See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/350024137

Assessment of Catastrophic Failure for Phase Control Thyristors in AC Pulsed

Applications

Article in IEEE Transactions on Industry Applications · March 2021

DOI: 10.1109/TIA.2021.3065896

CITATIONS READS

3 47

6 authors, including:

Fernando Amaral Sidelmo M Silva

Federal University of Minas Gerais Federal University of Minas Gerais
18 PUBLICATIONS 124 CITATIONS 94 PUBLICATIONS 1,232 CITATIONS

SEE PROFILE SEE PROFILE

Cláudio Alvares Conceição Braz Cardoso Filho

8 PUBLICATIONS 39 CITATIONS
Federal University of Minas Gerais
248 PUBLICATIONS 2,385 CITATIONS
SEE PROFILE
SEE PROFILE

Some of the authors of this publication are also working on these related projects:

TUPF - True Unity Power Factor rectifier View project

Insulation failures detection and protection View project

All content following this page was uploaded by Sidelmo M Silva on 27 April 2021.

The user has requested enhancement of the downloaded file.

 Universidade Federal de Minas Gerais Departamento de Engenharia Elé trica

Assessment of Catastrophic Failure for Phase Control Thyristors in ac Pulsed

Applications

Fernando Amaral; Sidelmo Silva; Claudio Conceição; Gustavo Fontoura; Cleber Inacio; Braz
Cardoso

Published in:
IEEE Transactions on Industry Applications, March 2021

Print ISSN: 0093-9994

Electronic ISSN: 1939-9367

F. Amaral, S. Silva, C. Conceicao, G. Fontoura, C. Inacio and B. Cardoso, "Assessment of

Catastrophic Failure for Phase Control Thyristors in ac Pulsed Applications," in IEEE
Transactions on Industry Applications

DOI: 10.1109/TIA.2021.3065896

General Rights
Copyright and moral rights for publications made accessible in the public portals are retained by the authors and/or
other copyright owners and it is a condition of accessing publications that users recognize and abide by the legal
requirements associated with these rights. Users may:
• download and print one copy of the publication from the public portal for the purpose of private study or
research.
• not further distribute the material or use it for any proCit-making activity or commercial gain
• freely distribute the URL identifying the publication in the public portal
If you believe that this document breaches copyright please contact us providing details, and we will remove access
to the work immediately and investigate your claim.
 1

Assessment of Catastrophic Failure for Phase

Control Thyristors in AC Pulsed Applications

Fernando V. Amaral, Student Member, IEEE, Sidelmo M. Silva, Member, IEEE, Claudio A. Conceição,
Non-member, Gustavo F. Guimarães, Non-member and Braz J. Cardoso, Member, IEEE

°C. If this threshold is exceeded, then the subsequent

Abstract—This paper presents a theoretical development,
supported by experimental results, which is useful for the
evaluation of the ride through capability of phase control
thyristors that should operate under high multi-cycle AC, power
frequency current surges with different amplitudes between
cycles. Despite being commonly found in the field of metals
industry applications, this is not a standard data sheet test
condition that can be readily used to predict if the device can
handle a given surge profile. The present work is devoted to
contributing to the fulfillment of this gap. A recursive, discrete-
time method is used for the calculation of instantaneous virtual
junction temperature of the devices tested. Such method is
advantageous over the traditional equivalent power pulse
technique since for any given time step only the immediately
previous calculation is needed instead of the overall series. It is
shown that the results obtained with these two methods are
practically identical. Besides, a complete study based on a
commercial power semiconductor device is given along with
experimental results that support the application of the
assessment method developed.
Index Terms—Power semiconductor switches, pulsed power
systems, prediction methods, industrial electronics, design
engineering
I. INTRODUCTION
reliability of the device operation cannot be guaranteed.
Thyristors have been frequently used in crowbar circuits in
which a capacitor bank is discharged into a load, like army
applications [6], electric guns [7], [8] and electromagnetic
launchers and flash lamps [9]. By forming an extremely low
impedance short-circuit they provide a freewheel path for the
load [10].
In the field of metals industry applications, these devices
are often used in pulsed power tools intended for removal of
surface layers and drilling [11], [12]. Another application of
power thyristor is as bypass switch for series-connected power
conditioners, for example dynamic voltage restorers [13]. In
this case, a bidirectional AC switch, composed by two
antiparallel-connected thyristors, is kept closed until a voltage
sag is detected on the supply side – when the AC switch then
opens, leading to the operation of the series power converter to
the load side voltage restoration. However, should a short-
circuit take place at the protected load side, the bypass switch
would be exposed to a huge current surge that lasts until the
operation of an upstream protective device – in general, an
electromechanical circuit breaker (CB).
More recently, it has been proposed in the literature that
such electronic AC switch may be applied for arc-flash

