Note : Les descriptions sont présentées dans la langue officielle dans laquelle elles ont été soumises.
This invention relates to spark sources for producing
electrical sparks for any desired purpose, particularly to ~ -
produce light for emission spectroscopy.
In electrical spark emission spectroscopy, electrical
sparks are produced across a spark gap between two spark elec-
trodes. A material to be analyzed is introduced into the sparks
so that the material is caused to emit light having a spectrum
which is characteristic of that material. The light is then
analyzed by the usual procedures of emission spectroscopy. ~ ~-
The material may be in the gaseous, liquid or solid
state when introduced into the sparks. When the material is a
solid, it is generally placed on or made the material of one or -
both of the electrodes. The material is then eroded and vapor- ~ ~
ized by the sparks so as to produce its characteristic spectrum. ~-
One principal object of the present invention is to
provide a spark source capable of producing a train or series
of repetitive sparks which are regulated in magnitude or energy
in a naw and improved manner, so that the magnitude of the
sparks can be kept constant or otherwise regulated.
It has been found that it is desirable to keep the
magnitude of the sparks constant in order to carry out spectro-
scopic analyses which are repeatable and quantitatively accurate ;
to a high degree. If the mangitude of the sparks is kept con-
stant, there will be a definite quantitative relationship bet-
ween the concentration of any particular constituent of the
material being analyzed and the amount of light in the spectral
lines produced by such constituent. With the control over the
magnitude of the sparks afforded by the present invention, ;
spectroscopic analyses can be carried out to determlne not
. .: .
only the identity of the various constituents of the material
being analyzed, but also the concentration of each constituent
with an improved degree of accuracy~ - -
.:
--1--
~ ,',
. , .
(3~6
The p~esent invention in Various aspe~ts is an improv~-
ment over the invention disclosed and claimed in the copending
Canadian patent application Serial No. 220,578, filed Feb. 21,
1975, and entitled "Spark Sources With Electronic Switching
Tubes". The present invention is applicable to spark sources
of the general construction disclosed in such copending appli-
cation. In such a spark source, the sparks are produced across
a spark gap between two electrodes, by discharging a capacitor
across the spark gap~ The production of each spark is con- -
trolled by the triggering of a thyratron or some other elec-
tronic switching device, connected into a discharge circuit
between the capacitor and the spark gap. The provision of ~-
the electronic switching device makes it possible to control
the timing of each spark with a high degree of precision.
After each spark, the capacitor is rechar~ed to a hi~h voltage
by a charging circuit comprising a hi~h voltage transformer
and a rectifier.
In accordance with the present invention, it has been
found that the magnitude of the individual sparks in a -train or
series of sparks can be regulated and controlled by varying
the timing of ~e individual sparksin the train. The variation
in the timing of each spark affects the voltage to which the
capacitor is charged for the next spark. By varying the timing
of the individual sparks, it is p~ssible-to charge the capacitor
t~ the same voltage for all of the sparks in the series or train.
When this is done, all of the sparks will be the same in magnitude
or energy. Moreoverr the waveform of the spark discharge current
will be the same for all of the sparks. Accordingly, the spec
tral characteristics of the light emitted by all the sparks
~- .
will be the same, b~th as to magnitude and detailed structure.
In some cases, it may be desired to vary the magnitude of
the individual sparks in a predete~mined m~nner. This can also be - ~-
-2- ~
accomplished by varying the timing of the sparks.
I-t has been found that if the timing of the sparks is
not varied, so that the repetition rate of the sparks is kept
constant, the magnitude of the sparks will not be constant, but
will vary throughout the cycle of the alternating current which
is employed to energize the high voltage transformer.
In order to keep the magnitude of the sparks constant,
the timing of the sparks may be varied under the control of a
computer, which may be programmed to vary the timing of each
spark in accordance with the instantaneous timing of the recti-
fied alternating current supplied by the transformer secondary
and the rectifier.
As disclosed in the previously mentioned Walters and
Bernier application, the thyratron employed as the electronic
switching device is preferably shunted by a reverse connected
diode for carrying the reversely polarized half-cycles of
the oscillatory spark discharge current in the discharge cir-
cuit. To improve the commutation between the thyratron and
the shunting diode, inductance may be inserted in series with
the thyratron. Such inductance may be sufficiently great in
magnitùde to sustain the thyratron current so as to prevent
deionization of the thyratron during the half-cycles~when the
9hunting diode is conductive. This construction minimizes or
obviates the production of voltage spikes and other transients
across the thyratron, and also minimizes or obviates irregu-
larities in the total time span of the spark discharges.
j Thus, in accordance with one aspect of the invention,
there is provided a spark source, comprising a high voltage
transformer having a secondary winding for supplying alter-
nating current at a high voltage, a storage capacitor, a charg-
. .
ing circuit including a rectifier connected between saidsecondary winding and said capacitor for charging said capacitor,
lC~9Z&j
spark gap electrodes having a spark gap therebetween, a dis-
charge circuit including an electronic switching device connect-
ed between said capacitor and said spark gap electrodes for
discharging said capacitor across said spark gap, said electron-
ic switching device having input means for receiving triggering
signals, and control means for supplying a sequence of variably
spaced triggering pulses to said input means for producing a
sequence of spark cycles in which said capacitor is charged to
substantially equal voltages for all of said spark cycles, said :
control means including variable timing means for producing
each triggering pulse at a time when said capacitor has been
charged to a particular voltage.
