4

Logic Inputs in Digital Protective Relays

4.1 Reliability of Logic Inputs in DPRs

Problems with the reliability of microprocessor-based protective devices
(MPDs) have arisen in connection with the worldwide moving away from
electromechanical and static protective relays to digital protective relays
(DPRs). As shown in References 1 and 2, the common misconception that
DPRs have reliability considerably exceeding that of electromechanical
and static protective relays is, in actual fact, no more than a myth fostered
through the years by the advertising publications of DPR manufacturers. It
is abundantly clear that such complex multipurpose structures as DPRs can-
not, even theoretically, be free of problems, cannot be absolutely reliable, and
cannot be devoid of failure statistics during 15–20 years of maintenance.
Previously, we have discussed the problems connected with insufficient
reliability of the output miniature electromagnetic relays of DPRs intended
for the direct tripping of high-voltage circuit breakers.3 This chapter addresses
a problem with the reliability of logic inputs in the DPR using the RE*316
series (REL, RET, REC, etc.) as an example. The RE*316 series has been widely
used over the past 10–15 years.
Digital (logic) inputs in the DPRs of these types consist of a set of com-
pletely identical elements functionally representing the logical function
Prohibition (see Figure 4.1). The direct input and prohibited input of each of
these elements are connected respectively to the input circuit of the DPR and
to the central processing unit (CPU) through optocouplers: Opt1 and Opt2.
The starting signal from optocoupler Opt2 logically repeats the presence or
absence of an input voltage. The functioning of this circuit can be blocked
(prohibited) through a CPU internal logic. Thus, the blocking signal from the
CPU through matching electronic circuits acts on the prohibiting input (an
input of optocoupler Opt1).
The problem of this type of logic cell consists in the high level of an input
signal (220–250 V DC) which needs to be attenuated up to a level of 1.5–2.0 V,
at which point the Opt2 optocoupler operates. The total current consumed
by the input circuit is split approximately 50/50 between resistors R1 and R2;
therefore, both of them, basically, should be equal. However, there is no place

149
 150
Opt1 Logic input
Opt2 +
VD3 230 V
R1 R2
C
VT Opt2

Digital Protective Relays: Problems and Solutions

“Prohibition”
input VD1
R2 RU
VT
R1
R4 C
Opt1
Z
R3
RU VD2
–

Figure 4.1
Logical inputs: printed circuit board (PCB) module of microprocessor-based protective device (MPD) RE*316 type, single-channel and IT circuit
diagram.
Logic Inputs in Digital Protective Relays 151