A MONG power semiconductor devices, bipolar thyristor
has been one of the most powerful ever produced, with
clear advantages in fabricating a high-power device in a single
wafer and the highly reliable press-pack (capsule) packaging
technology [1]. Typical press-pack silicon-controlled rectifier
(SCR) is commercially available up to 105 kA (non-repetitive
surge current) or 8.5 kV (repetitive reverse voltage) [2] at the
time of writing this paper. Some manufacturers offer even
highly rated devices, for example 12 kV [3].
The ratings of power semiconductor devices are extremely
important because of the high probability of catastrophic
device failure if they are exceeded [4], [5]. Almost all the
thyristor ratings are based on its virtual junction temperature
Tvj, which maximum calculated value is limited by the
blocking capability of the device. This is normally set at 125
suppression in switchgear and motor control centers (MCCs)
of metal industry facilities [14]. In this case, though, the AC
switch is shunt-connected and remains opened until a
developing arcing fault is detected into the switchgear. Once
the AC switch is closed, it provides an alternative, low
impedance path to the fault current, eliminating the arc until
the operation of an upstream protective device.
Since thyristors have been primarily designed for the
continuous operating mode, typical of power supplies, motor
drives or power conversion systems, their application in
pulsed switching is mostly unknown and therefore requires
special considerations. The best choice of power
semiconductor devices is dependent on the specific pulse
switching application, since in this case the current waveform
often goes beyond the ratings in data sheets [15]. In the bypass

This project was supported by the Brazilian agencies CAPES, CNPq and FAPEMIG, and was funded by Petróleo Brasileiro (Petrobras) under Agência
Nacional de Energia Elétrica (Aneel) Research Grant PD-00553-0052/2017.
F. V. Amaral (corresponding author) is with the Graduate Program in Electrical Engineering, Federal University of Minas Gerais – Av. Antônio Carlos 6627,
31270-901, Belo Horizonte, MG, Brazil (e-mail: famaral@ieee.org).
S. M. Silva and B. J. Cardoso are with the Department of Electrical Engineering, Federal University of Minas Gerais – Av. Antônio Carlos, 6627, 31270-901,
Belo Horizonte, MG, Brazil (e-mail: sidelmo@ufmg.br, braz.cardoso@ieee.org).
C. A. Conceição is with the Department of Electrical Maintenance, Petróleo Brasileiro S.A. – Av. Refinaria Gabriel Passos 690, 32669-205, Betim, MG,
Brazil (e-mail: claudioac@petrobras.com.br).
G. F. Guimarães is with Research Development Foundation (Fundep) – Av. Antônio Carlos 6627, 31270-901, Belo Horizonte, MG, Brazil (e-mail:
gustavofontoura.eng@outlook.com).

switch and arc-flash suppression applications mentioned
above, for example, the electronic switch must conduct a high
AC current when a short-circuit occurs, but this surge lasts for
no more than 3 to 5 power-frequency cycles, which is the
typical time period that the upstream CB takes to complete its
elimination [16]. In this scenario, the selection of the devices
based only on its non-repetitive ratings available in the data
sheets, like ITSM, is not good enough to the optimum
realization of the solution. These ratings are not readily
applicable.
Even though the operation of the AC switch under such
circumstances is relatively rare, the large current surge that
arises when it does occur may lead to the catastrophic failure
of the thyristors that compose the AC switch. In this sense,
these components must be designed considering their
anticipated withstanding to the expected short-circuit current
that may be clearly part of their mission profile. This scenario
is worsened by the fact that short-circuit currents may include
a DC-level component, leading to time-varying current
amplitude through the AC switch and consequently in
asymmetry between the current waveforms that each of the
two thyristors is subjected to.
The main contribution of this work is to provide guidance
for the specification of phase control thyristors (PCT) to be
used in AC pulsed power applications, given the current
waveform that it will be subjected to. The calculation strategy
presented in this paper allows for the prediction of the exact
catastrophic failure limit for a given PCT subjected to a given
AC current surge. Moreover, the virtual temperature junction
rise of the device is calculated using an original proposition
based on recursive convolution method. Simulation and
experimental results give support to the theoretical
development that will be presented in the next section.
II. ELECTROTHERMAL BEHAVIOR OF A THYRISTOR
A.Non-repetitive Ratings
Parameter “peak non-repetitive surge current”, ITSM, is the
maximum allowable non-repetitive surge current the device
will withstand at a specified pulse width. Manufacturers often
specify this parameter in the data sheets as the maximum 10
ms half-sine wave of current, following maximum load current
(case temperature Tc = 125 ºC) that the thyristor can conduct
without the device experiencing thermal runaway and the
silicon melting. Yet the temperature may temporarily increase
to as much as 400 ºC [17], [18]. This parameter is empirically
determined by the manufacturer by viewing the forward
voltage characteristic of the thyristor at high currents and
observing the onset of thermal runaway. Exceeding this limit
will damage the device. In other words, the surge current
rating is determined to be the maximum peak surge current at
which none of a significant sample of devices failed or
degraded in voltage classification [19]. Because the device is
so hot following the surge current pulse, it loses the ability to
block voltage, and therefore this rating is for zero reapplied
voltage. Also, the temperature excursion is so high that the
device will fail due to temperature cycling wear out if this
level of current is repeated several times in the life of the
device. For a surge duration of 8.3 ms, ITSM is about 10 %
higher than for a surge duration of 10 ms.
2
Manufacturers commonly present a curve relating
permissible on-state current surge peak as a function of the
surge duration. For events lasting longer than 10 ms it is
assumed that the current waveform is a series of half sine
waves of 10 or 8.3 ms (50 Hz or 60 Hz) occurring at a rate of
one every 20 or 16.6 ms [20]. The points of this curve are
obtained by calculating the current that gives the same peak
temperature as the 10 ms half-sine wave derived from the non-
repetitive physical rating (ITSM) of the device [18]. The
permissible on-state current surge peak is also called “limiting
overload characteristics” (IT(OV)) by some manufacturers. An
example can be seen in the data sheet of reference [21] for part
number SKT-553/18E (41 mm, 1.8 kV, 553 A) from Semikron,
which is the device that was adopted for the experimental
evaluation presented in this work.
B.Virtual Junction Temperature
The thermal properties of a thyristor are described by its
transient thermal impedance characteristic curve, which allows
the calculation of the temperature response to a unit power
step (1 W) [22]. It is obtained by thermal modelling of the
device using finite element method and is verified by
measurements [18]. Because of heat storage effects due to
thermal capacity, the transient thermal impedance for short
pulse time durations is much lower than that for long pulse
times or steady state.
The use of transient thermal impedance curve in rating
calculations allows the consideration of cyclic junction
temperature excursions. At power frequencies, these cyclic
excursions are appreciable because the current pulse widths
that occur in phase-controlled circuits are of the same order of
magnitude as the thermal response time of the semiconductor
crystal. A complete expression for time-dependent transient
thermal impedance Zth(j-c)(t) can be modelled as the sum of
some exponential terms [23], [24]
n
Rth,i ∙ (1 − e −t/Tth,i )
∑
Zth( j − c)(t) = (1)
i=1
where the subscript “j-c” stands for “junction to case”. The
curve fit parameters Rth,i and Tth,i, namely “thermal resistance”
and “thermal time constant”, respectively, are normally given
in a table in the data sheet [25], [26]. If not, they can be easily
estimated using curve fitting techniques.
The method that has been historically used for virtual
junction temperature calculation is based on (2), in which
equivalent power pulse waveform should be utilized for
effortlessness [18], [27]. This approach takes advantage of the
linear properties of the thermal system, for which the
superposition principle is valid. This means that a waveform
can be approximated with any desired accuracy by several step
waves [28].
m n −(tm − tμ−1)
( )
Tvj(tm) = Tc + Tth,i
∑ ∑
(Pμ − Pμ−1) ∙ Rth,i ∙ 1−e (2)
μ=1 i=1
where Tc is the device case temperature and Pµ is the
equivalent rectangular pulse power at instant tµ. The initial
junction temperature is equal to Tc assuming thermal
equilibrium before current surge. For pulsed power
applications up to 100 ms, no stationary temperature
difference develops between junction and case and therefore
 3