In accordance with another aspect of the invention
there is provided a spark source, comprising a storage capacitor,
charging means for charging said capacitor, spark gap electrodes
having a spark gap therebetween, a discharge circuit including
inductance means and an electronic switching device connected
in series with said capacitor and said spark gap, said electron- `. -
ic switching device being operable between a nonconductive ...
state and a unidirectionally conductive state to discharge said .~ :
capacitor through said inductance means and across said spark ~ ~
gap, a shunting diode connected in parallel with said elec- .
tronic switching device and polarized to be unidirectionally
conductive in a direction opposite ko the conductive direction
of said electronic switching device, said capacitor and said
lnductance means producing an oscillatory discharge current
which is conducted in one direction by said electronic switching .
device and in the opposite direction by said shunting diode,
and second inductance means in series with said electronic . ; ~
switching device for improving the commutation between said . ~ :
electronic switching device and said shunting diode. .
According to another aspect of the invention there is
-3a- ~:
provided a spark source, comprising a high vol~age transformer
having a primary winding and a high voltage secondary winding,
means for supplying alternating current to said primary
winding, a storage capacitor, a charging circuit including a
rectifier connected between said secondary winding and said
capacitor for supplying high voltage rectified pulses to said
capacitor for charging said capacitor, spark gap electrodes
having a spark gap therebetween, a discharge circuit including
an electronic switching device connected between said capacitor
and said spark gap electrodes for discharging said capacitor
across said spark gap, said electronic switching device having
input means for receiving triggering signals, and control
means for supplying a se~uence of triggering pulses to said
input means for producing a sequence of spark cycles in which
said capacitor is charged by said charging circuit and then is
discharged through said electronic switching device and across
said spark gap, said control means including variable timing
means for varying the timing of each of said triggering pulses
with respect to the phase angle of the alternating current to
vary the voltage to which said capacitor is charged for each
spark discharge.
Further objects, advantages and features of the
present inventi.on will appear from the following description,
taken with the accompanying drawings, in which:
Figure 1 is a schematic circuit diagram of a spark
source to be described as an illustrative embodiment of the
present invention.
Figure 2 is a schematic diagram showing the timing
control ~or the spark source of Figure 1.
Figure 3 is a fragmentary schematic circuit diagram
showing a modified construction utilizing a half-wave rectifier
rather than a full-wave rectifier, as in Figure 1.
-3b-
- .. .. : .
-
FIG. 4 iS a schematic circuit diagram showing a modi-
fied construction in which inductance is introduced in series
with the thyratron of Fig. 1.
FIG. 5 is a series of wave~orm diagrams showing the
ef~ects of changes in the spark timing upon the magnitude of
the voltage tv which the capacitor is charged for each spark.
FIG. 6 is a series o~ waveform diagrams showing the
charging of the capacitor for different values of the spark `~
timing, the capacitance of the capacitor, and the inductance
in the charging circuit.
FIG. 7 is a series oE wave~orm diagrams showing the
capacitor charginy voltage and the capacitor charging current
for different values of capacitance and resistance in the
charging circuit. -`
FIG. 8 is a series of waveform diagrams showing the
capacitor charging current and voltage ~or different values
of inductance in -the charging circuit.
FIG. 9 is a series of waveform diagrams similar to
those of Fig. 8, but for a different value o~ capacitance. ;
FIG. 10 is a wave~orm diagram representing the oscil-
latory discharge current in the capacitor discharge circuit.
FIG. 11 is a waveform diagram representing the modi- ; -
~ied discharge current across the spark gap.
FIG. 12 is a wave~orm diagram similar to Fig. 11 r . ~
but for a different value of capacitance. ~` "" `
FIGS. 13 and 14 are schematic circuit diagrams showing
modi~ied timing control means.
Fig. 1 illustrates a spark source 10 o~ the general
construction disclosed and claimed in the previously mentioned
Walters and Bernier application. In the spark source 10, the
electrical sparks are produced across a spark gap G between a -
grounded electrode 12 and an ungrounded electrode 14. The
--4--
sparks are produced by discharging a high voltage capacitor C
across a spark gap G. While -the capacitor C may comprise a
single capacitor unit, if desired, the illustrated construction
employs a plurality of capacitor units Cl, C2, C3, C4, c5, C6,
C7 and C8 adapted to be connected in parallel on a selective
basis by corresponding switches Sl-S8. By operating the
switches, any or all of the capacitor units cl-c8 can be con-
nected in parallel, so that the total capacitance of the
capacitor C can be varied.
