horst. “Applications of Zero Voltage CrossingOptically Isolated TRIAC Drivers,” AN982, Motorola Inc.,1987.

“Guide for Surge Withstand Capability (SWC) Tests,”ANSI 337.90A-1974, IEEE Std 472–1974.

“IEEE Guide for Surge Voltages in Low-Voltage AC PowerCircuits,” ANSI/IEEE C62.41-1980, IEEE Std 587–1980.

Ikeda, Shigeru and Tsuneo Araki. “The dIdt

Capability of

Thyristors,” Proceedings of the IEEE, Vol. 53, No. 8,August 1967.

Kervin, Doug. “The MOC3011 and MOC3021,” EB-101,Motorola Inc., 1982.

McMurray, William. “Optimum Snubbers For PowerSemiconductors,” IEEE Transactions On Industry Applica-tions, Vol. IA-8, September/October 1972.

Rice, L. R. “Why R-C Networks And Which One For YourConverter,” Westinghouse Tech Tips 5-2.

“Saturable Reactor For Increasing Turn-On SwitchingCapability,” SCR Manual Sixth Edition, General Electric,1979.

Zell, H. P. “Design Chart For Capacitor-Discharge PulseCircuits,” EDN Magazine, June 10, 1968.

AN1048/D

http://onsemi.com23

Notes

AN1048/D

http://onsemi.com24

USA/EUROPE Literature Fulfillment :Literature Distribution Center for ON SemiconductorP.O. Box 5163, Denver, Colorado 80217 USAPhone : 303–675–2175 or 800–344–3860 Toll Free USA/CanadaFax: 303–675–2176 or 800–344–3867 Toll Free USA/CanadaEmail : [email protected]

Fax Response Line *: 303–675–2167800–344–3810 Toll Free USA/Canada*To receive a Fax of our publications

N. America Technical Support : 800–282–9855 Toll Free USA/Canada

ON Semiconductor and are trademarks of Semiconductor Components Industries, LLC (SCILLC). SCILLC reserves the right to make changeswithout further notice to any products herein. SCILLC makes no warranty, representation or guarantee regarding the suitability of its products for any particularpurpose, nor does SCILLC assume any liability arising out of the application or use of any product or circuit, and specifically disclaims any and all liability,including without limitation special, consequential or incidental damages. “Typical” parameters which may be provided in SCILLC data sheets and/orspecifications can and do vary in different applications and actual performance may vary over time. All operating parameters, including “Typicals” must bevalidated for each customer application by customer’s technical experts. SCILLC does not convey any license under its patent rights nor the rights of others.SCILLC products are not designed, intended, or authorized for use as components in systems intended for surgical implant into the body, or other applicationsintended to support or sustain life, or for any other application in which the failure of the SCILLC product could create a situation where personal injury ordeath may occur. Should Buyer purchase or use SCILLC products for any such unintended or unauthorized application, Buyer shall indemnify and holdSCILLC and its officers, employees, subsidiaries, affiliates, and distributors harmless against all claims, costs, damages, and expenses, and reasonableattorney fees arising out of, directly or indirectly, any claim of personal injury or death associated with such unintended or unauthorized use, even if such claimalleges that SCILLC was negligent regarding the design or manufacture of the part. SCILLC is an Equal Opportunity/Affirmative Action Employer.

PUBLICATION ORDERING INFORMATIONASIA/PACIFIC : LDC for ON Semiconductor – Asia SupportPhone : 303–675–2121 (Tue–Fri 9:00am to 1:00pm, Hong Kong Time)Email : ONlit–[email protected]

JAPAN : ON Semiconductor, Japan Customer Focus Center4–32–1 Nishi–Gotanda, Shinagawa–ku, Tokyo, Japan 141–8549Phone : 81–3–5487–8345Email : [email protected]

ON Semiconductor Website: http://onsemi.com

For additional information, please contact your local Sales Representative.

AN1048/D

This datasheet has been download from:

www.datasheetcatalog.com

Datasheets for electronics components.