Mullard ThyratronsForIndustrialControl 1954
Mullard ThyratronsForIndustrialControl 1954
Circuit Specifications
Comprehensive Control System for a DC Motor (Fig. 24, page 23)
Rl = 47k52 1W * Pl 20kS2 SW
R2 470k52 0•SW P2 lOkSZ 1W
R~ 470k52 0•SW C1 2.0µF
Q similar circuit was published by A. L. Whiteley in "The Engineer", Vol. 186, 1948, p. 219, under the
title "Electronic Control of Motors."
Single-Cycle Timer for a Small Spot Welder (Fig. 25, page 24)
F~--S= 5 — 68
APPLICATIONS
Motor Control 20
Resistance Welding Control 23
Regulation of D.C. Power Supplies 25
Regulation of A.C. Power Supplies 25
Thyratron Inverters 26
Timing Circuits 27
Lighting Control 28
Electric Furnace Control 29
Control of Vibrators 29
Firing of Mercury Pool Rectifiers 29
Thyratrons as Uncontrolled Rectifiers 29
BIBLIOGRAPHY 32
Heating Time
The time required for a valve to attain minimum operating temperature
with normal voltage applied to the heating element. For a mercury vapour
valve this time is generally much longer than that required to bring the cathode
to the normal operating temperature.
Anode Voltage Drop or Arc Voltage
The potential difference between anode and cathode during conduction.
Control Characteristic
A relation, usually depicted graphically, between the critical grid voltage
.and the anode voltage.
Control Ratio
The slope of the straight part of the control characteristic.
Critical Grid Voltage
The instantaneous value of grid voltage at which anode current commences
to flow.
Critical Grid Current
The instantaneous value of grid current immediately before anode current
commences to flow.
Commutation Factor
The product of the rate of anode current decay (A/µsec) and the rate of the
initial inverse anode voltage rise (V/µsec) immediately following the current
decay. These values should, in practice, be taken as the average rate of
current decay over the last l0 µsecs of current conduction, and the average
rate of inverse voltage rise over the first 200V.
Deionisation Time
The time between the cessation of anode current and the instant when the
grid regains control.
Ionisation Time
The time required under specified conditions for the anode current to rise to
90 per cent of its rated peak value, the time being measured from the instant of
application of critical grid voltage.
Maximum Averaging Time
The longest period of time over which it is permissible to compute the
maximum average value of the characteristic under consideration.
Condensed Mercury Temperature
The temperature of the external surface of that part of the valve envelope
at which the mercury is seen to condense during normal operation of the valve.
6
GENERAL DESCRIPTION AND
OPERATING PRINCIPLES
7
the critical grid voltage being positive with respect to thus to reduce the extent to which the voltage of the
the cathode. first grid will follow any sudden change in the anode
Industrial thyratrons for control applications voltage. This obviously reduces the risk of random
usually have a negative control characteristic. Such striking. Often the second grid serves also as a thermal
thyratrons require a negative grid bias when it is screen for the first grid. If the temperature of the first
desired that the tube shall be non-conducting. This grid is allowed to rise unduly it will emit electrons
disadvantage is far outweighed by the fact that the which will result in a spurious signal across the grid
valves do not, ideally, pass any grid current before resistor, with a possible loss of control.
conduction. The grid control voltage can therefore Many thyratrons incorporate a shield grid which
be provided by a high impedance source such as a is connected internally to the cathode. The valve is
voltage amplifying pentode of high mutual con- then nominally a triode. When, however, the shield
ductance. Only thyratrons with negative control grid connection is brought out to a separate base pin,
characteristics are considered in this publication. the control characteristics can be modified by applying
From curve A in Fig. 2 it will be seen that when the different potentials to this grid. A family of control
control grid is at cathode potential, the thyratron will
strike (i.e. become conductive) at an anode voltage
Vgt. This voltage is an important characteristic since
it represents the minimum anode voltage at which the
thyratron retains its negative control characteristic.
If it is desired to operate a thyratron at an anode
voltage below Vst, conduction must be initiated by
producing aloes-current arc between the control grid
and the cathode. With these conditions, often termed
positive firing, it is necessary to feed the control grid
from a source of lower impedance.
—V9 O +Vg
For any particular type of thyratron there will
FIG. 3
inevitably be small variations in critical grid charac-
teristics, both from valve to valve and during the life characteristics, with the shield grid voltage as para-
of individual valves. These are due to dimensional meter, is given in Fig. 3. They show that by con-
variations in the electrode system, and to changes in necting the shield grid directly to the cathode, normal
valve temperature and cathode emission during triode operation is obtained. If other potentials are
operation. To indicate the extent of these variations, applied to the shield grid, the control grid charac-
valve manufacturers usually quote a spread of teristic is changed. In fact the thyratron can be given
critical grid voltage for each type of thyratron. The a positive control characteristic if desired. An
circuit designer must ensure that the particular grid alternative mode of operation is to use the shield
circuit can accommodate any critical grid voltage grid as a normal control electrode.
within this band. It follows that, in general, the grid.
Grid Control
signal voltage must be large in comparison with th:;
The normal method of striking a thyratron is to
critical grid voltage. Since, however, the gre-a.L
reduce the negative grid bias to the critical grid
majority of thyratrons have similar grid characteris-
voltage value. If a d.c. voltage Va is applied to the
tics, standard control circuits can be designed for use
anode, the critical grid voltage Vgccrit~ can be repre-
with different types of thyratron.
sented by a straight line as indicated in Fig. 4. If the
Triode and Tetrode Thyratrons grid voltage is raised at the instant t from its hold-off
Although the majority of thyratron applications bias value Vg to a value above Vg~crit>, the valve will
require only one control electrode, many thyratrons strike. The anode voltage will then fall to the arc
are provided with a second electrode or shield grid voltage, Varc, and an output voltage, Va—Varc, will
which partially screens the control grid. The most be available. To regain control the grid voltage must
important function of the second grid is to reduce the be made more negative than the critical value and the
capacitance between the first grid and the anode, and anode voltage must be reduced below the arc voltage.
8
the valve strikes, the duration of this delay depending
Va Va
on the instant at which the grid voltage was increased
and the value to which it was raised. At the end of the
positive half-cycle of Va the anode voltage will be less
than the arc voltage and the valve will automatically
Varc
0 t, extinguish. With a single valve there is no conduction
TIME
I
I
during negative half-cycles of the supply.
---Vg (crit~ In many control applications thyratrons are fed
Vg
from an alternating supply but are arranged to give
Vg
Ibb9 a variable output voltage (either direct or alternating)
FIG. 4 by altering the time during which the thyratrons
conduct in each half-cycle. This can be achieved by
When the anode of a thyratron is fed with a varying the instant at which each thyratron strikes.
sinusoidal alternating voltage, the critical grid voltage Fig. 7 shows the basic connection for a bi-phase
Vgccrit~ will vary with the instantaneous value of Va.
Vo
— Vg ~-- ~ TIME
~ /~ --vy
z_ ~,. FIG. 7
1670 ~g
FIG. 5 rectifier employing two thyratrons. At (a) the valves
are made to strike early in each half-cycle and a large
unidirectional output is obtained. At (b) striking is
This variation can be plotted from the control charac-
delayed and the output is smaller. The arrangement
teristic as shown in Fig. 5. Tf now a steady negative
in Fig. 8 is similar, but the valves are connected
direct bias Vg is applied to the grid, the valve can be
back-to-back to give an alternating output.
struck by neutralising the grid bias,as indicated in Fig. 6.
In-this case, however, a time tl —t must elapse before
1673
a ~1671
Joao
Vg
- Vg
FIG. 6 FIG. 8
9
As already explained, the control characteristics bias is equal to the peak-to-peak value of the alter-
of a particular type of thyratron vary both from nating grid voltage, control can be exercised over the
valve to valve and also for any one valve during its full 180° of the positive half-cycle of Va. In the
life. Allowances for this must be made in the design horizontal shift method of control illustrated in Fig. 11,
of the grid circuit. if consistent striking is to be the direct grid bias is maintained constant, the striking
achieved, the striking signal must intercept the critical
grid characteristic at a steep angle to the time scale.
FIG. 12
FIG. 10
For a vertical shift system where a constant phase
shift of 90~ is required, the value of R is constant and
1
In Fig. 10 the alternating grid voltage has a perma- equal to The variable direct bias is applied
nent phase shift with respect to the alternating anode cuC
voltage—usually 90°—and is in series with an adjust- between the point X and the cathode of the thyratron,
able direct bias. By changing the direct bias the point Y being connected to the grid.
point at which the grid voltage curve intercepts the For manually-operated horizontal shift control
critical grid characteristic is changed, thus altering the resistor R would be a potentiometer. In a servo
the striking angle. This is sometimes called vertical system R could be simulated by ahigh-vacuum valve.
shift control. If the range of adjustment of the direct A common variant of this arrangement is the use of a
10
Iixed value of R, and the replacement of C by an Vt
inductor whose incremental permeability is controlled
by a winding carrying a direct current.