on a printed circuit board (PCB) for two large resistors such as R1 and R2
(including the distance between them, which is necessary for cooling).
Therefore, to dissipate the superfluous power of this element, the DPR
designers have assigned to a transistor, VT, working not in a switching mode,
as is usual in such a circuit, but in an amplification mode (linear mode). It
is natural that the resistor for R2 is therefore selected as being of low power
and small size. In practice, such aspirations to miniaturization lead to serious
problems: resistor R2 often completely burns out, leading sometimes to the
burnout of sections of the PCB and even those components in its vicinity.
The first problem of this arrangement centers on the transistor, VT. In
operating in amplification mode and continually dissipating the super-
fluous power, it heats up to temperatures of 70–80°C. Unlike the usual
resistor, the resistance of which slightly increases with increase of temper-
ature, resistance of the n-p-n transition of the transistor with the increase
of temperature essentially decreases owing to the drift of the operating
point (referred to as the quiescent point) on the characteristic and increases
in the coefficient of amplification. This leads to an increase of the collector
current, that is, the current through resistor R2. With the simultaneous
heating of many resistor (R1) and transistor (VT) combinations in 10–15
input circuits, the temperature inside of the input section (in the case that
the DPR is divided into sections by internal diaphragms) can essentially
increase. This leads to a further drift of a quiescent point of the transis-
tor, further increasing its collector current and the overheating of resistor
R1. Thus diode VD, intended for stabilizing the quiescent point of the
transistor, appears ineffective because its temperature and the tempera-
ture of the transistor discriminate on 50–60 grades. Thus, the aspiration
of the manufacturer to reduce MPD sizes, using a transistor instead of a
powerful resistor for superfluous power dissipation, has led to a decrease
of MPD reliability. The problem of insufficient reliability of this circuit
connected with aspirations for miniaturization is not restricted only to
this problem.
The second problem is the miniature disk ceramic capacitor, C, with a
capacitance of 2.2  nF. Ceramic capacitors are considered to be among the
most resistant to the effects of electric operational loads and are stable over
long durations. However, in ceramic capacitors with badly isolated gaps
between electrodes, there is the possibility of a decrease in the insulation
resistance and even an electric breakdown due to the migration of metal ions
of facings (especially silver) on the butt end of the capacitor, especially in
conditions of a wet tropical climate. At an input voltage of the DPR below 100
V, capacitors such as these give no indication of the defects and at procall, a
usual tester, show full serviceability. But already at voltages of 180–230 V, an
outflow current through the capacitor increases (Figure 4.2) such that normal
operation of the transistor and optocoupler becomes impossible. Moreover,
there is a long-range influence of an input voltage of 220–230 V because of the
152 Digital Protective Relays: Problems and Solutions

Resistance of capacitor C
in dependance to applied
R, kOhm DC voltage
400

300

200

100

0
0 100 200 U, V

Figure 4.2
Reason for R2 burning is increasing leakage current through capacitor C.

increased outflow current through this capacitor; the power dissipation on

resistor R2 increases, and it simply burns out (Figure 4.2).
The third problem is the internal cutoff failures appearing in the course of
time in the powerful R1 resistors (see Figure 4.1). This phenomenon presumably
originates owing to a contact fault between metal cups at end faces of the resistor
and a resistive layer on a surface of the ceramic cylinder because of its shelling in
cheap resistors used in the DPR. The fourth problem is mass failures of the Opt2
optocouplers. For example, once we had to replace seven optocoupler units of
type CNY17-2 on only a single PCB of a logic inputs unit of relay REC316!
Optocouplers of this type have following parameters:

• Average forward current (a current through a light-emitting element,

or IF): 60 mA
• Maximum forward current (IF): 100 mA
• Rated collector current of the output transistor: 50 mA
• The maximum collector current: 150 mA

It appears that these optocouplers operate in the DPR in a mode that is far from
nominal; rather, they operate with very small input (forward) and output (col-
lector) currents, practically on the bottom boundary line of these currents.
Besides, as a result of research carried out by us, it has been found that
optocouplers, desoldered from nonworking inputs of the DPR, actually are
Logic Inputs in Digital Protective Relays 153

TABLE 4.1
Voltage Drop (V) on Output Transistors (Collector-Emitter)
of CNY17-2 Optocouplers at Collector Current 0.5 mA and
Input (Forward) Currents 2 and 2.5 mA
IF = 2.0 mA IF = 2.5 mA
Optocouplers: Extracted from Nonworking Inputs of REC316
1.471 0.992
1.027 0.344
0.574 0.188
1.452 0.967
1.315 0.769
0.734 0.201
0.634 0.178

New Optocouplers: Same Types, but Other Manufacturer

0.134 0.120
0.143 0.127
0.132 0.119
0.144 0.127
0.139 0.123
0.139 0.124
0.134 0.120
0.143 0.125