The capacitor C is adapted to be charged by a charging
circuit 16 comprising a high voltage transformer Tl, a diode
ractifier Dl and a series resistor Rl. The transformer Tl ~:
comprises a low voltage primary winding TlP and a high voltage
secondary winding TlS.
The diode rectifier Dl may be of the full-wave type ~-
or the half-wave type, and may comprise one or more diode
units. As shown, the diode rectifier Dl is of the ~ull-wave
type, utilizing four diode stacks DlA, DlB, DlC and DlC in a
bridge circuit. Each diode stack comprises a plurality of
diode units DlU connected in series. In each stack, the diode
units DlU are shunted by high value resistors DlR which pro-
vide a voltage divider for equalizing the voltage dlstribution-
along the diode stack.
While the resistor Rl may comprise a single resistor
unit, the illustrated construction utilizes two resistor units
RlA and RlB connected in parallel, to double the wattage
rating.
In the charging circuit 16, an inductance LT is shown ~;
in series with the secondary winding TlS of the high voltage
transformer Tl. This inductance LT represents the effective
inductance in series with the charging circuit, such effective
inductance being produced by the leakage inductance of the
.
-5-
transformer Tl and an~ addltional or ~umped inductance in-
serted into the chargin~ circuit. Typically, no additional
lumped inductance is provided, so that the inductance LT
represents the e~fective inductance produced by the trans-
former Tl. Such effective inductance is generally the
primary inductance multiplied by the square of the turns
ratio of the transformer Tl.
The primary winding TlP of the high ~oltage trans-
former Tl is adapted to be energized from a pair of alter-
nating current power lines 18 which may supply alternatingcurrent at 240 volts and 60 Hz or any other suitable volt-
age and frequency. While the primary winding TlP could be ;
connected directly to the power lines 18, the illustrated
construction provides a primary circuit 20 which makes it
possible to control and vary the voltage applied to the
primary TlP. As shown, a voltmeter Ml is connected across
the primary winding TlP to measure the primary voltage,
while an ammeter M2 is connected in series with the primary
winding to measure the primary current. The opposite sides
of the primary winding TlP are connected to a pair of two-
.............. .
position selector switches 22A and s, whereby the primary
winding can be connected to either the output terminals of
a variable autotransformer T2 or the secondary winding of a
transformer T3. The input connections of the variable auto-
transEormer T2 and the primary winding of the transformer T3
a~e connected in parallel to leads 24A and B. It will be
seen that a switch 26 is connected between the lead 24B and
one of the power lines 18. The other power line is connected
to a two-position selector switch 28. In one position, the
switch 28 establishes a direct connection between the power
line and the lead 24A. In the other position, the switch 28
establishes a connection through a bank of resistors Rl-R8,
-6-
connected in parallel to pro~ide the desired wattage rating.
It will be unders~ood that the inser~ion of the resistors
Rl-R8 reduces the voltage supplied to the primary winding
TlP of the high voltage transformer Tl.
The spark source 10 of Fig. 1 includes a discharge
circuit 30 which includes the capacitor C and the spark gap
G, for discharging the capacitor across the spark gap. The
discharge circuit 30 includes an electronic switching device,
illustrated as a thyratron gaseous discharge tube 32. In -~
the illustrated circuit, the thyratron 32 is polarized with
its anode connected to one side of the capacitor C, and with
its cathode connected to ground, and thus to the grounded
electrode 12 of the spark gap G. The thyratron 32 could be
connected elsewhere in the discharge circuit 30, with the
polarization of the thyratron reversed, but it is advantage- ~`
ous to connect the thyratron 32 so that its cathode is
grounded. As shown, the heater of the thyratron 32 is
energized from the secondary of a transformer T4, having
its primary connected to the alternating current power lines
18 through a switch 34. The control electrode of the thyra-
tron 32 is adapted to receive triggering pulses to cause the
thyratron to become conductive, as will be described in
greater detail presently.
In the illustrated spark source 10, the discharge
eireuit 30 includes induetors or induetance elements Ll and
L2, connected in series between the high voltage side of the
capacitor C and the ungrounded electrode 14 of the spark gap G.
The inductance elements Ll and I.2 are preferably variable or
adjustable,~ as shown, but may also be fixedO A junc ion lead
36 is provided between the inductance elements L1 and L2.
As described in the previously mentioned Walters and
Bernier application, a shunting diode D2 is preferably con~
-7- -
:`'
~ ~r~2~
nected in a shunting path 38 which extends between the ~unction
lead 36 and ground. Thus, the shunting path 38 is arranged to
shunt the series combination of the spark gap G and the induct-
ance element L2. The shunting path 38 preferably includes a ;
third inductance element or inductor L3, connected in series
with the shunting diode D2. The induc~ance element L3 is
preferably variable or adjustable, as shown, but may also be ~ -
fixed.