Tc may be appropriately considered constant during this evaluated in this paper is 9 kA for Tvj = 25 °C. Since this work
interval [18]. will be executed using 60 Hz instead of 50 Hz, ITSM is 10 %
However, the standard method of providing sine wave higher, i.e. 9.9 kA. The initial temperature considered will be
ratings for thyristors for time dependent power dissipation is 40 °C instead of 25 °C, since this is a common practice for
the convolution integral shown below [29], [30]: industrial systems.
t If logarithm scale is adopted for both axes and they are

∫
Tvj (t) = Tc + i(τ) ∙ v(τ) ∙ Żth( j − c)(t − τ) ∙ dτ (3) swapped, then the curve shown in Fig. 1 will be obtained.
The least square regression method may be applied to
0 obtain the coefficients a and b of the linear function given by
where i and v are respectively the device current and voltage. (6).
C.Voltage Drop log(t) = − a ∙ log(IT(OV )) + b (6)
Device voltage drop is often given in the data sheet as two For the points shown in Fig. 1, (7) is obtained.
separate, current-dependent curves, one for Tvj = 25 ºC and the TABLE I
other for Tvj = 125 ºC. Some manufacturers give an equation POINTS SELECTED FROM THE CURVE IN SKT-553/18E DATA SHEET
that is suitable for the calculation of the voltage drop as a t [ms] IT(OV) [kA] a
function of the current for a fixed temperature as follows: 109.9
407.9
v(i ) = A + B ∙ ln(i ) + C ∙ i + D ∙ i (4) 100 6.9
where A, B, C, and D are given constants. No information is 300 5.9
available to allow for simultaneous incorporation of current 10004.9
and temperature effects on the device voltage drop. Reference aFor no re-applied voltage after surge, i.e. 0·VRRM.

[31] proposes an empirical method which depends on data

from experimental testing. This is impractical since it requires
that a thyristor be submitted to a high current test bench before
being considered for a new design. Reference [32] presents an
alternative method that rely on fitting curves to the 1000
information given in the data sheets according to (5)
v(i, Tvj) = (A0 + AT ∙ Tvj) + (B0 + BT ∙ Tvjk ) ∙ i + (C0 + CT ∙ Tvj) ∙ i (5)

where A0, AT, B0, BT, k, C0, and CT are constants to be

calculated according to the curves given by the manufacturer,
t [ms]

i.e., v(i, Tvj = 25 ºC) and v(i, Tvj = 125 ºC). This approach will 100
be adopted in this work.
D.Failure under Surge Conditions
Reference [33] has presented a graph containing a “survival
curve” that relates the number of presumable safe current
surges to a given virtual junction temperature variation caused 10
by the event. This curve is a straight line on the log-log scale, 1 10
but it must be empirically verified for any thyristor model, IT(OV) [kA]
which is a serious drawback. Furthermore, the presented Fig. 1. A plot of the limiting overload characteristics of part number
technique is applicable only to cases in which the temperature SKT-553/18E using logarithmic scale for both axes, for t > 10 ms.
variation is the same for all the surges, which is not always
true. Next section proposes an alternative technique that is −log(9.9) 1 log(10)
based exclusively on information contained in the data sheet −log(7.9) 1
and which is applicable to waveforms composed by a log(40)
combination of AC semi cycles not necessarily with the same −log(6.9) 1 ∙ [a b] = log(100) (7)
amplitudes. Additionally, this procedure does not depend on
the calculation of the virtual junction temperature elevation, −log(5.9) 1 log(300)
which is not even measurable, but rather an estimate.
−log(4.9) 1 log(1000)
III. PROPOSED METHOD FOR CATASTROPHIC FAILURE By applying the pseudo-inverse, one can easily find the
PREDICTION result shown by (8).