Both the vertical shift and the horizontal shift
methods of grid control are widely used, but they
have the disadvantage that if a very stiff control is
required, and the unit does not incorporate feedback,
the alternating grid voltage must be large. Abetter
solution is to feed the grid from a pulse source in
1 C
which the phase of the pulses is variable. If the
pulses have a steep leading edge a very stiff control can
be obtained with pulses of quite small amplitude, as OUTPUT
indicated in Fig. 13. Pulse control has the added
advantage that it does not produce positive grid current
when the anode potential is negative (see page 15).
The pulse source most commonly used is a peaking
transformer fed from aphase-shifting network. 1679
FIG. 14
1676
Usually, however, the anode of the thyratron is fed
from an alternating supply and the valve is therefore
automatically extinguished at the end of each positive
half-cycle. The grid may then re-strike the valve if
required. The maximum time delay in the response
of such a system using a bi-phase circuit is the period
of half a cycle.
FIG. 13
11
Mounting of gas-filled types. The latter may, therefore, be
Gas-filled thyratrons may be mounted in any operated at higher frequencies, but they may have to
position between horizontal and vertical with the base be protected from high commutation factors leading
downward, and are therefore suitable for use in mobile to gas clean-up, i.e., the absorption of gas into the
equipment. Mercury thyratrons, on the other hand, electrodes. This subject is dealt with more fully under
may be mounted in a vertical position only. Gas Filling and Clean- Up on page 18. Gas clean-up
is usually more severe at the higher operating fre-
Ionisation and Deionisation Time quencies.
The ionisation and deionisation times of mercury Mercury vapour thyratrons have. a reservoir of
vapour thyratrons are normally longer than those mercury and therefore do not suffer from clean-up.
12
RATINGS AND DESIGN CONSIDERATIONS
Comprehensive data for the range of Mullard characteristic. In these circumstances the grid circuit
thyratrons, including ratings (limiting values) and must allow sufficient positive grid current to flow to
electrical and mechanical characteristics, are given in ensure that the discharge is transferred to the anode.
the final section of this publication. In the present Recommendations for specific applications where this
section the significance of the data, the reasons for mode of operation is necessary may be obtained on
the various operating limits, and other factors application to Mullard Ltd.
affecting circuit design, are considered.
CATHODE CURRENT RATING
CHOICE OF VALVE TYPE It is essential that the maximum peak and average
The particular features of the two general types of current ratings of the thyratron are never exceeded.
thyratron are indicated in the following table: The maximum average current must be calculated for
the published maximum averaging time T. The value
TABLE I obtained must not exceed the published limiting
value. For any shorter period t the following
MERCURY GAS-FILLED relationship must be satisfied
1pk Iay.
Type of equipment Static Static or mobile C t
For continuous sine wave operation Figs. 15 and 16
Approximate 15°C to 50°C —55°C toy-70°C may be used to indicate the ratio between the peak and
ambient tempera- average cathode current. Confirmation by actual
ture range measurement should, however, be made.
Maximum Peak Cathode Current
Operating Low frequen- Up to approxi-
frequency cies only mately 5 kc/s The peak cathode current, which is most con-
veniently measured with an oscilloscope, is limited
Heating-up time Long Short by the peak emission from the cathode and the pressure
of the conducting gas or vapour. Any attempt to
Circuit Simple circuit Circuit must exceed the quoted value will cause the arc voltage to
characteristics design not cause gas rise and will reduce the life of the valve.
clean-up For certain thyratrons a second, lower limit to
thepeakcathodecurrentrating is imposed for operation
ANODE VOLTAGE RATING at frequencies below 25 c/s. This is to prevent damage
to the cathode, which may occur if a high peak current
Maximum peak voltage ratings, both forward and
is drawn for relatively long periods.
inverse, are quoted for each type. They include any
surges which may occur. If these ratings are exceeded, Maximum Average Cathode Current
uncontrolled striking may take place. The average cathode current, as measured by a
In some applications it may be necessary to operate d.c. moving coil meter, determines the valve dissi-
a thyratron at an anode voltage which is lower than pation due to load current as indicated below. If
the maximum strike voltage indicated by the control v and i are instantaneous values of the arc voltage and
]3
Ipc°k Ipc°k
Imcan I r.m. s.
25 6
1681 1680
20 5
IS 4
10 3
5 2
O 30° GO° 90° 120° I50°
FIG. 16
14
MAXIMUM CONTROL GRID CURRENT
One reason for setting maximum permissible values
for the control grid current is to reduce the risk of gas
clean-up and electrode sputtering. This aspect is
dealt with more fully on page 18. In addition, when
positive grid current is drawn with the anode more
positive than —10 volts, it is necessary to limit the
grid current in order to keep the grid dissipation to a
reasonable level. Normally the specified limits can
be met with ease. FIG. 17
CONTROL GRID RESISTOR must be maintained within the —10 volt limit, say at
Both a maximum value and a recommended mini- —8 volts. From the grid ion-current characteristic of
mum value of the control grid resistor are quoted in the XR1-3200, Fig. 18, it is seen that a grid ion-
thyratron data. current of —53mA flows when the average anode
current is 2.4 amperes and the grid voltage is --8 volts.
Minimum Value
Thus a total of —60 —(-8) —52 volts must be
The minimum value for the control grid resistor dropped across Rg when a current of —53 mA flows
is determined by the need to keep the grid current through it. The minimum value for Rg is given by:
below a safe figure. In addition, there are certain — 52
cases, e.g. pulse firing, when the control grid signal 10-3 ' 1 k S2.
Rgcmin>--53 X
tends to hold the grid .circuit negative during con-
N 996 XRI-3200
duction, and it is essential that in these circumstances
the actual grid should be more positive than —10 volts a
(see page 11). The grid voltage is a function of the E
rn
grid resistor. In the grid ion-current graph of the H
15
Maximum Value below the striking voltage a limit to the maximum grid
At first sight it might be considered that a very circuit resistance may be imposed by the need to
large value of grid resistor should be used with a ensure sufficient positive grid current to cause reliable
thyratron in order to prevent any load falling on the initiation of the arc.
source of the grid signal, and also to reduce grid
conduction currents. However, uwing to the effects HEATER SUPPLY
of grid emission, leakage current, and grid-anode The voltage of the heater supply should be kept
capacitance, a limit must be set to the maximum within the rated limits specified in the valve data, to
value. give consistent performance and to ensure good life.
In thyratrons there is usually a slight amount of If the supply voltage is too high, the temperature
emissive material on the control grid. When the of the cathode, and thus the rate of evaporation of
anode is positive, electrons will flow from the control emissive material, will be excessive. Some of this
grid to the anode, and the corresponding grid current material will be deposited upon the grid, and, as the
will produce a voltage drop across the grid resistor, grid temperature will also be high due to heat radia-
making the grid more positive than is desired. This tion from the cathode, excessive grid emission may
voltage drop tends to strike the valve. The effect occur, with its attendant problem of loss of grid
increases with higher values of grid resistor. control.
Similarly, the anode-grid capacitance will cause an If, on the other hand, the heater voltage is too low,
anode-grid current to flow and will lead the anode the cathode temperature will be low and emission
voltage. This current will also generate an unwanted will be reduced. This will result in an increase of the
voltage across the grid resistor. Here again, a low arc voltage, and in cathode sputtering which will
value of grid resistor will reduce this unwanted grid shorten the life of the valve. It is therefore very
signal. important to adjust the heater voltage to the correct
The maximum values of grid resistor quoted for value. When checking this voltage, particularly in
Mullard thyratrons are chosen so that the spurious the larger valves, the measurement should be made
signal produced by the above effects is negligible and at the valve pins rather than at the socket terminals.
is accommodated within the critical voltage spread. The heater voltage should be set to within X2.5 % of
the nominal value. The valve will, however, accommo-
It will be noticed that for the larger thyratrons a date temporary mains fluctuations of up to ~6%.
maximum grid resistor of the order of 100 k S2 is
In directly-heated valves there is a tendency for
quoted. However, in certain small thyratrons, such
the part of the cathode which instantaneously has the
as the EN91, both the interelectrode capacitances
maximum potential difference with respect to the
and the grid emission currents are so small that a
anode, to be overloaded. If maximum valve life is
maximum control grid resistor of 10 M S2 may be
desired, this effect should be eliminated by arranging
specified.
that the filament supply voltage is 90° out of phase
When the application demands the use of such a with the anode supply.
high value, care must be taken to ensure that the
thyratron base and holder are kept clean and dry HEATING TIME
in order to minimise leakage currents between the It is essential that the cathode shall have reached its
anode and the control grid pins. Similarly, when a full operating temperature, and is therefore capable
thyratron is fed with an alternating anode voltage of its rated emission, before cathode current is drawn.
and has a very high value of control grid resistor, This condition is reached in the time specified in the
the circuit capacitance between control grid and valve data as the minimum cathode heating time.