not damaged. The difference between “working” and “nonworking” in DPR

optocouplers has appeared only in their sensitivity to very small, lowest-
limit forward currents (IF); see Table 4.1.
Judging by the data in Table 4.1, the collector-emitter voltage drops of opto-
couplers from nonoperating input circuits of the DPR considerably exceed
typical values (see Figure 4.3; 0.18 V at forward current IIF = 2 mA and col-
lector current 0.5 mA), while the same values for new optocouplers of other
manufacturers are below the typical value.
Unfortunately, many optocoupler manufacturers (e.g., CNY17-2 type
optocouplers are manufactured by the companies Agilent, Toshiba, QT
Optoelectronics, Fairchild, Vishay, Liteon, Everlight, Isocom, and Opto Inc.,
among others) simply do not present these major characteristics of photocou-
plers in their datasheets; therefore, any claims regarding them are impossible
to present. The claim can be made to designers of the electronic circuits (in
our case, to the designers of RE*316 from company ABB) that the operation
mode of the equipment (the DPR) is on the boundary line of the char­acteristic
of electronic components (optocouplers); therefore, the usual slightest tech-
nological dispersions in parameters of these electronic components lead to
full loss of the working capability of such crucial equipment as DPRs.
The fifth problem is the manufacturer selecting too wide a range of operat-
ing voltages (82–312 V) for logic inputs. This wide range of operating voltages
154 Digital Protective Relays: Problems and Solutions

CNY17-2 TA = 25°C
Agilent
5
IC = 0.5 mA

4 IC = 1 mA
VCE(SAT)

3
IC = 2 mA

2 IC = 3 mA

IC = 5 mA
1

0.18
0 2.0 5.0 7.5 10.0
IFmA

Figure 4.3
Voltage drop on the output transistor VCE(SAT) in dependence from the input current of the opto-
coupler (IF) for different collector current values.

is the source of two problems: low efficiency of overvoltage protection by

means of the varistor and a problem of false operations.
The first problem is caused because by maintaining the working capability
at a voltage of 312 V (which, needless to say, is absolutely not realistic in the
practical condition with a DC network based on 220 V batteries), the clamp-
ing voltage of the varistor RU (7K391 type, in our case; see Figure 4.1) should
be selected as not less than 650 V (because of the actual characteristics of
varistors for a working voltage of 312 V DC). This means that all electronic
circuit components will be exposed to the overvoltage reaching up to 650 V
(for comparison, the maximum collector-emitter voltage allowed for transis-
tor VT 2N3439 type does not exceed 350 V).
The second of these two problems is caused by too low a value of the bot-
tom boundary line of an operating voltage range: 82 V. When one of the poles
(positive or negative) in the DC power network, in which the DPR is con-
nected, is grounded to earth, there is a voltage pulse equivalent to half the
value of the rated battery voltage (that is, nearly 110 V) caused by the dis-
charge of the capacity of wires of a DC network that is produced. As shown
in Reference 4, at such conditions the spontaneous activation of logic inputs
of DPRs occurs if an operating threshold of logic inputs is below half of the
mains voltage of a DC network (that is, 110 V).
In just a fragment of the DPR circuit containing only ten electronic com-
ponents, we have shown how many of the problems connected with DPR
Logic Inputs in Digital Protective Relays 155

reliability can originate in actual conditions of operation. It confirms the the-

sis put forward by us that modern DPRs are not the model of perfection and
reliability at all as advertising catalogs try to present them. Further, they
represent a most complicated and not always reliable enough method to be
employed. These problems should not be ignored as is done today; rather,
they should be widely discussed.