{b = 7.643
Despite the diversity verified between the form of a = 6.693
presentation of permissible on-state current surge peak curve
(8)
by different manufacturers, it can be separated into two parts: t Now, (6) may be rewritten by applying basic logarithm
< 16.6 ms (half sine wave under 8.3 ms) and t > 16.6 ms (one properties as
or more consecutive half sine waves). For t > 16.6 ms, the a
points shown in Table I were promptly read from the curve
IT(OV )∙t =c (9)
given by reference [21]. The value of ITSM for the thyristor where c = 10b = 4.39·107. For a given current waveform
 4

composed by half sine waves of different amplitudes, one can V. SIMULATION RESULTS
account for the effect of each one as will be demonstrated
A.Catastrophic Failure Prediction
next.
The results that will be presented in this section will be
A.Single Half Sine Wave based on the surge current waveforms given in Table II. It
For a current waveform containing only a half sine wave of
amplitude I1, the inequation presented in (10) must be satisfied TABLE II
so that this waveform can be safely applied without incurring AMPLITUDE (KA) OF EACH 60 HZ CURRENT HALF-CYCLE OF THE SURGE
WAVEFORM
in thyristor catastrophic failure.
Surge # Semicycle
I1a ∙ 10 ≤ c (10) 1 2 3 4 5
In other words, any half sine wave with amplitude not 13- - - - 1
greater than IT(OV) is acceptable. 15- - - - 2
B.Two or More Half Sine Waves 4.0 6.0 - - - 3
5.7 8.7 - - - 4
For a current waveform containing two half sine waves of 7.0 7.0 7.0 - - 5
amplitudes I1 and I2, (11) must be satisfied. 7.0 7.5 9.0 - - 6
I1a ∙ 20 + I2a ∙ 10 ≤ c (11) 7
7.3 7.3 7.3 7.3 -
For N half sine waves of amplitudes I1, I2, I3, …, IN: 7.3 7.5 7.8 8.0 - 8
(1 N−1)
a a a a a
I + I2 + I3 + … + I ∙ 20 + I N ∙ 10 ≤ c . (12)
6.7 6.7 6.7 6.0 6.7 9
Practical examples will demonstrate the effectiveness of this includes single half-cycle surges and multiple half-cycle
method. surges.
For the single half-cycle surges, #1 and #2, the following
IV. AN ALTERNATIVE METHOD FOR VIRTUAL JUNCTION results are obtained by using (10):
TEMPERATURE CALCULATION 136,693 ∙ 10 = 2.85 ∙ 108 > 107.643 = 4.39 ∙ 107 (15)
Despite being the reference method for virtual junction 156,693 ∙ 10 = 7.44 ∙ 108 > 107.643 = 4.39 ∙ 107 . (16)
temperature calculation, the integral of (3) is not easily Then a thyristor SKT-553/18E submitted to surge #1 or
applicable to discrete-time simulation results or experimental
data recorded. Discrete-time recursive convolution may be 100%
used in this case [34]. If the device voltage cannot be
measured, it should be calculated by using (5).
For a current signal i(t) through a thyristor which transient
thermal impedance is a sum of exponential terms as in (1), the 68%
discrete-time recursive convolution can be applied along with 3
the superposition theorem and the trapezoidal integration rule 4
so that Tvj can be calculated as [34]: 35% 5
∑{ }
n
T (t) = T +
vj c p ∙ [T (t − ∆ t) − T ] + q ∙ i(t) ∙ v[i(t), T (t)] + q ∙ i(t − ∆ t) ∙ v[i(t − ∆ t), T (t − ∆ t)] (13)
i vj C i vj i vj 6
i=1

where v[i(t),Tvj(t)] ≈ v[i(t),Tvj(t-Δt)] can be assumed if the time 7

8
step Δt is made sufficiently small and 2% 9
Rth,i 10
ki = Tth,i
1 -30%
ai = Tth,i 0 23 45 68 90
. (14) t [ms]
2 − ai ∙ ∆ t
pi = 2 + ai ∙ ∆ t
Fig. 2. Theoretical percentage degradation for part number SKT-553/18E
submitted to multi half-cycle surges of Table II.
ki ∙ ∆ t
qi = 2 + ai ∙ ∆ t
surge #2 is not expected to survive. For multi half-cycle
surges, the direct application of (12) results in the curves
A comparison between (2) and (13) demonstrates that in the shown in Fig. 2. The vertical axis is presented as a percentage
first case all the past terms must be accounted for the of 10b = 4.39·107.
calculation at any instant, whereas in the second one only the For surges #3 and #4, no catastrophic failure is expected
immediately previous term is needed. This reduces the since the numeric result is greater than 0 at the end of the
computational cost without any impact in the result. A interval. For surge #5, the result is well greater than 0, which
comparison between these strategies will be given ahead in indicates the ride through surge survival condition of this
this paper. thyristor for the corresponding waveform. This result is in
accordance with the SKT-553/18E data sheet: for 0·VRRM and
50 ms, the curve leads to a value around 0.77·9 kA·1.1 ≈ 7.6
kA (remember: the factor 1.1 is adopted for 60 Hz half-sine
waves). It means that this thyristor survives to a current surge
 5

of three consecutive 60 Hz half-sine waves of 7.6 kA. Since 2,0

the amplitude of surge #5 is 7.0 kA, then it can be anticipated
that the thyristor will not be damaged. For surge #6, though, it
is predicted that the device would not survive. The reader can 1,7
perform similar analysis for surges #7 and #8 as well as for #9