The illustrated shunting path includes a pair of two-
position selector switches 40A and s for reversing the polarityof the shunting diode D2 . When the switches 40A and s are in
the positions shown in Fig. 1, the diode D2 is polarized to
conduct current between ground and the junction lead 36. When
the switches 40~ and B are in their other positions, the diode
D2 is polarized to conduct curren~ between the junction lead
36 and ground. The shunting diode D2 could possibly include
only a single diode unit having a sufficiently high inverse
voltage rating, but it is preferred to employ a diode stack
having a plurality of diode units D2U in series. The diode
units D2U are preferably shunted by corresponding high value
resistors D2R to provide a voltage divider for evenly distri-
buting the inverse voltage along the diode stack D2. ~Shunting
capacitors D2C are also provided across the respective diode
units D2U to provide a similar voltage dividing action.
Due to the presence of capacitance, conductance and
resistance in the discharge circuit 30, the capacitor dis-
charge current ls of a damped oscillatory character. Fig. 10
shows a damped oscillatory waveform which is a typical re-
presentation of the capacitor discharge current. It will be
understood that the thyratron 32 acts as a rectifier and thus
conducts the oscillatory discharge current in one direction
only, during the half-cycles of the discharge current which
-8-
~ t3~ ~
are polarized to cause current to flow between the anode and
the cathode of the thyratron 32. To carry the discharge
current during the reverse hal~-cycles, the thyratron 32 is
preferably shunted by a shunting diode D3, polarized to con-
duct current between ground and the capacitor C. The diode
D3 could possibly comprise only a single diode unit having a
suf~iciently high inverse ~oltage rating, but it is preferred
to employ the illustrated diode stack, comprising a plurality
of diode units D3U connected in series. Shunting resistors
D3R of high value are preferably connected across the respect-
ive diode units D3U to provide a voltage divider which will
equalize the distribution of the inverse voltage along the
diode stack D3. Shunting capacitors D3C of small value are
also preferably connected across the diode units D3U to pro-
vide a similar voltage dividing action.
As shown in Fig. 1, a protective resistor R9 is
preferably connected in series with the control electrode
of the thyratron 32. A shunting resistor R10 is preferably
connected between the spark gap electrodes 12 and 14 to dis-
charge the capacitor C completely when the spark source 10is shut downl and to prevent premature ignition of the spark
across the spark gap, before the thyratron 32 is triggered
into a conductive state. Such premature ignition o the ;~
spark may be produced due to distributed capacitance between
the anode of the thyratron 32 and groundl as explained in the ~`
previously mentioned Walters and Bernier application.
As previously indicated, the thyratron 32 is triggered
into condition by pulses or other suitable triggering signals
supplied to the control electrode of the thyratron 32 through
the resistor R9. Thus, the thyratron 32 acts as an electronic
switching device. When the th~ratron 32 becomes conductive,
the capacitor C is discharged through the discharge circuit 30.
, . ,
_g_ : :
The discharge o~ the capacitor C produc~s a spark across the
spark gap G. Due to the presence of capacitance, inductance
and reSistance in the discharge circuit 30, the discharge cur- -
rent is oscillatory and has a damped oscillatory waveform, as
represented in a general way in Fig. 10. The thyratron 32
conducts the capacitor discharge current ~uring the half-
cycles of one polarity, while the shunting diode D3 conducts
the discharge current during the half-cycles of the opposite
polarity.
During the oscillatory discharge of the capacitor C,
the waveform of the spark current across the spark gap G is
modified by the presence of the shunting diode D2, as ex-
plained in the previously mentioned Walters and Bernier appli- ;
cation, and also in the Walters patent No. 3,749,975, issued
July 31, 19730 The shunting diode D2 is conductive in one
direction only. ~he effect of the shunting diode D2 upon the
spark current depends upon the polarity of the diode D2. As
shown in Fig. 1, the shunting diode D2 is polarized so as to
be conductive during the first half-cycle of the capacitor
discharge current. If the inductance of L3 is very low,
approaching zero, and if the impedance of the shunting diode
D2 is sufficiently low, the spark gap G may not even be broken
down into conduction during the first half-cycle of the capa-
citor discharge current. If the impedance of the shunting
diode D2 is higher, or if L3 has a sufficient inductance, the
spark gap G will be broken down into conduction during the ~ -
irst half-cycle. It is generally preferred to keep the
impedance of D2 and the inductance of L3 to a minimum, so
that little or no current will flow across the sp~rk gap G
during the first half-cycle of the capacitor discharge current.
During the second half-cycle of the capacitor dis-
charge current, the shunting diode D2 becomes nonconductive,
-10
'', ,' ' '
39~
due to the reversal of the current, so thak the spark gap G
breaks down and becomes conductive. The spark current flows
through the induc-tance L2, so that energy is stored in the
magnetic field which is produced by the current in the induct-
ance L2.
During the third half-cycle, the shunting diode D2
again becomes conductive, but the conducti~n across the spark
gap G is also maintained, with the direction of the spark cur-
rent unchanged, by the inductance L2. The energy stored in
the magnetic field of the inductance L2 is released as the
magnetic field gradually collapses, with the result that a
voltage is induced in L2 so as to sustain the spark current
across the spark gap G. It is generally preferred to make
the inductance of L2 sufficiently high to sustain the spark
current across the spark gap G for the remainder of the damped
oscillatory discharge current of the capacitor C. Thus, the
spark current has a unidirectional pulsating waveform during
the second and subsequent half-cycles of the capacitor dis~
charge current.