4.2 Increasing Noise Immunity of the Logical Inputs in DPRs

Earlier, I offered solutions for solving the problem regarding the insuf-
ficient noise immunity of microprocessor-based protective relays.2,5 For
example, in 1996 the way of increasing the noise immunity for analog
inputs of the microprocessor based on overcurrent protection using fast
(0.8–1.5 ms) reed relays with high-voltage (3–5 kV) insulation was offered.6
Fast relays such as these should short-circuit while constantly measuring
inputs of the microprocessor relay until the input current reaches some
threshold value. Besides analog (measuring) inputs, microprocessor-based
protective relays are supplied with many logical inputs. Depending on the
internal protection logic incorporated in the relay, in some cases activat-
ing some of these inputs may cause a false trip of the high-voltage circuit
breaker, a high-voltage line, the transformer, or other important power
equipment.
In the event that such an input activation is a consequence of interference
affecting the relay, false disconnection of high-power equipment is a serious
crash, accompanied by significant damage and requiring time-consuming
analysis of the reasons for its cause. Unfortunately, in many cases it is not
possible to reveal the true reason for a false operation of protective relay.
Investigation of several such situations in which I participated has shown
that the frequent reason for false operations of microprocessor protection
on power substations is grounding of wires (poles) of a DC network. Thus,
nothing terrible happens with the network, as both its poles are in normal
conditions and well isolated from earth, and the condition of the insula-
tion is constantly monitored by special devices. However, this insulation
causes significant capacity branching and long DC networks, especially in
large substations, and is the origin of the pulses of a discharge current (and,
accordingly, voltage) at the grounding of one of the wires (poles) of a network
(see Figure 4.4). The maximal value of the voltage originating at the capac-
ity of a discharging network can reach, basically, up to half of the value of
the substation battery voltage, that is, 110 V. Actuating the logical inputs
of the microprocessor relays of various types occurs at various levels of the
initial voltage, since 50–60 V is the typical voltage for single-pole grounding
156 Digital Protective Relays: Problems and Solutions

1 I 2 I I
– +

Figure 4.4
The circuit diagram for discharging a DC network capacity through logic input of the protec-
tive relay (2) with the open control contact (1) at single-pole grounding.

in a DC network. Moreover, power and pulse duration of a discharge current

appear quite sufficient not only for actuating the logical inputs of micropro-
cessor relays, but also even for energizing some types of auxiliary electrome-
chanical relays with a rated voltage of 220 V and a minimal pickup voltage
near 70–80 V.
Considering the limited pulse duration of a discharge current, technically
it is possible to inject (through computer software) the additional program
timer, providing a time delay of the input signal on 20–40 ms into a logical
circuit of the microprocessor relay. A timer such as this will play the role of
the filter, prohibiting pickups of the microprocessor relay of the troublesome
short-time pulses of a current at the discharging of a DC network capacity.
It is possible to use a resistor-capacitor (RC) circuit which is connected to the
inputs of the relay and carries out the same function.
However, it is necessary to make the decision of, basically, whether or not
doubling the time reaction of the protective relay to a fault mode (short cir-
cuit) is admissible. In our opinion, such delay of the protective relay actua-
tion is inadmissible. In connection with what has been stated above, another
solution for the problem is offered, that is, increasing the minimal threshold
of logical input pickups to a level exceeding half the value of the DC network
voltage. The selected level of the voltage is to be 150 V.
To implement this idea, we suggest using an elementary module as shown
in Figure 4.5, consisting of two electronic components: a Zener diode with a
rated voltage of 150 V and rated current of 5 mA, and a power thyristor for
a maximum current of 7.5 A, maximum voltage of 800 V, and required gate
current of 5 mA. At voltage levels below 150 V, the device is in the inactive
condition; therefore, voltage pulses with magnitudes of up to 110 V, originat-
ing in a DC network ground, do not pass to logical inputs of the relay.
On the turning ON of the operating contact S, a voltage of 220 V will be
abruptly applied to the Zener diode and it will instantly open, passing a
current in the gate of the thyristor. The thyristor will open, and by the low
forward resistance shunts the Zener diode and the gate circuit.
Now all operating current passes to logical input of the relay through the
anode-cathode circuit of the thyristor. The voltage drop on the open thyris-
tor does not exceed a fraction of a volt, which, in combination with the small
current carried through it (15–50 mA), causes very insignificant dissipating
power and does not heat the power thyristor.
Logic Inputs in Digital Protective Relays 157

+250 V
S
Trigger module
1
1 Trigger module
2

Protective
Logical relay
input

3 2

Figure 4.5
The circuit diagram and a design of the trigger module. 1: thyristor BT151-800L type; 2: Zener
diode 1N5383 type; and 3: two terminal panels of the Wieland 9700A/6S35 type.