anode must be kept low. This may be achieved by For gas-filled thyratrons this is the only condition
connecting the control grid resistor directly to the which must be met, since the gas pressure is sub-
grid terminal of the holder, and (in the case of a stantially constant over the permissible range of
tetrode thyratron) by connecting the shield grid to ambient temperature. With a mercury vapour thyra-
the cathode at the holder.
tron, however, it is also essential that the condensed
When thyratrons are operated at anode voltages mercury vapour temperature reaches the minimum
16
value specified in the valve data before cathode IONISATION AND DEIONISATION TIMES
current is drawn. This temperature determines the
Maximum Frequency Limitation
pressure, and hence the current-carrying capacity of
the mercury vapour. If an attempt is made to When the voltage at the grid reaches the critical
draw the maximum rated peak cathode current at value and the thyratron starts to conduct, the anode
reduced mercury temperature, the arc voltage rises, voltage does not drop instantaneously from the supply
resulting in bombardment of the cathode and reduction voltage to the arc voltage. The time required to reach
of valve life. the arc voltage is termed the ionisation time, and
depends upon the grid signal as well as upon the
The time taken to reach a given condensed mercury
geometry of the valve and the nature of the gas filling.
temperature from the instant of switching on the
heater supply can be ascertained from the appropriate Similarly, when the anode voltage drops below the
heating and cooling curves, which are reproduced in arc voltage, the positive ions are not instantaneously
the section on thyratron data. If, for example, the neutralised. These positive ions form a sheath
specified minimum condensed mercury temperature around the grid which prevents the grid from control-
is 40°C and the ambient temperature is 18°C, the ling the initiation of the arc. After a certain time,
condensed mercury temperature must be raised by called deionisation time, the ion density will have
40°C-18°C=22°C before drawing cathode current. decreased to a sufficiently low value for the grid to
The term "ambient temperature " is rather regain control. The deionisation time, like the
indefinite when applied to conditions in the neigh- ionisation time, is a function of the electrode potentials,
bourhood of an electronic valve, and it is therefore the valve dimensions, and the nature of the gas filling.
preferable to work entirely in terms of condensed Generally speaking, high negative voltages on both
mercury temperature and to measure this temperature grid and anode assist rapid deionisation, but any
under operating conditions. The temperature- advantages so gained must be balanced against
measuring device must be one having a low thermal possible electrode sputtering and gas clean-up.
capacity, e.g. a fine wire thermocouple; and it should The deionisation time imposes a limit upon the
be attached to the valve in such a way as to cause the maximum frequency at which a thyratron may be
minimum interference with normal heat dissipation. operated. If the non-conducting periods are shorter
If a mercury vapour thyratron is to be operated at than the deionisation time the grid is unable to regain
a peak current below the rated maximum, it is control and the valve will function as an uncontrolled
permissible to run the valve at a somewhat reduced rectifier. In general, mercury vapour thyratrons are
condensed mercury temperature. Information as limited by deionisation time to a maximum operating
to the reduction permissible for Mullard thyratrons frequency of 500 c/s, while xenon-filled thyratrons lose
may be obtained on application. control at frequencies of the order of 5 kc/s. However,
It is usually unsatisfactory to rely upon an operator for certain valves a considerably lower limit is specified
to observe the necessary time delay between switching to guard against the incidence of gas clean-up and
on the heater supply and allowing cathode current to sputtering.
be drawn from the thyratron. Some form of auto-
In view of the great dependence of ionisation and
matic delay should be employed. For short delays an
deionisation times on circuit conditions and electrode
electronic timing circuit can be used. Some notes on
potentials, any purely theoretical calculations which
suitable thyratron timers will be found on page 27.
make use of these times should be treated with some
The delay device can operate in either of two ways. caution.
It can be used to close the anode circuit of the thyratron
at the expiration of the required delay period; but
this may necessitate the use of a heavy-duty switch. CONDITIONS FOR LONG LIFE
The alternative method is to connect all supplies to The electrode structure and glass envelope of a
the thyratron simultaneously, but with an over-riding thyratron have an almost infinite expectation of life.
negative grid bias voltage which prevents conduction. There are, however, two potentially expendable com-
The delay circuit eventually removes this bias and ponents in a thyratron—the cathode, and (where
allows the thyratron to operate. mercury vapour is not used) the gas filling.
17
The Cathode feeding a purely resistive load may be considered. If
Thyratrons, whether mercury vapour or gas filled, the peak anode current is 80 amperes and the peak
invariably have oxide-coated cathodes. If cathodes of inverse voltage 1500 volts, the commutation factor at
this type are not bombarded by swiftly-moving positive 5 kc/s will be of the order of 120.
ions, their lives are extremely long. It is essential, In some applications, however, one thyratron is
therefore, never to permit the arc voltage to exceed the fully conductive until another takes over the load. At
published value. A high arc voltage will occur if the that instant a large inverse voltage is applied to the
peak cathode current is too high, if the tolerances on anode of the first thyratron. This condition arises in
cathode and heater voltages are not maintained inverter circuits and in three-phase thyratron circuits
within the specified limits, or if an attempt is made to feeding inductive loads. It is likely to impose a high
draw current before the cathode has reached its commutation factor.
correct operating temperature. With a mercury If the commutation factor for a particular appli-
vapour valve it is also necessary to ensure that the cation is excessive, it can be reduced by means of a
condensed mercury temperature is within the limits snubbing circuit. Atypical arrangement is shown in
set by the manufacturer.
18
value for the positive grid current which may safely at the anode. This information is essential in the
be drawn while the anode is more negative than —10 design of reliable circuits. Care must be taken to
volts. Similarly, if, during anode-to-cathode con- ensure that the stray capacitances of the oscilloscope
duction, the grid is more negative than —10 volts, it are small, particularly when measurements are being
will be bombarded by positive ions. The value of made on ahigh-resistance grid circuit.
grid resistor must therefore be sufficiently high to keep
the grid voltage within the published limits. Measurement of Commutation Factor
Mercury vapour thyratrons do not suffer from gas In order to measure the commutation factor
clean-up, but, under conditions of rapidly increasing obtaining in a particular circuit it is necessary to
negative voltage in the presence of positive ions, these examine the anode voltage and anode current over a
ions will be accelerated towards the negative electrode. period of approximately 30 microseconds about the
Since they are of large mass, they may damage the instant of commutation. It is therefore essential to
electrode by impact. For this reason mercury vapour use a timebase with a sweep of that order, and to
thyratons are often limited to a maximum operating arrange that it can be synchronised to trigger at the
frequency of 150 c/s, although the frequency limitation required instant ; and it is customary to obtain the
which would have been set by consideration of triggering signal from the supply line voltage, feeding
deionisation time alone is of the order of 500 c/s. it to the oscilloscope via a suitable delay device or
simple phase-shifting network (see Fig. 20).
It may be safely claimed that the care taken in the
design and manufacture of modern thyratrons has
Rasa,
resulted in valves of this class having long lives. POWER
6UPPLr
TNYRATRON
CIRCUIT
However, if extremely long life is the prime con- DEFLECTOR PLATES
FOR VOLTAGE TRACE
sideration in a particular application, it is a wise DEFLECTOR
PLATES FOR ~~ ,/
policy to under-run the thyratron by, say, 20 % in VARIABLE
PNASE
PULSE
FORMING
CURRENT TRADE
OSCILLOSCOPE
SNIFT NETWORK
respect of both current and voltage ratings. This ~ TRIGGER
~ AL ERNATIVE
reduces the thermal stresses on the valve structure and srs TEM
PULSE , , DELAY ,
diminishes the possibility of grid emission. It also ~---y FORMING ~---~
~ NlTWORKI
LINE J
SIGNAL FOR
MEASURING TECHNIQUES CURRENT TRACE
19
APPLICATIONS
The field of application of the thyratron is too wide Control of A.C. Machines
for complete classification, but it may be said that For synchronous machines, thyratron control of the
thyratrons can be used whenever it is desired to control field excitation can be used to determine important
currents up to some tens of amperes at voltages up to characteristics. With alternators, for example, thyra-
about 2 kilovolts. The thyratron is not the only tron control of the excitation can form the final stage
device suitable for such applications, but when the of an electronic voltage regulator. Similarly, with
control source impedance is high and the operating synchronous motors operating on variable load,
delays may be less than one microsecond, the thyratron thyratrons can be used to ensure automatically the
usually offers the only practicable solution. optimum values of pull-out load and power factor.
Applications fall into two main categories : first, With a.c. motors in general, speed control is
direct power control in heavy industry, where thyra- possible by feeding the machines from a variable
trons with current ratings above 0.5 amperes are used frequency source. This can be generated by a thyra-
for such purposes as motor and welding control; tron inverter. Moreover, by thyratron control of the
and second, light industrial applications such as inverter power supply, both a variable frequency and a
process timing control and the operation of electro- variable voltage can be obtained. Notes on inverters
magnetic devices from high impedance or low current will be found on page 26.
signal sources such as photocells and light metallic
contacts. When it is desired to vary the r.m.s. value of the
voltage supplied to a motor, two thyratrons may be
Some of the principal applications are dealt with connected back-to-back, or alternatively a thyratron
in the following pages. The circuit diagrams given in can be used with aseries-connected transformer to
this section are only for illustration and general form a variable impedance. Further information on
guidance. They are not intended to be directly suit- these applications is given on page 25.
able for every specific application.