The waveform of the spark current is represented by
Figs. 11 and 12, which show that the spar]c current is uni-
dlrectional and pulsating during the second and subsequent
half-cycles. In the situation represented by Figs. 11 and 12,
there is a reverse spark current during -the first half-cycle,
but this reverse current can be eliminated by ma~ing the
impedance of the shunting diode D2 and the inductance of L3
sufficiently low.
The waveEorms of Figs. 11 and 12 are similar, but
specifically different, because the waveform of Fig. 12 was
o~tained with a higher value of the inductance Ll than was
employed to produce the waveform of Fig. 11. Thus, Fig. 12
represents a condition in which Ll had a value of 150 micro-
--11--
henrys, while Fig. 11 represents a condition in which ~1 hada ~alue o~ 85 microhenrys.
~ s explained in the previously mentioned Walters and
sernier application, the unidirectional spark current during
the second and subsequent half-cycles is advantageous, because
the material to be analyzed is eroded and vaporized almost
entirely from the grounded spark gap electrode 12, which
functions as the cathode during the second and subsequent
half-cyles.
The frequency of the oscillatory capacitor discharge
current is inversely related to the values of the capacitance
and the inductance in the capacitor discharge circuit. ~peci-
fically, the oscillatory frequency is inversely proportional
to the square root of the inductance. Thus, as will be evi-
dent from a comparison of Fi~s. 11 and 12~ the oscillatory
frequency is greater for the condition represented in Fig. 11,
in which the inductance of Ll is less than in the condition
represented by Fig. 12. To restate the same proposition in
another way, the period of the oscillatory current is less
20 for the condition of Fig. 11, in which the inductance is less. `
As will be evident from Fig. 10, the oscillatory capa- `
citor discharge current is damped, because the energy initially ;
stored in the capacitor C is dissipated by the effective re-
sistance of the spark across the spark gap G, and also by the
.. . .
resistances of the inductances and conductors in the discharge
circuit. Thus, the oscillatory capacitor discharge current
dies out after a brief period of time, comprising a number of
cycles of the oscillatory capacitor discharge current. Thus,
the spark gap G becomes nonconductive. The thyratron 32 also ;~
becomes nonconductive because the triggering pulse supplied
to its control electrode is terminated.
As sho~m in Fig. 1, the trlggering pulses are supplied
.
-12-
. ~:
to the control electrode o~ the thyratron 32 by a timing control
device or means 44, which supplies the pulses between the con-
trol electrode and the grounded cathode of the thyratron. As
will be explained in greater detail presently, the timing of
the pulses supplied by the timing control 4~ is related to the
phase angle of the alternating current supplied by the alter-
nating current power lines 18. Thus, in the construction of
Fig. 1, the alternating current power lines 18 are connected
to the timing control 44 through a switch 46.
When the capacitor discharge current dies out and the -
spark gap ~ becomes nonconductive, the capacitor C is re- -
charged by the charging circuit 16, comprising the secondary
winding TlS of the high voltage transformer Tl, the diode
rectifier Dl and the series resistor Rl. The waveform of
the charging current and the rapidity with which the capacitor `
C is recharged are determined by the volta~e developed in the ~-
secondary winding TlS, the effective charging inductance LT,
and the charging circuit resistance afforded primarily by Rl, ~
but also by the diode rectifier Dl and the secondary winding ;
TlS. For simplicity of discussion, Rl may be regarded as
incorporating not only its own resistance, but the resistances
contributed by the diode rectifier Dl and the transformer
secondary TlS.
It has been found that the timing of the sparks has a
surprisingly great effect upon the magnitude of the voltage to
which the storage capacitor C is charged between sparks. The
timing of the sparks can be varied by changing the tLming of
the triggering pulses supplied to the control electrode of the
thyratron 32 by the timing control means 44
The effects of changes in the spark timing are graphic-
ally illustrated in Fig. 5. Figs. 5A and B represent situations ;
in which the spark triggering pulses are uniformly or equally
-13- ~ -
: . .,:
a3~
spaced. Specifically, Figs. 5A and B represent situations in
which the spark triggerin~ pulses are supplied at a repetition
rate of 1000 pulses per second, so that the pulses are timed
at intervals of 1 millisecond. Such evenly spaced pulses are
represented at 50 in Figs. 5A and B.