Both of the elements are placed on a small dielectric plate constructed

with a glass-cloth-base laminate and measuring 70  × 40 mm. The plate is
clamped by means of two screws between two panels of the standard ter-
minal block, intended for installation on a standard Deutsches Institut für
Normung (DIN, the German national organization for standardization) rail.
After the installation of elements on a plate, they are covered by a layer of
water-­resistant varnish and are isolated by means of dielectric shrink tubing,
providing mechanical protection of the elements on the plate. This device
represents a small, easily constructed module (see Figure 4.6) having a very
low cost (the thyristor together with the Zener diode cost less than $2.00).
The module is not connected to all the logical inputs, only to those that are
especially critical. Upon activation it may affect pickups of a microprocessor
relay and tripping high-voltage line or power electric equipment. For a single
microprocessor relay, two to four modules may be required. The proposed
device can also be used together with the usual auxiliary electromechanical
relays when their real pickup voltage is less than half of the DC network volt-
age, that is, when there is a threat of their faulty operation.
It is necessary to emphasize that use of the proposed device does not
reduce the reliability of relay protection, as electronic elements are com-
pletely deener­gized at the opening of the operating contact S, and therefore
cannot pickup spontaneously, for example owing to a breakdown or a spon-
taneous switching ON of the thyristor. Even allowing the probability (theo-
retically only) of a breakdown of the electronic elements of the module, the
protective relay is simply returned to the mode of operation in which it was
before using the module. The module can only have an effect on reliability of
the relay in the breaking of the internal circuit. However, according to statis-
tics, internal breakages in semiconductor devices originate only in the pres-
ence of very high carrying currents (when there is a burnout of the internal
158 Digital Protective Relays: Problems and Solutions

Figure 4.6
External view of the single module mounted on a standard Deutsches Institut für Normung
(DIN) rail.

semiconductor structure) and thus constitute no more than 5% of all faults.

Under normal conditions of real currents carried through the thyristor in
a real circuit of the module (15–50 mA), burnout of the internal thyristor
structure is not possible. Some internal soldering joints and additional wire
splices by means of the terminal block used in the module can really affect
the reliability of relay protection. Therefore, the ensemble of these elements
(in which latent defects are possible) is not used on the PCB, and leads of the
thyristor and Zener diode are fixed before soldering by means of a thin cop-
per tin-coated conductor that keeps the probability of failure of the module
extremely low.

References
1. Gurevich, V. I. How to Equip a Relay Protection: Opinions of Russian Experts
and a View from Outside. Electric Power News, no. 2, 2007, pp. 52–59.
2. Gurevich, V. Reliability of Microprocessor-Based Relay Protection Devices:
Myths and Reality. Engineer IT, part 1: no. 5, 2008, pp. 55–59; part 2: no. 7, 2008,
pp. 56–60.
3. Gurevich, V. Peculiarities of the Relays Intended for Operating Trip Coils of the
High-Voltage Circuit Breakers. Serbian Journal of Electrical Engineering, vol. 4, no.
2, 2007, pp. 223–237.
4. Gurevich, V. I. Increasing Noise Immunity of the Logical Inputs in Microproces­
sor Based Protective Relays. Electronics-Info, no. 11, 2008, pp. 26–27.
Logic Inputs in Digital Protective Relays 159

5. Gurevich, V. I. Principles of Increase of Noise Immunity of Static Overcurrent

Protective Relays. Power Engineering and Electrification, no. 2, 1992, pp.16–18.
6. Gurevich, V. I. Some Ways for Solving Problem of Electromagnetic Compatibility
of Protective Relaying. Industrial Power Engineering, no. 3, 1996, pp. 25–27.