20
controlled to ensure constant power and a torque voltage can be connected so that the voltage supplied
characteristic which falls hyperbolically with increase to the armature by the thyratrons is increased by an
of speed. amount equal to IR. In an alternative method, which,
ARMATURE FIELD
however, does not differ in principle, the current
REGULATION ~- REGULATION transformers may be replaced by resistors.
POWER CONSTANT
Constant Operating Speed
Q TORQUE CONSTANT If a constant operating speed is required a further
measure of compensation is necessary, since the
Z armature compensation leaves the system subject to
supply voltage variations. By comparing the armature
a~'
NOMINAL SPEED
SPEED
negligible. This may not be so if very constant speeds
FIG. 22 are required. A small tacho-generator can then be
used, the output voltage of which is compared with a
One of the great attractions of this form of control. is reference voltage and the difference fed back to the
the ease with which deviations from ideal character- control circuit.
istics can be compensated, or any required charac- Table II indicates which Mullard thyratrons are
teristic can be achieved. suitable for armature control of d.c. fractional H.P.
Armature Current Limitation motors. The thyratron rating should be checked
In any control system for d.c. motors the armature against the particular circuit conditions. It should
current must be limited in order to avoid excessive also be noted that when armature control is applied
starting current and to protect the motor and control to a d.c. motor, the r.m.s. current in the armature
gear from overload. In electronic controllers this can windings will be higher than when operating on a
be achieved by the use of a current transformer whose smooth d.c. supply. Care must therefore be taken to
primary is in series with a thyratron. The voltage in ensure that the permissible copper losses for the motor
the secondary, which is proportional to the armature are not exceeded. For a d.c. motor which is not
current of the motor, is rectified and compared with a specifically designed for this mode of operation, it
reference voltage. The difference between the two may be necessary to reduce the rating of the motor.
voltages is amplified and fed to the grid circuit of the
thyratron so that the armature voltage is auto- Control Circuit for D.C. Fractional H.P. Motor H~ith
matically reduced if the armature current exceeds a Singlephase Supply
predetermined value. A simple and inexpensive circuit for manually
controlling a d.c. fractional H.P. motor, working on a
Armature Drop Compensation
230 volt single-phase supply, is shown in Fig. 23. l~'o
The speed of a d.c. motor operating with constant
excitation is not proportional to the applied armature
voltage V, but to V—IR when IR is the voltage drop in
the armature. Although IR may be small and even
negligible at high speeds, it can be comparable with V
at low speeds. Thus to obtain a fairly linear voltage/
speed characteristic over a wide speed range, it is
necessary to compensate for the IR drop.
An additional secondary winding on the transformer
used for armature current limitation will provide a
voltage proportional to the armature current. This FIB. 23
21
TABLE II
TYPE OF THYRATRON
input transformer is required. The circuit provides connected to terminals A and L, and the field coil
armature control only, and the field excitation is across F and L. Constant field excitation is obtained
obtained by half-wave rectification. Field control from V6. The armature supply voltage is varied by
could be introduced if required. the thyratrons V4 and V5, the control grid signal being
Variable phase or "horizontal "grid control of the a variable d.c. voltage with a superimposed a.c.
thyratron is employed, a 50 kS2 potentiometer being voltage delayed 90° with respect to the appropriate
used as the control element. The 200 pF capacitor anode voltage.
connected between the grid and cathode of the
thyratron prevents random striking caused by transient The speed of the motor is varied by the potentio-
voltages induced in the mains supply or in the control meter Pl. " IR drop " compensation and armature
circuit by associated equipment. Choice of thyratron, current limitation are determined by the preset
1
field current rectifier, and smoothing circuit, will potentiometer and resistor, PZ and Rls respectively.
depend upon the size of the motor and the particular All these actions are combined in the double triode
application. Automatic control of the heating-up V2. The left hand section V2A controls the speed
time of the thyratron is not indicated in the diagram, and IR drop compensation, whilst the right hand
but must be included in practical circuits. section V2B is used for armature current limitation.
Comprehensive Control System fora D.C. Motor, The armature voltage is used as a measure of the
incorporating IR Drop Compensation and Armature speed of the motor. It is compared with the fraction
Current Limitation of the field voltage determined by the setting of the
The circuit diagram Fig. 24 shows a simple but potentiometer Pl. If the speed falls, the armature
comprehensive motor controller, the armature being voltage falls below the reference voltage, and negative
22
0 0
T2
0 4Q944444Q.4Q.44_Q.9
FIG. 24
bias will then be applied to V2A and the striking of variation of about 10: 1. Provided the field is not
V4 and VS will be advanced to restore the armature saturated, the variation of speed due to mains
voltage. The reverse will apply for an increase in fluctuations of ~ 10 % will not exceed ~ 2 %. For
speed. By changing the setting of Pl the speed of the small motors, V6 can be a vacuum diode such as a
motor can be varied over a wide range. Mullard type GZ32. For larger equipments a gas-
The armature voltage drop compensation is added filled rectifier should be used.
to the speed control by connecting the grid of V2A
to the slider of the potentiometer P2. This is adjusted RESISTANCE WELDING CONTROL
to give a linear speed characteristic at low speeds, In many applications the full benefit of resistance
the compensation remaining operative over the entire welding can be realised only if the welding current and
speed range. A delay circuit C3 R 4, prevents instability its duration can be accurately controlled over a current
due to too rapid compensation. scale ranging from nearly zero to the maximum
Armath.ire current limitation is obtained by a current obtainable from the equipment, and a time
current transformer T3 and a rectifier V3 which produce scale covering from 10 milli-seconds to several
a voltage proportional to the d.c. line current. At seconds. Exact repetition of welding conditions is
normal values of current V2B is held non-conducting, also of great importance. Electro-mechanical timing
but if the limit is exceeded it will conduct and the and control methods are restricted to long minimum-
thyratron striking will be delayed, so decreasing the time intervals, and they tend to be unsatisfactory when
applied armature voltage. used in highly inductive circuits. Electronic methods,
It will be seen that the primaries of the current however, permit accurate control of current and
transformers also carry the field current. This, with duration over a wide range.
the relay R/2, ensures that the field is established The supply to the welding electrodes can be directly
before the armature voltage can be applied. The controlled by two thyratrons connected back-to-back.
thyratrons will be cut off until V2B is biased back. For very large welders, thyratrons may be used to
This control system will give a substantially flat control the firing of ignitrons or other mercury-pool
speed/torque characteristic over a range of speed devices.
23
SW 2 HaiorCool welo
MAIN F All components includ¢0
SWITCH MrMostal
eat contra in this section eSWI arc
¢xt¢rnal to TIMER
I WELDING TRANS.
IBOV
Nominal L
56kn PKNG TRANS
C
a T3
RI
3knbw Initiating ~ q
switch
4pF~ SWI IOOk n~
v r DE LAv 25OV R2
ISkn
D To Neural TO2yF
II CIRC UIT 23OV SS 1 ISOV
r r= =+7 3w
vl vz I
54 (SV) CI
~i — e8 Z3O0T 54
I 0 5rF 1267
i I 25aV 47kn (Leader)
AG- - lo%
I SS (14V) Iw IBOk
Irr= I 7 •_ V3
I 53 (SV)
RU —+3
I B• b
~3 i ni kn (Follower)
S3 (14V) PKNG Westinghox ~
I TRANS 22kn 16k4 OV
I 4
T2 «~o~ czl
05uF
I I Skn < ISOVT
I Iw
tJ¢u~t,rJ(ol Mains Trans. I T ? 53
TI
(n conne~ted to chassii~
FIG. 25
It is possible to apply the normal power supply the control of ignitrons in larger equipments the
directly to the welding electrodes by using one of the XG1-2500 mercury vapour thyratron and the
control methods described on page 25. The thyratrons XR1-3200 gas-filled thyratron have peak current
can be struck at any instant in the positive half-cycles ratings of the required value, namely, 40 amperes.
of the supply voltage, so that full control of the
welding current is obtained. Single-Cycle Timer for a Small Spot Welder
Control of the duration of the welding current A control circuit for asingle-cycle welder is shown
may be achieved in the following ways: in Fig. 25. It was developed for use in valve
(1) By using a variant of the RC timing devices manufacture but can be used for any type of small
described on page 27. assembly work where short and reproducible welding
(2) By using aflip-flop circuit with a variable times are required.
change-over time. This is a variant of The current in the primary winding of the welding
method (1). transformer is controlled by two thyratrons, V2 and
(3) By counting the cycles of the supply frequency V3, which are connected back-to-back. The design
with a decade scaler such as the Mullard can, if required, be extended to give welding times of
type E1T and allowing the welding time to more than one cycle.