It will be understood that the high voltage transformer
secondary TlS and the rectifier Dl supply rectified alter- -
nating current pulses, corresponding to half portions of a
sine wave, as represented at 52 in Figs. 5A-F. The half-period
of the 60 cycle waveform amounts to about ~.3 milliseconds, so
that each of the triggering pulses 50 is in a different relat-
ionship to the phase angle of the half-sinusoidal pulse 52. ~ ~ -
Each triggering pulse 50 initiates the discharge of
the capacitor C, which produces a spark across the spark gap -~
G. The damped oscillatory capacitor discharge current, as
represented by Fig. 10, dies out after a small fraction of a
millisecond. The spark gap G then becomes nonconductive,
with the result that the recharging of the capacitor C is ~ ;
commenced by the high voltage transformer Tl, acting through
the rectifier Dl and the resistor Rl. The rate at which the
capacitor C is recharged is directly related to the instant-
aneous voltage produced by the secondary winding TlS, and is
inversely related to the inductance represented by LT, the
resistance represented by Rl and the capacitance value of
the capacitor C. The charging of the capacitor C continues
until the next spark is triggered, or until the capacitor
charging current attempts to reverse its direction, such
reversal being prevented by the diode rectifier Dl, so that
the charge is retained on the capacitor C. Due to the
presence o both inductance and capacitance in the charging
circuit, the capacitor charging current tends to be oscil-
latory, so that the current rises to a peak, decreases to
-14-
.
';
39 ~6
zero and tends to reVerse, at which point the diode Dl prevenks
any such reversal. Due to the presence of both inductance and
capacitance in the charging circuit, the capacitor charging
voltage can rise to a value which is greater than the peak
voltage developed by the secondary winding TlS of the high :
voltage transformer T l .
In Fig. 5, the waveform of the capacitor charging
voltage is represented by the series of pulses 54, each of
which rises gradually-to a peak and then drops substantially
to zero when the next spark is ~riggered by the next trigger~
ing.pulse 50. For the situations represented by Figs. 5A and
5B, in which the triggering pulses 50 are evenly spaced at .
one millisecond intervals, the capacitor charging voltage
pulses 54 are of unequal magnitude, due to the fact that each
pulse 54 has a different relationship to the half-sinusoidal
voltage waveform 52 developed by the secondary winding TlS, ~ :
acting through the rectifier Dl. The magnitude of the capa- ..
citor charging voltage pulses 54 depends directly upon the
instantaneous voltage represe~ted by the half-sinusoidal
waveform 52 during the capacitor charging time span.
The capacitor charging voltage pulses 54 of Fig. 5A .
were developed with a capacitor value of 0.0055 microfarad. .
It will be seen that most of the capacitor voltage pulses 54 .. ~
rise to a peak voltage which is greater than the instantan- .:
eous voltage developed during the charging time span by the :~
transformer secondary TlS, as represented by the half
sinusoidal waveform 52. .
In Fig. 5B, the capacitor charging voltage pulses 54
. ,
were produced with a higher capiacitor value of 0.0137 micro- .~
30 ~arad, with the result that the voltage pulses 54 rose less ::;;
rapidly to lower peak values than in the case of Fig. 5A. In
.. . .. .. .
both Figs. 5A and B, each capacitor charging voltage pulse 54 `.
~ .
-15-
.
has a different peak value, due to its uniquely different
relationship to the half-sinusoidal voltage waveforrn 52.
~ igs. 5C, D, E and F illustrate the manner in which
the spark timing may be varied, in accordance with the
present invention, to cause the capacitor charging v~ltage
pulses 54 to be constant or uniform in peak magnitude. It
will be seen that in the waveform diagrams of Figs. 5C-F,
the timing of the spark triggering pulses 50 is nonuniform,
so that the pulses 52 occur at unequal or variable intervals.
These intervals are chosen or compu-ted so that the capacitor
voltage pulses 54 have the same magnitude, for each of the
set of conditions represented by the individual waveform
diagrams of Figs. 5C-F.
The waveforms o~ Fig. 5C were produced with a storage
capacitor C having a value of 0.0055 microfaradl the same as -
for Fig. 5A. However, the timing intervals of the tri~gering
pulses 50 were varied to cause the peak voltage of the capa-
citor charging pulses ~4 to be held constant at 5,000 volts.
This was achieved by providing relatively large timing inter-
vals between the triggering pulses 50 for the early and late
portions of the half-sinusoidal transformer voltage waveform
52, while providing smaller time intervals between the trigger-
ing pulses 50 for intermediate portions of the half-sinusoidal
waveform 52, when the instantaneous value of the half-
sinusoidal waveform was at or near its maximum. Thus, more
time was allowed for the capacitor C to charge when the
instantaneous value of the rectified transformer voltage 52
was low, than when such rectified transformer voltage was
high, so that the capacitor C was charged to the same voltage
in all cases. Thus, the same energy was stored in the capa-
citor C to produce each of the sparks in the spark train.
In the case represented by Fig. 5D, the conditions
'.- ~': -.
-16-
' ~ '
. . . . . . . . .
- ~3~ 26
remained the same as in Fig. 5C, except that the timing of
the spark triggering pulses 50 was changed so as t~ raise -the
peak value of the capacitor charging pulses 54 to 8, 000 volts .
To achieve this higher voltage, the timin~ intervals between
the triggering pulses 50 were increased, so that more time
was allowed to charge the capacitor C in each instance. The
time intervals were different from spark to spark, to allow
for the different instantaneous values of the half-sinusoidal
waveform 52, during the successive charging intervals of the
capacitor C.