V [V]
and #10. 1,3 dts25
B.Virtual Junction Temperature Calculation fit25
The results produced by the application of the discrete-time dts125
1,0 fit125
recursive convolution method will be compared to the ones
given by the equivalent power pulse method. Surges #3 and #4
were adopted as the waveforms to be simulated. The transient 0,6
thermal impedance and the simultaneous current- and 0 500 1000 1500 2000
temperature-dependent voltage drop curves must be modeled I [A]
according respectively to (1) and (5) for the specific thyristor Fig. 4. Voltage drop of thyristor SKT-553/18E at 25 ºC and 125 ºC: points
directly read from data sheet (dts25 and dts125) and points obtained by using
TABLE III
NUMERIC COEFFICIENTS FOR THE TRANSIENT THERMAL IMPEDANCE CURVE
OF THYRISTOR SKT-553/18E 6000
i v v

i(t)[A], v(t)[mV]
Rth,i [mΩ] Tth,i [ms] akdtrc akepp

1 2 3 4 1 2 3 4 4000
1.75 5.07 28.0 110 7.08·10-1 1.18·101 3.78·102 2.36·103
2000
0,05
0
-0.015 -0.01 -0.005 0 0.005 0.01 0.015
100
p p T T
p(t)[kW], T vj(t)[ºC]
dtrc epp vjdtrc vjepp
0,0375 dts
1
Zth(j-c) [°C/W]

2 50
3
0,025
4
fit
0
-0.015 -0.01 -0.005 0 0.005 0.01 0.015
0,0125 t[s]
Fig. 5. Voltage drop, power and virtual junction temperature calculated for a
thyristor SKT-553/18E submitted to multi-cycle surge #3, using discrete-time
0 recursive convolution (dtrc) and equivalent power pulse (epp) methods.
0 0 0,1 1 10 100
power pulse calculation, the simulation was performed
t [s]
Fig. 3. Transient thermal impedance curve of thyristor SKT-553/18E plotted 10000
from points read from datasheet (dts) and from (1) using coefficients shown i v v
i(t)[A], v(t)[mV]

akdtrc akepp
in Table III.
part number adopted in this work. The coefficients for both 5000
curves were calculated using “cftool” from MATLAB. Table
III and Fig. 3 present the result for transient thermal
impedance, while Table IV and Fig. 4 for the voltage drop.
0
Fig. 5 presents the comparison between calculated voltage -0.005 0 0.005 0.01 0.015 0.02 0.025
150
drop, power, and virtual junction temperature for surge #3. p p T T
p(t)[kW], T vj(t)[ºC]

dtrc epp vjdtrc vjepp

The sample time is 25 µs. Discrete-time recursive convolution
was calculated using this time interval. For the equivalent 100

TABLE IV 50
COEFFICIENTS FOR CURRENT- AND TEMPERATURE-DEPENDENT VOLTAGE
DROP CALCULATION OF THYRISTOR SKT-553/18E
0
Coefficient Value -0.005 0 0.005 0.01 0.015 0.02 0.025
A0 0.96 t[s]
AT -2.21·10-3 Fig. 6. Voltage drop, power and virtual junction temperature calculated for a
B01.08·10-4 thyristor SKT-553/18E submitted to multi-cycle surge #4, using dtrc and epp
BT -7.99·10-5 methods.
C01.38·10-2 considering rectangular power pulses with duration of 1 ms.
CT 4.95·10-5 The initial virtual junction temperature is Tc = 25 ºC. As one
k -9.35
 6

can notice, the difference between the results is practically source gate-drivers have been employed for firing the
negligible. thyristors. Fig. 8 shows the complete test bench, including
Fig. 6 shows the results obtained for surge #4. Again, no control and measurement apparatus.
significant difference can be noticed between the results. Switches S1, S2 and T1/T2 are initially closed to allow
transformer magnetization. Then the secondary voltage is
VI. EXPERIMENTAL RESULTS measured, and a phase-locked loop algorithm is applied to
choose the point of wave at which T3/T4 are fired so that the
T1 test starts. A preset time interval is waited for to open switch
T1/T2. These parameters were calculated to reproduce surges
S1 S2 #1, #2, #3 and #4. Each surge was applied to a brand-new
T3
P S
switch.

T2 T4 A.Surge #3
Fig. 9 shows the registered waveforms for secondary current
Fig. 7. Simplified schematic of the experimental setup. The device under test
is composed by T3 and T4.
(Ch3), current through thyristor T4 (Ch4) and current through
thyristor T3 (Ch2), while primary voltage (Ch4) and secondary
Fig. 7 shows a schematic diagram of the experimental setup voltage (Ch3) waveforms are shown in Fig. 10. Thyristor T4 is
that was employed in this work. A multi-tap the one that was submitted to a current waveform with the
660/600/550/480/440:15 transformer was utilized to step-up characteristics of surge #3. Fig. 11 shows a comparison
secondary current to the levels compatible with the surges
previously defined in Table II.
The switch S1 is an electromechanical CB and the switch S2
is an electromechanical contactor. In addition to these
components, an electronic switch composed by the thyristors
T1 and T2 connects the power supply to the transformer
primary, which is the high-voltage, low-current side. The
device under test is composed by switches T3 and T4 and is

Fig. 9. Secondary current (Ch3 - yellow), T4 current (Ch4 - green), and T3

current (Ch2 - magenta) waveforms registered during surge #3.