continue for a certain number of cycles. The method of operation is as follows: when the
For welding operations with a duration of only one main switch is closed a delay circuit energises relay
cycle (one-fiftieth of a second) it is possible to control RLl at the end of the heating-up time appropriate to
only the starting point. The welding current itself the thyratrons. A d.c. supply of 180 volts can then
can be varied only by changing the impedance or the be applied to the anode of the cold-cathode trigger
voltage of the welding transformer secondary. tube V1 by the weld initiating switch SWl. The
The Mullard range of thyratrons adequately covers trigger electrode of V1 is held at ~ 57 volts, which is
the requirements of electronic welding control. For below its strike voltage, and superimposed on this is
direct control, valves are available with peak current the output voltage from the peaking transformer T2.
ratings up to 80 amperes (Type XR1-6400), while for This transformer is fed from aphase-shifting network
24
C1, R 1 and R2, giving approximately the same phase the average duty of the series valves will be con-
delay as the loaded welding transformer. The values siderably decreased by the effective reduction of the
shown correspond to a delay of 35°. When SW1 is mains variation by the thyratrons. The hard valves
closed V1 will strike when the supply voltage reaches
vncuvw vuve
a positive value of ~/2 x 230sin35° volts. As a I REGULATOR
25
waveform (and hence r.m.s. voltage) can be obtained With the switch open, both thyratrons are prevented
by suitable adjustment of the bias. from striking by a negative bias of approximately 20 volts
obtained from the metal rectifiers. When the switch
1692 is closed, the grid of each thyratron is connected to the
mid-point of a 3.3-10-10-3.3 kS~ potentiometer
across the valves. Thus, during a given half-cycle one
thyratron will conduct, and the arc voltage will appear
across the other during most of the half-cycle. The
two thyratrons will conduct alternately in successive
half-cycles.
THYRATRON INVERTERS
Static alternating current generators of medium
power, capable of operating at either constant or
FIG. 28
variable frequency, have many applications. For
example
Fig. 28 shows an alternative method of control.
The load is connected to the mains in series with 1. Generation of alternating current at normal
the primary winding of a step-up transformer. The supply frequency to enable a.c. mains-
secondary winding is connected to a thyratron, or operated equipment to be used where only a
two wired back-to-back. If the arrangement is biased d.c. supply is available.
to cut-off, only the comparatively small magnetising 2. Production of variable frequency power for
current of the transformer will flow through the load. speed control of a.c. motors.
If, however, the thyratron circuit is made to conduct, 3. Low frequency power generation for vibrators.
the voltage across the transformer primary will be 4. Generation of power for induction heating.
equal to the arc voltage of the thyratron divided by
the turns ratio of the transformer; thus the greater part These and other requirements for power outputs up
of the supply voltage will be applied to the load. By to about 40 kW can be met by using thyratrons in
multi-phase inverters. The principle of the single
delaying the striking of the thyratron, the time for
which the greater part of the supply voltage is applied phase parallel inverter can be explained by reference
to the load is decreased; therefore the r.m.s. voltage to Fig. 30, which shows the basic circuit of the so-called
constant voltage inverter.
applied to the load also decreases. With a suitable
choice of transformer ratio the thyratron can be A variable frequency oscillator may be used to
operated near its full current and voltage ratings, trigger the thyratrons alternately through the grid
thus a high valve utilisation factor is obtained. transformer. When thyratron Vl is conducting, a
voltage applied to the grid of the second thyratron
An On-Off Switch for A.C. Circuits Vz will cause this valve to strike. There is a rapid fall
The circuit diagram reproduced in Fig. 29 is that of
an electronic on-off switch using two thyratrons con-
-~ LOAD
nected back-to-back, the switching operation being 3'3 kn
IOkn ~/(~/~
consequent excessive wear of switch contacts. The 50 ~.
SUPPLV
~ IOkn
26
220 volts. It operates in the manner described above,
except that self-excitation is obtained from the winding
WS2 via the phase-shifting network R5, C2.
To allow the cathodes of the thyratrons to heat up,
operation of the anode circuit is delayed for 10
seconds by relay R/ 1. During this period the heaters
are fed from the d.c. input via the dropping resistor
R2. R/1 is a thermal relay which, after the l0~second
delay, switches on the main circuit of the inverter.
The voltage induced in the winding WS3 then
energises relay S/ 1 which disconnects the heaters and
R/1 from the d.c. mains and connects them to WS2.
This circuit will operate at 100 c/s if RS is reduced
and the value of C1 is changed from 0.75 µF to 0.3
FIG. 30 µF. The action of the device is also reversible, so that
by feeding the transformer from a 220-volt 50 c/s
supply a d.c. output of 0.2 ampere at 180 volts can be
in the anode potential of V2 to a value equal to the arc obtained. This inverter was designed for operation
voltage. In a mercury vapour thyratron this at low d.c. and a.c. voltages and therefore does not
voltage is about 12 volts. The fall of potential is give optimum efficiency.
transmitted to the anode of Vl by means of the com-
mutating condenser C. The anode potential of Vl at
Varc (since the valve is already struck) will be forced
below this value and the valve will extinguish. The
anode voltage must remain below Varc long enough
for the valve to deionise. The voltage across Vl then
rises at a rate determined by the circuit constants.
Before the next trigger pulse is applied to Vl its anode
potential must be above the striking voltage (about
40 to 50 volts) for the inverter to operate. This
switching of the thyratrons causes alternating pulses
of current to flow in the transformer primary, and an
alternating voltage is induced in the secondary winding.
The frequency is determined by the trigger repetition
rate.
The circuit is externally excited by the grid pulse
oscillator, but it is possible, particularly in constant-
load, constant-frequency applications, to supply the FIG. 31
control grid circuit from the output, and thus to make
the circuit self-exciting. In general, however, the
externally excited inverter is preferable by reason of TIMING CIRCUITS
its greater stability. Small thyratrons such as the Mullard types EN91
and EN32 may be operated with high impedance grid
circuits, and are therefore suitable for use in timing
Low-power Self-excited Inverter circuits where this is of primary importance. They
The circuit of aloes-power, self-excited inverter can also operate quite large relays or contactors
using a pair of EN91 thyratrons, is given in Fig. 31. directly.
It is intended for operation from 220-volt d.c. mains, Typical examples of such use are resistance welding,
and it gives a usefu150 c/s output of about 24 watts at punch pressing, and pre-set heat cycle control.
27
A basic circuit for many such systems is shown in the half-cycles when the grid is positive with respect
Fig. 32. With the switch in position A, the thyratron to the cathode. Closing switch SW 1 completes the
grid is biased to —100 volts and the valve cannot anode circuit, and the voltage across C1 appears as a
conduct. If the switch is thrown to position }3, the negative bias on the control grid. Conduction will
capacitor C will start to discharge through resistor R, start after an interval determined by the values of RV1
and the grid will become less negative at a rate deter- and C1. As the thyratron supplies pulses and not a
mined by the time constant RC. When the critical steady direct current, the capacitor C2 is required to
grid voltage is reached the valve will strike and will prevent relay chatter. An additional resistor R4,
initiate action in the anode circuit. in series with the relay coil, may be needed to limit the
average anode current.
In designing this circuit care must be taken to en-
sure that the maximum permissible pre-conduction
grid voltage of the thyratron is not exceeded. This
imposes a voltage limit between points A and B of
70 volts r.m.s. for thyratron type EN91 and 180 volts
r.m.s. for type EN32. Where a higher supply voltage
1697
FIG. 32
28
controller, a simple "ringing "circuit is usually em- strike voltage the value of the resistor should be as
ployed to strike the lamps. Thyratrons determine the small as possible, but the maximum grid current
point in each half-cycle at which the circuit rings, and rating should not be exceeded.
they also carry the subsequent lamp current.
CONTROL OF VIBRATORS
There is a considerable demand for controlled
FIG. 34
vibrating equipment, both for electromagnetic con-
veyors and for programmed mixing. Thyratrons can At high supply voltages a high value of strike
provide both switching and continuous control. voltage, say 200 volts, may not be objectionable, and
the control grid may be connected directly to the
FIRING OF MERCURY POOL RECTIFIERS cathode.
For accurate firing of ignitrons and grid-controlled For operation at medium voltages the control grid
mercury-pool rectifiers it is necessary to feed the should be connected to a potentiometer between anode
igniting or control electrode with a large current signal and cathode as shown in Fig. 35 To limit the inverse
which has a steep leading edge. Thyratrons are very
suitable for this purpose, and they also present a high
input impedance to the signal source. This is par-
ticularly important in a feedback system with fine
control.
For ignition duty the peak-to-mean current ratio is
considerably higher than for other applications. For
certain thyratrons, therefore, e.g. XG1-2500, a special
ignitron firing rating is quoted. Designers who wish
to use other types of thyratron for this service, where
the published ratings are likely to be exceeded, should
FIG. 35
Consult the valve manufacturer.
current the total resistance RI-~R2 should be large.