The waveform diagrams of Fig. 5E represent a condition
in which the value of the capacitor C was increased to 0.0137 ~ -
microfarad. The timing of the spark triggering pulses 50 was
varied to maintain the peaks of the capacitor charging pulses
54 at 5,000 volts, as in the case represented in Fig. 5C.
Because of the increased capacitance in the charging circuit,
the intervals between the triggering pulses 50 were increased
to allow more time for the capacitor to charge to the desired
5,000 volts. The individual intervals were adjusted to allow
Eor the different instantaneous values of the half-sinusoidal
transformer voltage waveform 52 during the successive intervals.
In the case represented by Fig. 5F, ~he capacitor value
was the same as in the case of Fig. 5E, but the spark timing
intervals were increased to raise the peak capacitor voltage
to 8,000 volts. The timing of the individual intervals was
~aried to allow for the differenk instantaneous values of the
half-sinusoidal waveform 52.
It will be evident from Figs. 5C F that the peak volt~
..:
age to which the capacitor C is charged for each spark can be
held constant by varying the timing of the sparks. The volt~
. ", . .: ~ .
age to which the capacitor is charged can be raised by increas-
ing the timing intervals between the sparks, so as to allow
-17-
,
more time ~or the capacitor to charge. Conversely, the peak
capacitor voltage can be decreased by decreasing the timing
inter~als between the sparks. If the value o~ the capacitor
is increased, the timing intervals must be increased to main-
tain the same capacitor peak voltage, because more time will
be required to charge the capacitor to the same voltage. The
individual timing intervals between the sparks are adjusted -
to allow for the different instantaneous values of the half- -
sinusoidal voltage developed by the transformer.
It is usually desirable to charge the capacitor C to
the same peak voltage for each spark in the spark train, as
represented by the waveform diagrams of Figs. 5C-F. For any
particular value of capacitance, the energy stored in the
capacitor C is a function of the voltage to which the capa-
citor is charged. Thus, if the capacitor is charged to the
same voltage for each spark, the energy stored in the capa
citor is the same, and the energy dissipated in each spark
is the same. Accordingly, the total amount of light produced
by each spark is the same. Moreover, the damped oscillatory ~ ,t
capacitor discharge current, as represented by Fig. 10, has
the same maximum amplitude and waveform for each spark. It
follows that the maximum intensity of the light produced by
each spark is the same. Moreover, the detailed structure of -~
the spectrum produced by each spark is the same, because such ~ -
detailed structure is determined by the maximum amplitude and
waveform of the spark current, both of which stay the same,
as represented, for example, by Figs. 11 and 12O
Thus, the charging of the capacitor C to the same
voltage for each spark makes it possible to achieve results
which are fully~repeatable and accurate.
Accordingly, with the present in~ention, spectral
analyses can be carried out with a high degree of repeat-
-18- ;
3~
ability and accuracy.
Fig. 2 illustrates details of the timing control means
44 of Fig. 1~ It will be seen that the timing control means 44
of Fig. 2 may utilize a computer 60, which may comprise a suit-
ably programmed general purpose computer, or a special computex
programmed solely to control the timing of the spark triggering
pulses 50. The timing of the pulses is related to the instant-
aneous phase angle of the half-sinusoidal rectified transformer
voltaye 52 of Fig. 5. Accordingly, information with regard to
10 such phase angle is fed into the computer 60. In this case, ;-
such information is developed by a zero-crossing detector 62,
which receives an alternating current input from the power
lines 18, through a switch 64. The zero-crossing detector 62 -
supplies a timing signal to the computer 60 whenever the alter-
nating current waveform goes through zero. The computer 60
produces trigger control pulses at programmed or computed
intervals after each zero-crossing signal. The interval
be-tween each pair of successive spark triggering pulses is
programmed or computed to charge the capacitor C to the
desired voltage, considering the instankaneous voltage which
exists in the alternating current waveform during such inter- ;~
val. It is not necessary to feed the half-sinusoidal wave-
form 52, representing the rectified transformer secondary
voltage, as shown in Fig. 5, directly to the computer 60,
because the instantaneous magnitude of such waveform is a
simple sine function of the phase angle, so that such instant-
aneous amplitude can readily be computed by the computer 60.
The computer 60 can then compute the charging time which is
needed to charge the capacitor C to the desired voltage,
considering the existing values of the capacitor C, the
inductance LT and the resistance Rl in the charging circuit.
When the desired time interval has elapsed, the computer ;
-19- ~
: '
~ ~3~
generates a control pulse to trigger the next spark. To
assist the computer 60 in carrying out its timing functions,
the computer may be connected to the output of a clock pulse
generator 661 as shown in Fig. 2. It will be understood that
the clock pulse generator 66 may be incorporated into the
computer 60, if desired.