T3/T4

T1/T2
Fig. 10. Primary voltage (Ch4 - yellow) and secondary voltage (Ch3 - green)
XFMR waveforms registered during surge #3. Ch2 (red) and Ch1 (blue) are the
measured gate-to-cathode voltages of the thyristors.

S1 between measured and theoretical voltages across T4, as well

Fig. 8. Overview of the built test bench including control and measurement
devices.
connected to the transformer secondary, which is the low-
voltage, high-current side. Isolated, 6 kHz pulse voltage-
as between measured and theoretical power and virtual
junction temperature. Theoretical quantities were calculated
both by discrete-time recursive convolution and equivalent
power pulse methods. As one can notice, minimal difference is
verified among these results.
As previously calculated, thyristor T4 did not fail due to
surge #3. The electrical resistance of the device was measured
 7

Fig. 13. Primary (Ch4 - yellow) and secondary (Ch3 - green) voltage – surge
#4. Ch2 and Ch1 are the measured gate-to-cathode voltages of the thyristors.

10000

i(t)[A], v(t)[mV]
Fig. 11. Comparison between measured and theoretical voltage, power, and i ameas vakmeas vakdtrc vakepp
virtual junction temperature, using dtrc and epp techniques, for thyristor T4
5000

TABLE V
ELECTRICAL RESISTANCE OF THYRISTOR T4 MEASURED AT 40 °C, BEFORE 0
AND AFTER THE APPLICATION OF SURGE #3 40
T [°C] Rak [MΩ] Rka [MΩ] pmeas pdtrc pepp

p(t)[kW]
Before After Before After 20
40354 318 341 292
0
200
before and after surge, at 40 ºC. The results are presented in
Tvj(t)[ºC]

Table V. A small difference is observed. Tvj Tvj

dtrc
Tvj
epp
100
B.Surge #4
Fig. 12 and Fig. 13 present the waveforms recorded during 0
surge #4. Fig. 14 shows a comparison between measured and -0.015 -0.01 -0.005 0 0.005 0.01 0.015
t[s]
theoretical quantities calculated for thyristor T4. Again,
Fig. 14. Comparison between measured and theoretical voltage, power, and
calculated values show great concordance with measured virtual junction temperature, using dtrc and epp techniques, for thyristor T4
ones. There is practically no difference between electrical submitted to surge #4.
resistances of the device measured before and after the
application of the current surge. Table VI presents the
TABLE VI
ELECTRICAL RESISTANCE OF THYRISTOR T4 MEASURED AT 60 °C AND 40 °C,
measured values, which were obtained at 60 ºC and at 40 ºC.
BEFORE AND AFTER THE APPLICATION OF SURGE #4
T [°C] Rak [MΩ] Rka [MΩ] C.Surge #1
Before After Before After Fig. 15 shows the registered waveforms for primary voltage
6055.8 53.3 59.3 56.1 (Ch1), secondary current (Ch3) and current through thyristor
40277 280 289 292 T3 (Ch4), while secondary voltage (Ch3) and current through
thyristor T4 (Ch4) are shown in Fig. 16. As one can notice, T4

Fig. 12. Current waveforms recorded during surge #4: secondary (Ch3 - Fig. 15. Primary voltage (Ch1 - blue), secondary current (Ch3 - yellow) and
yellow), T4 (Ch4 - cyan), and T3 (Ch2 - magenta). T3 current (Ch4 - cyan) during surge #1.
 8

104
The actual temperature reaches almost 300 ºC and the
i(t)[A], v(t)[mV] 2 device fails catastrophically. Measured values of electrical
i
ameas resistance of the device before and after the test are presented
v in Table VII. A measurement at 25 ºC was included. It shows
akmeas
1 v
akdtrc that the resistance is highly dependent on the temperature even
vakepp after the catastrophic failure and that it may not be easily
0 detectable at ambient temperature, since the value of the
200 resistance is highly below its original value but at the same
p time is extremely above the typical value verified for a short-
p(t)[kW]

meas
100 p
dtrc circuit condition.
p
epp
D.Surge #2
0
500 Fig. 18 shows the registered waveforms for this test. As it
Tvj(t)[ºC]

Tvj can be seen, the device failed catastrophically immediately

Tvj
dtrc after the 15-kA peak. As in surge #1, the current waveform
Tvj
epp changes its pattern after 15 kA: instead of falling, its value
0 keeps rising even under decreasing voltage. This change has to
-0.01 -0.005 0 0.005 0.01
t[s]
Fig. 19. Voltage, power, and virtual junction temperature – measurement x
theoretical dtrc x theoretical epp for thyristor T3, surge #2.

carries current firstly, while the secondary voltage is positive,

and then T3, when the voltage becomes negative. At some
point in the waveform, a transient is verified where the
resistance of T3 appears change to a lower value (the voltage
decreases while the current increases). When the voltage
becomes positive again, thyristor T3 keeps carrying current.
The peak current through T4 is now extremely below the value
15000
i(t)[A], v(t)[mV]

i
ameas
10000 vakmeas
v
5000 akdtrc
Fig. 18. Primary (Ch1 - blue) and secondary (Ch2 - magenta) voltage,
v
akepp
0
100 TABLE VIII
p ELECTRICAL RESISTANCE OF THYRISTOR T3 MEASURED AT DIFFERENT
p(t)[kW]

meas
TEMPERATURES, BEFORE AND AFTER THE APPLICATION OF SURGE #2
50 p
dtrc T [°C] Rak [MΩ] Rka [MΩ]
pepp
Before After Before After
0 6066.7 - 65.0 -
300
Tvj
40406 - 400 -
Tvj(t)[ºC]