THYRATRONS AS UNCONTROLLED The maximum value of R1 is determined by the grid
RECTIFIERS current required to strike the valve at a low anode
The thyratron is essentially a rectifier with the voltage. The ratio R1/R2 is set by the maximum
conduction period in each positive half-cycle con- permissible grid-to-cathode pre-conduction voltage.
trolled by the potentials applied to the grid. It can, The values of R1 and R2 are thus dependent on the
however, be used to provide uncontrolled rectification type of rectifier circuit, the supply voltage, and the
by means of simple passive grid circuits which allow characteristics of the thyratron.
conduction over most of the positive half-cycle. The Table III shows suitable R1/R2 ratios for different
high inverse-to-forward impedance ratio inherent in valves and for different values of transformer
gas-filled valves should be retained. secondary voltage. The table applies to the circuits
At low supply voltages the control grid may be given in Table IV. For the single phase bridge
connected to the anode via a resistor of suitable value system the R1/R2 ratios relate to twice the transfor-
as in Fig. 34. This method must not be used where the mer voltage quoted in Table III.
peak inverse voltage exceeds the rated maximum Table III is based on a maximum R2 value of
permissible grid-to-cathode voltage. To obtain a low 100 k S2. Lower values can be used, but. the ratio
29
TABLE III
GRID CIRCUIT RESISTANCE RATIOS FOR UNCONTROLLED RECTIFIER CIRCUITS
0.1 0.2
1.0 5.7
1.5 9.2
TABLE IV
CHARACTERISTICS OF RECTIFICATION SYSTEMS
Three-
Single- Single- Three- Three- phase Hexaphase
Type of Circuit phase full phase phase half phase full double Y half wave
wave bridge wave wave with reactor
No. of Valves 2 4 3 6 6 6
Peak valve current 1 xIa~ 1 xIa~ 1 xIa~ 1 xIa~ 0.51 xIa~ 1 xIac
Average valve current 0.51 xIa ~ 0.51 xIa~ 0.33 xIa~ 0.33 xIa~ 0.167 xIa~ 0.167 xIao
30
R1/R2 must not be decreased. The ratios quoted regulation or other voltage losses in transformer or
permit the use of resistors with a tolerance of ~ 10 %. valves.
In Tables TIT and IV all voltages have been quoted For sine wave operation the resistor R1 can be
in terms of the transformer r.m.s. voltage instead of replaced by a capacitor, with an improvement in the
the more usual d.c. output voltage—which is depen- inverse voltage characteristics. of the system. How-
dent on load current. Errors may occur if the peak ever, any transient voltages which appear during the
inverse voltage per valve is determined from the negative half-cycle may give rise to excessive grid-to-
measured d.c. output voltage. Table IV assumes no cathode voltages.
31
BIBLIOGRAPHY
GENERAL
Industrial Electronics Reference Book.
Westinghouse Electric Corporation, 1948.
Applied Electronics.
Electrical Engineering Staff. Massachusetts Institute of Technology. John Wiley and
Sons, 1943.
Industrial Electronics.
R. Kretzmann. Philips Technical Library, Cleaver Hume Press, 1953.
MOTOR CONTROL
Inverter Action on Reversing of Thyratron Motor Control.
H. L. Palmer and H. H. Leigh. Transactions of the American Institute of ~ Electrical
Engineers, Vol. 63, April 1944, pp. 175 to 184.
32
CIRCUIT DESIGN
Grid Control of Gas-Filled Tubes.
W. D. Cockrell. Electronics, Vol. 17, No. 6, June 1944, pp. 124 to 128.
RECTIFIER CIRCUITS
Critical Inductance and Control Rectifiers.
W. P. Overbeck. Proceedings of the Institute of Radio Engineers, Vol. 27, October 1939,
pp. 655 to 659.
A Graphical Analysis of the Voltage and Current Wave Forms of Controlled Rectifier Circuits.
P. T. Chin and E. E. Moyer. Transactions of the American Institute of Electrical
Engineers, Vol. 63, July 1944, pp. 501 to 508.
RESISTANCE WELDING
Thyratron Control Equipment for Resistance Welding.
H. L. Palmer. Part I: Spot and Seam Welders. General Electric Review, Vol. 40, May
1937, pp. 229 to 235. Part II: Timing and Heat Control. July 1937, pp. 321 to 325.
INVERTERS
Parallel Inverter with Resistance Load.
C. F. Wagner. Electrical Engineering (Transactions of the American Institute of
Electrical Engineers), Vol. 54, November 1935, pp. 1227 to 1235.
MISCELLANEOUS
Timing Machine Operations with Small Thyratrons.
S. C. Rockafellow. Product Engineering; November 1950, p. 85.
33
TECHNICAL DATA FOR MULLARD THYRATRONS
The final section of this publication contains full determined by the available safe cathode emission,
technical data for the Mullard ranges of mercury whereas the average current is limited by its heating
vapour and gas-filled thyratrons. The data should effects. For periods less than the maximum averaging
be read in conjunction with the following general time it is permissible to draw average currents greater
recommendations. than the maximum rated value. In this case it should
Installation be understood that this maximum value must not
Mercury vapour thyratrons should always be be exceeded over any period equal to the maximum
mounted vertically with the cathode connections at the averaging time. When different values of peak current
lower end. When a mercury vapour thyratron is first are quoted for different frequencies of operation,
installed, and before it is put into service, it should be this must be taken into consideration.
run for at least half an hour at its normal heater or The surge (fault protection) current rating indicates
filament voltage but without any electrode voltages the maximum value of transients, resulting from a
applied, in order to vaporise any mercury which may sudden overload or short circuit, which the valve will
have been deposited on the electrode assembly during pass for a period not exceeding 0.1 seconds without
transit. This precaution should also be taken before immediate destruction. Several overloads of this
putting into service a mercury vapour valve which nature will, however, appreciably reduce the life of
has been out of use for any considerable time. the valve.
Gas-filled thyratrons have much less rigid limita- Use of Control Characteristics
tions of mounting position. The smaller types may The control characteristic is shown by upper and
be operated in any position. lower boundary curves within which all valves may be
Limiting Values expected to remain during life. The control charac-
It is important that these limits are never exceeded. teristic of a particular valve may move within these
Such variations as mains fluctuations, component boundaries, although, as a rule, these limits should be
tolerances, and switching surges must be taken into considered as extreme cases. This should be taken
consideration in arriving at actual valve operating into consideration when designing grid excitation
conditions. circuits for thyratrons.
Filament Supply Effects of Positive Ion Current
The heater or filament voltage of a thyratron should When a thyratron is conducting, a positive ion
be maintained within ~ 2.5 % of its specified value. current of magnitude proportional to the cathode
The valve will, however, accommodate temporary current is generated. This current will, in general,
mains fluctuations up to ~ 6 %. The filament trans- flow to the electrode which is at the most negative
former for directly-heated types should be centre- potential during conduction. In order to prevent
tapped. damage to the valve it is necessary to ensure that the
Current Ratings voltage of this electrode is more positive than —10 V
For each rating of maximum average current, a during this phase. This precaution will prevent: an
maximum averaging time is quoted. This is to ensure increase in electrode emission due to excessive
that an anode current greater than the maximum electrode dissipation; sputtering of electrode material;
permissible average value is not drawn for a time long changes in the control characteristics caused by shift
enough to produce an excessive temperature within in contact potential; and, in inert gas-filled valves, a
the valve. The maximum peak anode current is rapid gas clean-up.
34
In circuits where the control grid is held negative determined from the heating and cooling character-
during anode conduction, a suitable choice of resistor istics. This precaution ensures that the cathode is at
in series with the grid will maintain an effective grid the correct temperature for electron emission and also
bias more positive than —10 V. The correct value that the gas pressure, in mercury vapour valves, is
of resistor may be determined from the grid ion within the required limits. In applications where the
current characteristic. If the instantaneous value of ambient temperature is unduly low, some means of
anode current is low then the restriction on grid bias auxiliary heating may be required to assist in attaining
does not apply. In general, the grid should be more the requisite minimum condensed mercury tempera-
positive than —10 V for all values of anode current ture.
greater than 10 per cent of the rated maximum average It is necessary to provide adequate ventilation round
current. the valve so that the maximum ambient or condensed
In circuits where the anode potential changes from mercury temperature is never exceeded. This avoids
a positive to a negative value and the control grid is the danger of arc-back or loss of grid control.
at a positive potential, thereby drawing cathode When it is necessary to check the condensed mercury
current, a small positive ion current flows to the temperature of mercury vapour thyratrons the
anode. In such a case the inclusion of a high value of following procedure should be adopted. A tempera-
anode resistor is precluded by circuit requirements and ture indicator of low thermal capacity, such as a fine-
the anode will usually reach a high negative potential. wire thermocouple, should be attached to the valve
It is therefore essential to limit the magnitude of the at the mercury condensation point by means of the
positive ion current by severely restricting the current minimum amount of suitable adhesive or cement.
flowing from cathode to grid. This may be effected Care should be taken to ensure that other conditions
by using the maximum permitted series grid resistor of operation, such as load current, ambient tempera-
and (or alternatively) keeping the positive grid voltage ture of the air outside the equipment, and ventilation,
swing as low as possible. remain unchanged during the measurement.
In circuits where the anode potential changes very
rapidly from a positive to a high negative value, as Power Supply Frequency Limitations
with inductive loads fed from polyphase supplies, In general, when thyratrons are operated at fre-
there will be residual positive ions which will be drawn quencies below 25 c/s, a lower maximum peak anode
towards the anode with considerable energy. In an current is applicable. This is necessary to ensure that
inert gas-filled valve this would result in excessive gas cathode fatigue does not occur.
clean-up, therefore it is necessary to observe the The maximum frequency at which a thyratron will
limitations imposed by the appropriate commutation operate satisfactorily is dependent upon the deionisa-
factor. tion time and therefore upon the conditions of opera-
Valve Temperature Limitations tion. At higher frequencies the valve will fail to
To ensure that the requisite valve heating time has operate because of arc-back and loss of grid control.
elapsed before electron current is drawn from the At high frequencies the commutation factor should be
cathode, it is necessary for the heater or filament to kept as low as possible in order to ensure satisfactory
be run for the appropriate time which may be life.
35
EN32
a
91
~0 92
- 92 ~ ® O ~
91 ~~~
~ ~ h
h k
Octal Bas¢
PRELIMINARY DATA
The EN32 is a tetrode thyratron with an inert gas CHARACTERISTICS
filling and an indirectly-heated cathode, mounted on Electrical
an international octal base. With the recommended Heater voltage 6.3 V
connection of the shield grid to the cathode the valve Heater current at 6.3 V 0.95 A
has a negative control characteristic. The small con- Control ratio
trol-grid current allows a high value grid resistor to be g2 to k and Rgl=O S2 275
used, so that the EN32 can be controlled from a high gl to k and Rg2=0 S2 370
impedance signal source such as a photocell. The Arc voltage 10 V
valve is suitable for use in relay circuits, regulated
Mechanical
power supplies, and general industrial applications.
Type of cooling Convection
The Services type number is CV2253. Mounting position Any
Max. average positive control-grid current for anode voltage more positive than --10 V (averaging
time one cycle) 20 mA
Max. peak positive control-grid current for anode voltage more negative than — ] 0 V 1.0 mA
36
E N 32
LIMITING VALUES (continued)
Max. average positive shield-grid current for anode voltage more positive than —10 V (averaging
time one cycle) 20 mA
Max. peak heater-cathode voltage with .heater positive 25 V
Max. peak heater-cathode voltage with heater negative 100 V
Min. valve heating time 15 secs.
Ambient temperature limits —75 to -{-90 °C
Where circuit conditions permit the shield-grid should be connected directly to the cathode.
IIu11N1
.....
1111111Nllllllllill~~! 111111
...■......■................1.........
a
....... . ....■■■■■■■■■■■■■■■■■■■■II■■■■■■■■■
11111111
I Nlllllil11111l111111111il11111l111
IIIIIIN 1pIIIIIIIIIIIIIIIIIIIIIIIINIIIIII
iiiiiliiiiilniiiiiiiiiiiiiiiii■i~iiNiiiai
11111111 IIIIIIIIIIIIIIIIIIIINI!1~IIN111111
■■■ ■■■ ■■■■■N■■■■■■■■■■■■■■■■II■ ■■■■■■ ~~~~~~~~~ Rangc for Rg -10M[1
O
■■11■■■■■■■■■■■■■ O
Range to Rgl-=IOOkf)
11■■■1:'IC::CC::::1:::=:::IC:CN1:1~.N::.:CC 1111111111111111 1
11111IIIIIIIIIIIIIINIS~~IINIINII
''!"~!'I""
11111 111N~ II~~11111!11111111111111!111111 illl!1111111
1111.11111111►IIIIIliiilllllllllll1111111!111 0
Ia-30mA- ................::i:rie:::::■■■:
~~~- ■■■■■■■■■■■■■■■■N~111■■■■■N■
11111
.■...
NIIIIIIIIIIIIIIIIII~~IIINI
■■■.................... 1111
1 a
■■■■■ ■■■■■■■■■■■■■■■■■■w!S~■■~/I 11 ■
■■■■■ _ ■■■■■■■■■■■~~~i.■■■■■~■Illll ■
■■■■ IOOmA N~~~ai~■■■■■■■■■■■■■■'It111■ ■
■MI N■■■■■■■■■■■■■■■■■■■■III■ 1■ ■~
~■■
■q ■
.w■■...■ ■■■■■■■■■■
■ ■■■■~►~~ ■ 1
■■■■■■■N■■■■■■N■■■■~~11■ 1■
■■■■■
■N■1
.III
■■■■■■■■N■■ ■■■■■■■■■■■■■■I■■■■■■■ ■■■■■■■■■
111111111111111~1111IIIIIIIIII11111IlQlll~llll
0
Nlil""""'iii_iiiiii;illllllll111111111111
■■■ 300mA ~~~~~~•~■■■■■■■■■■■■■■■■■■■■■■■■■■
w■■■■■■■■■■■■■■■■■■■■■■■n■■■■■■■■
■■■■■■■■■■■■■■■■■■■r■■■■~~■■■■■■■■■■■■■n■11~■
IIIIIIIIIIIIIIIIIIIICl11111~■■■■■'■■'■■■'~'°••
1N11
lN11 ~~
vq(y) -2 -O
N1~!l~11!l~~~~1~~~~~1~!!~I
-8 -6 -4
! !! !
-2 4
■■■■■■■■■■■■■■■rr■■r■~■■~.~~~■■~_—
■■■■■■■■■■■■■■■■■■■■■■■■■■■■~l~e~e~!www~www■
iiiiiiiiiiiiiiiiiiiiiiiirwiiiiiiiiiiiiiiiiiiii 0
V91 V~ -8 -6 -4 -2. O
CONTROL CHARACTERISTIC
The curves indicate the spread in characteristics
due to
(a) Variations in characteristics due to changes in
heater voltage.
(b) Variations in characteristics during life.
(c) Variations in grid resistor.
37
EN70
92 k
h ~ o
92 0
91 92
a 92
B8D BASE
~IOmm. subminiatur¢~
11~
38
E N70
LIMITING VALUES (continued)
Max. control-grid resistor 10 M S2
Recommended min, control-grid resistor 100 k S2
It is not desirable that the control-grid should be positive when the anode is more negative than
—10 V, but where this condition is unavoidable the control-grid resistor may need to be
greater than the recommended minimum value.
Max. negative shield-grid voltage before conduction 100 V
Max. negative shield-grid voltage during conduction 5.0 V
Max. average positive shield-grid current for anode voltage more positive than —10 V (averaging
time one cycle) 700 µA
~ EN70 oss
E
e.
>°
i /,
Heater voltage 6~3 V
R.M.S. anode voltage 150 V 0 f
0
Shield-grid voltage 0 V /. r,
R.M.S. control-grid voltage 5.0 V
(180° out of phase with anode voltage)
0
*Peak control-grid signal voltage 5.0 V 0
N
I
Control-grid circuit resistance 1.0 M SZ
r
Anode circuit resistance 3.75 k S2
*The frequency of the signal is high compared 0
with 50 c/s. 0
0
Vgi(V)-3 —2 O
CONTROL CHARACTERISTIC
39
E N 70
w
O Vh = 6.3 V
O RQ = IOKf2
1 R92 = O
Rye = IOOKII
1
N
O
SPREAD OF CONTROL O
CHARACTERISTIC WITH
SHIELD-GRID VOLTAGE AS
PARAMETER
1
O
O
O
Vgt(V)-5 —4 —3 —2 -I O 2 3
I N7O I I I I I I I 1026
O Vh = 6.3V
O
V92 = OV
O ~//
I]~I'
i Ia= 2 mA
O 1 //,
O
I _//
N ~~~
O
O IOmA ~~~_
i GRID ION CURRENT
CHARACTERISTIC
O
O
~dw1i6i oos oab
Vg~(V)- 9 —8 —7 —6 —5 —4 —3 —2 O
40
EN91
h
92
9i
B7G BASE
41
EN91
LIMITING VALUES (continued)
Max. control-grid resistor 10 M S2
Recommended min. control-grid resistor 100 k S2
It is not desirable that the control-grid should be positive when the anode is more negative
than —10 V, but where this condition is unavoidable the control-grid resistor may need to be
greater than the recommended minimum value.
Max. negative shield-grid voltage before conduction 100 V
Max. negative shield-grid voltage during conduction 10 V
Max. average positive shield-grid current for anode voltage more positive than —10 V (averaging
time one cycle) 10 mA
-~~
.'
~N
120 ■ EN91 ~■ E
f
~..