200
Tvj 25- 1.00·10-6 - 1.00·10-6
dtrc
100 Tvj epp
do with the huge difference seen between measured and
0
-0.015 -0.01 -0.005 0 0.005 0.01 0.015 calculated quantities, which are now even more accentuated
t[s] than in case of surge #1, as it can be seen in Fig. 19. Another
Fig. 17. Voltage, power, and virtual junction temperature – measurement x difference is that the resistance of the device is now low even
theoretical dtrc x theoretical epp for thyristor T3, surge #1. after it has cooled down to ambient temperature, as it can be
it has reached at the beginning of the test. T3 failed seen in the data presented by Table VIII.
catastrophically, as it was anticipated. The device has carried Fig. 20 shows a photograph of the thyristor after being
13 kA, but it failed at some point after this. divided into two parts. No visible signal of damage was found.
Fig. 17 shows that there is an expressive difference between
measured and calculated voltage, power, and virtual junction VII. CONCLUSION
temperature. In this paper, a procedure for the calculation of the ride
TABLE VII through condition of a thyristor submitted to single- or multi-
ELECTRICAL RESISTANCE OF THYRISTOR T3 MEASURED AT DIFFERENT cycle current sinusoidal waveform with high amplitude was
TEMPERATURES, BEFORE AND AFTER THE APPLICATION OF SURGE #1
developed and evaluated both from simulation and
T [°C] Rak [MΩ] Rka [MΩ] experimentation. This procedure allows for the prediction,
Before After Before After during design phase, of the possibility of catastrophic failure
6067.2 15.0·10-8 61.9 15.0·10-8 of a given thyristor submitted to any AC sinusoidal current
40327 24.0·10-8 313 24.0·10-8 waveform. Such calculation is based on an equation that has
25- 55.0·10-2 - 66.0·10-2 the form Ia·t, where the current I is the peak value instead of

RMS and the exponent a is not 2, but rather depends on the
parameters given by the manufacturer in the data sheet. The
physical meaning of the constants a and c is being deeply
studied and will be presented in a future paper, as well as a
microscopic evaluation of the device shown in Fig. 20 along
with more experimental work that is being currently
conducted. The work presented in this paper has been
experimentally verified by testing a specific SCR
commercially available device, but it may be useful for any
part number of any thyristor type. It was also shown that the
device resistance is a particularly important failure indicator,
which may be useful even for degradation detection.
REFERENCES
1. A. Q. Huang, "Power Semiconductor Devices for Smart Grid and
Renewable Energy Systems," in Proceedings of the IEEE, vol. 105, no.
11, pp. 2019-2047, Nov. 2017.
2. Dynex Semiconductor Ltd, "Power Semiconductor Product Guide
2019", DYN-PS-19002, 2019.
3. Mitsubishi Electric Corporation. (2020, Nov. 10). Product Information /
Fig. 20. A photograph of the thyristor T3, after becoming damaged by the
current surge #2.
High Power Devices / Thyristors. [Online]. Available: https://
www.mitsubishielectric.com/semiconductors/products/highpower/
thyristors/index.html.
4. J. S. Read and R. F. Dyer, "Power thyristor rating practices," in
Proceedings of the IEEE, vol. 55, no. 8, pp. 1288-1301, Aug. 1967.
5. D. Chamund and C. Rout, "Understanding i2 Phase Control Thyristor
Datasheets", Application Note AN 5950. Dynex Semiconductor Ltd,
Sept. 2009, pp.1–23.
6. H. J. Boenig, J. W. Schwartzenberg, L. J. Willinger, D. E. Piccone, D. A.
Lopez and H. A. Smolleck, "Design and testing of high power,
repetitively pulsed, solid-state closing switches," IAS '97. Conference
Record of the 1997 IEEE Industry Applications Conference Thirty-
Second IAS Annual Meeting, New Orleans, LA, USA, 1997, pp.
1022-1028 vol.2.
7. R. Pastore, M. Weiner, H. Singh, S. Schneider, R. Fox and G. Ayres,
"Evaluation of SCRs as Millisecond Switches for Electric Gun
Launchers," Ninth IEEE International Pulsed Power Conference,
Albuquerque, NM, USA, 1993, pp. 360-.
8. T. F. Podlesak, F. M. Simon and S. Schneider, "Single shot and burst
repetitive operation of thyristors for electric launch applications," in
IEEE Transactions on Magnetics, vol. 37, no. 1, pp. 385-388, Jan. 2001.
9. C. Liu et al., "A Novel Insulated Gate Triggered Thyristor With Schottky
Barrier for Improved Repetitive Pulse Life and High-di/dt
Characteristics," in IEEE Transactions on Electron Devices, vol. 66, no.
2, pp. 1018-1025, Feb. 2019.
10. P. Pawar and A. B. Patil, "Use of diode as a crowbar for electromagnetic
launcher," 2016 ICACDOT, Pune, 2016, pp. 563-567.
9
11. H. Bluhm, "Pulsed Power Systems: Principles and Applications", 1st ed.,
Springer, 2006, pp.1–6.
12. H. Akiyama, T. Sakugawa, T. Namihira, K. Takaki, Y. Minamitani and
N. Shimomura, "Industrial Applications of Pulsed Power Technology,"
in IEEE Transactions on Dielectrics and Electrical Insulation, vol. 14,
no. 5, pp. 1051-1064, October 2007.
13. F. V. Amaral, S. M. Silva, J. A. S. Brito and B. J. C. Filho, "Analysis and
characterization of an active bypass switch for series connected power
conditioners," 2015 IEEE 13th COBEP/SPEC, Fortaleza, 2015, pp. 1-5.
14. F. V. Amaral, M. H. M. Z. Oliveira, C. A. Conceição, S. M. Silva, C. O.
Inácio and B. de J. Cardoso Filho, "On the Application of a Power
Electronics-based Arc-Flash Suppressor," 2019 IEEE 15th COBEP/
SPEC, Santos, Brazil, 2019, pp. 1-6.
15. J. Vitins, J. L. Steiner and J. A. Welleman, "High power semiconductors
for pulsed switching," 7th Pulsed Power Conference, Monterey, CA,
USA, 1989, pp. 352-357.
16. J. A. Kay and L. Kumpulainen, "Maximizing Protection by Minimizing
Arcing Times in Medium-Voltage Systems," in IEEE Transactions on
Industry Applications, vol. 49, no. 4, pp. 1920-1927, July-Aug. 2013.
17. I. L. Somos, D. E. Piccone, L. J. Willinger and W. H. Tobin, "Power
semiconductors-a new method for predicting the on-state characteristic
and temperature rise during multicycle fault currents," in IEEE
Transactions on Industry Applications, vol. 31, no. 6, pp. 1221-1226,
Nov.-Dec. 1995.
18. A. Wintrich, U. Nicolai, W. Tursky and T. Reimann, “Application
Manual: Power Semiconductors”, 2nd revised ed. Semikron International
Gmbh, 2015, pp.146–156.
19. J. W. Motto, "A New Quantity to Describe Power Semiconductor
Subcycle Current Ratings," in IEEE Transactions on Industry and
General Applications, vol. IGA-7, no. 4, pp. 510-517, July 1971.
20. Semikron International Gmbh, "Technical Explanations: Discrete
Thyristors and Diodes”. 2005, pp.1389–1398.
21. Semikron International Gmbh, “SKT 553 Data sheet". 2007, pp.1–4.
22. F. Volle, S. V. Garimella and M. A. Juds, "Thermal Management of a
Soft Starter: Transient Thermal Impedance Model and Performance
Enhancements Using Phase Change Materials," in IEEE Transactions on
Power Electronics, vol. 25, no. 6, pp. 1395-1405, June 2010.
23. ON Semiconductor, "Thyristor Theory and Design Considerations
Handbook", Rev. 1. HDB855/D, 2006, pp.218–233.
24. Infineon Technologies AG, "Transient Thermal Measurements and
Thermal Equivalent Circuit Models", Rev. 1.1, 2018, pp.1–12.
25. B. Allard et al., "Application of averaged models to real-time monitoring
of power converters," 2001 IEEE 32nd Annual Power Electronics
Specialists Conference (IEEE Cat. No. 01CH37230), Vancouver, BC,
2001, pp. 486-491 vol.2.
26. Semikron International Gmbh, "Technical Explanations: Rectifier
Diodes and Thyristors". 2015, pp.1–19.
27. Dynex Semiconductor Ltd, "Effects of Temperature On Thyristor
Performance", Rev. 3.1. AN4870, 2002, pp.1–5.
28. E. T. Schonholzer, "Fuse Protection for Power Thyristors," in IEEE
Transactions on Industry Applications, vol. IA-8, no. 3, pp. 301-309,
May 1972.
29. J. W. Motto, "Thyristor steady state current ratings past, present and
future," Conference Record of the 1993 IEEE Industry Applications
Conference Twenty-Eighth IAS Annual Meeting, Toronto, Ontario,
Canada, 1993, pp. 1256-1263 vol.2.
30. ABB, "Surge Currents for Phase Control Thyristors". Application Note
5SYA 2102-00, 2014, pp.1-5.
31. J. W. Motto, W. H. Karstaedt, J. M. Sherbondy and S. G. Leslie,
"Electro-thermal modeling of multi-megawatt power electronic
applications using PSPICE," Conference Record of 1998 IEEE Industry
Applications Conference. Thirty-Third IAS Annual Meeting (Cat.
No.98CH36242), St. Louis, MO, USA, 1998, pp. 982-990 vol.2.
32. M. Cepek, "Novel techniques and procedures for the assessment of fault
current withstand capability of power thyristors," in IEEE Transactions
on Power Electronics, vol. 14, no. 2, pp. 323-328, March 1999.
33. I. L. Somos, D. E. Piccone, L. J. Willinger and W. H. Tobin, "Power
semiconductors empirical diagrams expressing life as a function of
temperature excursion," in IEEE Transactions on Magnetics, vol. 29, no.
1, pp. 517-522, Jan. 1993.
 10

34. F. Amaral, S. Silva, C. Conceição, G. Guimarães, and B. Cardoso, Pulsed Power Applications," presented at the 55th IEEE Industry
"Catastrophic Failure Prediction for Phase Control Thyristors in AC Applications Society Annual Meeting, Virtual, Oct. 10-16, 2020.

View publication stats