4
0
0
M
I
■1 OPERATING RANGE OF
CRITICP,L GRID VOLTAGE
O
1 N
1
1
,`
Vgl(V} 6 -4 -2 O
42
E f~0 91
Vh= 6.3 V
R92 : Of1
R9 = OA
A
O
O SPREAD OF CONTROL
CHARACTERISTIC WITH
SHIELD-GRID VOLTAGE
AS PARAMETER
8V
fJ
0 s ~o is vgl(v,
~SOmA_
N
T IOOmA va•2sv
m
y 150mA
Zo
400V
~~e ---~.■
_~~■■ ■
■■~~~~■
■
~~ N
600V O
■
— 200mA v
O
0
300mA
x
ao ■ x
Vg~(V~ -6 -4 -2 O vgl(v) - 12 -8 -~ O
CONTROL-GRID CURRENT PLOTTED AGAINST CONTROL- CONTROL-GRID CURRENT PLOTTED AGAINST CONTROL-
GRID VOLTAGE DURING CONDUCTION GRID VOLTAGE BEFORE CONDUCTION
4~
76 mm.
XGI-2500 ~ L°
max.
(MT57)
a
185mm.mcx.
E .
Ex
O~
r
M¢dium 4 -pi n
Bas¢ '~/Jith Bayon¢t Catch ~
91~
Max. peak forward anode voltage (condensed mercury temperature 40 to 75°C) 1.0 kV
Max. peak forward anode voltage (condensed mercury temperature 40 to 80°C) 1.0 kV
Max. condensed mercury temperature rating for intermediate anode voltages may be
determined by linear interpolation.
44
XG 1-2500
LIMITING VALUES (continued)
Max. average positive control-grid current for anode voltage more positive than —10 V (averaging
time 15 secs.) 250 mA
Max. peak positive control-grid current for anode voltage more positive than —10 V 1.0 A
Max. peak positive control-grid current for anode voltage more negative than —10 V 100 mA
i'n_' '■■
HEATING AND COOLING CHARACTERISTICS
XG I- 2500 919
~!_!!!
..._...
r~1
■■■■
iii■►,ii■ii■iii■iii■~~►~■i ■i■
■■■ ■■■■■
■■~_■~■■■■■_■■■~1~
■ • ~■■■■■■■_■■■_■.1
Supp 1¢s switch¢d off
■~n■~■■■_■■■w■■■_■■■__■~i►_~~~■■■■■
0
'ii■ii'iiiii■iii■iii■■'~iiin
■I~■_■■ _■■■_■■■_■■.~■. .ate-.
O Ib 20 30 40 O 10 20 30 Time (minutes
TYPICAL EXAMPLE OF THE APPLICATION OF therefore to wait until the excess mercury temperature
HEATING AND COOLING CHARACTERISTICS has risen by 25° before applying the h.t. supplies.
The minimum permissible condensed mercury From the heating characteristic it will be seen that
temperature for XG1-2500 is 40°C and, assuming this entails a delay of at least 24 wins. After applying
an ambient temperature of 15°C, it is necessary the h.t. the valve is operated at a mean anode current
45
of 2.5 A for a time sufficient for the condensed mercury XGI - 2500 9 18
temperature to reach equilibrium conditions. If, now,
both h.t. and filament supplies are switched off for a
period of 20 mins. it is observed from the cooling
characteristic that the excess temperature at the end
of this period will be 5.3°.
Upon switching on the filament again the heating-up
time indicated on the heating characteristic for an
excess temperature of 5.3° (i.e. 5 mins.) may be
deducted from the total heating up time of 24 mins.
originally required for an excess temperature of 25°
to give a reheating time of 19 mins. before applying
the h.t. supplies again.
In each case the time required for re-heating exceeds
the minimum cathode heating time of 5 minutes.
When the normal operation of the valve is at a
lower peak current than the maximum rating, it may
be permissible to reduce considerably the preliminary
valve heating time. If such a reduction is proposed,
however, reference should first be made to Mullard
Ltd.
XGI-2500 917
0 Vg (V)-10 -5
GRID ION-CURRENT CHARACTERISTIC
O
O
~/O
0
0
e
r! CONTROL CHARACTERISTIC
46
.r 62mm.
max. XGS-500
~ La (MT17)
a
E .
Eo
`f E
o~
~o
M¢dium 4 -pin ~
Base With Bayon¢t Catch 19121
47
XG5-500
LIMITING VALUES (continued)
Max. average positive control-grid current for anode voltage more positive than —10 V (averaging
time 15 secs.) 50 mA
Max. peak positive control-grid current for anode voltage more positive than —10 V 250 mA
Max. peak positive control-grid current for anode voltage more negative than —10 V 50 mA
Min. valve heating time—see heating and cooling characteristic on page 49.
Condensed mercury temperature limits -{-35 to x-70 ~C
Max. power supply frequency 150 c/s
XG5-500 it 913
O
O 914 XGS-500
N
aE
0
0
,~ Io ~ 50m A
IOOmA
0 0
0
0 200mA
300mA
0 0
N
i
400mA
500mA
0 0 I
m
Vg (V) -I 5 -10 - O 600mA
CONTROL CHARACTERISTIC
The shaded area between the curves indicates the
spread in characteristics due to:— 0v
(a) Initial differences between individual valves. Vg (V) -F5 -10 O
-5
(b) Variations in characteristics during life.
GRID ION CURRENT CHARACTERISTIC
(c) Variations in characteristics due to changes in
heater voltage.
(d) The effects of circuit loading.
The effects of different values of series grid resistor
have been ignored.
48
X G5-500
gm XGS-500 915
An
o>n
~p
l l l~ l
p3 HEATING AND COOLING CHARACTERISTICS
~~
0
n~
a
Io(av)=0•SA
N HEATING COOLING
0
Io (OV)'~
HEATING AND COOLING
CHARACTERISTICS
O 0
_4
O ~v
O 10 20 30 40 O 10 20 30 Time (mi~utesZ
TYPICAL EXAMPLE OF THE APPLICATION OF Upon switching on the filament again the heating-up
HEATING AND COOLING CHARACTERISTICS time indicated on the heating characteristic for an
excess temperature of 5.2° (i.e. 4.5 mins.) may be
The minimum permissible condensed mercury
deducted from the total heating-up time of 11 mins.
temperature for XGS-500 is 35°C and, assuming an
originally required for an excess temperature of 15°
ambient temperature of 20°C, it is necessary therefore,
to give a reheating time of 6.5 mins. before applying
to wait until the excess mercury temperature has risen
the h.t. supplies again.
by 15° before applying the h.t. supplies.
From the heating characteristic it will be seen that In each case the time required for re-heating exceeds
this entails a delay of at least 11 mins. After applying the minimum cathode heating time of 5 seconds.
the h.t. the valve is operated at a mean anode current
of 500 mA for a time sufficient for the condensed When the normal operation of the valve is at a
mercury temperature to reach equilibrium conditions. lower peak current than the maximum rating, it
If, now, both h.t. and filament supplies are switched may be permissible to reduce considerably the
off for a period of 15 mins. it is observed from the preliminary valve heating time. If such a reduction
cooling characteristic that the excess temperature at is proposed, however, reference should first be made
the end of this period will be 5.2°. to Mullard Ltd.
49
_~97mm.
XGQ2-6400 a
max,-
(MT105) a
g 9~
32
B40 BASE
988
Max. average positive control-grid current for anode voltage more positive than —10 V (averaging
time 15 secs.) 250 mA
Max. peak positive control-grid current for anode voltage more positive than —10 V 1.0 A
50
LIMITING VALUES (continued) X G 612-6400
Max. control-grid resistor 100 k S2
Recommended min. control-grid resistor 10 k S2
Max. average positive shield-grid current for anode voltage more positive than —10 V (averaging
time 15 secs.) 500 mA
Max. peak positive shield-grid current for anode voltage more positive than —10 V 2~0 A
XGQ2-6400 I 989
0
>
4 Vi a OV
V
N
1
%/
O
Vg (Vl ZO -I O O 10
CONTROL CHARACTERISTIC
51
XRI-3200 a
(MT5544) 0
g
B4D BASF
52
LIMITING VALUES (continued) X R1-3200
Max. average positive control-grid current for anode voltage more positive than —10 V (averaging
time one cycle) 200 mA
Max. peak positive control-grid current for anode voltage more positive than —10 V 2.5 A
Max. peak positive control-grid current for anode voltage more negative than —10 V 25 mA
nnn m~n~ nm ~
0
00 ,~~r ~
°o
uwi~nn%~
~,~ r ~wi u
0
~~ i~
~i~~/~
a
N
53
XRI-6400 a
67mmy ,
max.
14.6mm.
~+ a~"- ~ 112.7mtn.
(MT5545 ) nnn.
g
0
E
Eo Ex
=E 0
N NE
N
B4D BASE
f f
997
54
LIMITING VALUES (continued) X R I -6400
Max. average positive control-grid current for anode voltage more positive than —10 V (averaging
time one cycle) 200 mA
Max. peak positive control-grid current for anode voltage more positive than —10 V 2.5 A
Max. peak positive control-grid current for anode voltage more negative than —10 V 25 mA
I9 (mA1 50,
XRI^6400 995 XR - 6400 998
O
O_
•o O
- Ia=O.64A
O
v O•eA
O i•6A
O O
in
O 3.2A
O
4 C'1
m
4.8A
0
g 6.4A
O
O
N e•OA
O
O
f~l
55
4
Printed by Wightman & Co. Ltd., 179, Tottenham Court Road, London, England.