The trigger control pulses from the computer 60 ~re
typically of small amplitude. As shown in Fig. 2, such
pulses are fed to a triggerpulse generator 68 which produces
corresponding spark txiggering pulses of sufficient amplitude
to trigger the thyratron 32 o Fig. 1. These pulses are fed
to the control electrode of the thyratron 32 through the
protective resistor R9.
Fig. 3 illustrates a modified construction which
employs a half-wave rectifier circuit 70, instead of the
full-wave bridge rectifier circuit of Fig. 1. As shown, the ;~
half-wave rectifier circuit 70 comprises only a single diode
stack DlA. The other three diode stacks of Fig. 1 are
omitted. As before, the diode stack DlA comprises a plural- ~ -
ity of diode units DlU connected in series, with individual
shunting resistors DlR of high value to equalize the distri-
bution of the inverse voltage along the stack DlA. All of
the previous discussion is applicable to the half-wave recti-
~ier 70 o~ Fig. 3, except that the charging voltage is avail-
able during only alternate half-cycles of the alternating
voltage ~rom the high voltage transformer secondary TlS.
Fig. 4 illustrates a modified electronic switching ~ `
circuit 72, in which an additional inductance unit ~4 is -~
connected in series with the anode-cathode path of the
3~ thyratron 32. Otherwiser the electronic switching circuit
72 is the same as illustrated in Fig. 1. The shunting diode
rectifier D3 is connected across the thyratron 32 in the same
-20-
-
.. . .. . ..
.3~3~
manner as described and illustrated in Fig. 1.
The provision of the addi~ional inductance el~ment L4
in Fig. 4 is advanta~eous because it improves the commutation
between the thyratron 32 and the shunting diode D3. I~ Will
be recalled that the thyratron 32 carries the capacitor dis-
charge current in one direction, while the shunting diode D3 ~ -
carries such current in the opposite direction. Thus, the
thyratron 32 and the shunting diode D3 carry the oscillatory
capacitor discharge current of Fig. 10 during alternate half-
cycles, so that such current is commutated between the thyra~
tron 32 and the shunting diode D3 at the transitions between
such hal~-cycles. The inductance element L4 preferably has
sufficient inductance to maintain a small current through the
thyratron 32 during the half-cycles in which the shunting ~-
diode D3 is conductive. By maintaining such small current,
the deionization of the thyratron 32 is prevented, so that
the thyratron 32 is ready to conduct the main capacitor dis-
charge current when such current reverses so that the shunting
diode D3 is no longer conductive.
It may be helpful to offer a more detailed explanation
of the action of the inductance element L4~ During the half-
cycles when the thyratron 32 carries the main capacitor dis-
charge current, such current also passes through the induct-
ance L4 and builds up a magnetic field around the inductance.
When the main capacitor discharge current reverses, it ~ -
switches to the shunting path through the shunting diode D3,
because the thyratron 32 will not carry the reverse current.
However, the magnetic field around the inductance L~ col-
lapses and induces a continuing forward voltage in the induct- ;
ance L4. Such forward voltage produces a continued forward
current between the anode and the cathode of the thyratron 32.
Such forward current circulates through the closed loop formed
3~
by the thyratron 32, the shunting diode D3 and the inductance
element L~. If the inductance element L4 is large enouyh in
inductance, the forward current through the thyratron 32
persists throughout the half-cycles when -the main capacitor
discharge current is being carried by the shunting diode D3.
ThuS, the thyratron 32 never deionizes, so that it is ionized
and ready to carry the main capacitor discharg~ current when
such current reverses so that it is switched from the shunting
diode D3, back to the thyratron 32.
With the modified construction of Fig. 4, the com-
mutation between the thyratron 32 and the shunting diode D3
is extremely smooth in both directions. In the absence of the
inductance element L4, a commutation spike tends to be produced ~ `
at the heginning of each half-cycle when the thyratron 32
carries the main capacitor discharge current, because a greater .
voltage is required across the thyratron to reionize the thyra- ::
tron than to maintain the ionization after it has become con-
ductive. The commutation spikes have the disadvantage of
tending to cause radio frequency interference in other elec-
tronic and electrical equipment which happens to be near the
spark source.
The provision of the inductance element L4 also has
the advantage that the commutation between the thyratron 32
and the shunting diode D3 is always consistent and repeatable,
eVen though the frequency of the damped oscillatory capacitor
digcharge current may vary over a wide range. There is no
ambiguity or ~rregularity in such commutation, so that the ~:
waveform and duration of the damped oscillatory capacitor
discharge current are free from any conse~uent ambiguities
or irregularities.
It has been found that, in the absence of the
inductance element L4, such ambiguities or irregularities
-22-
'
are noticeable when the frequency o~ the oscillatory capacitor
discharge current is such that the hal~-period o~ the oscil-
latory discharge current is o~ the same order of magnitude as
the deionization time of the thyratron 32. For higher fre-
quencies, when such half-period is substantially less than . :
the deionization time, the thyratron 32 remains ionized during
the half-cycles when the shunting diode D3 is carrying the
main oscillatory capacitor discharge current. Thus, the
thyratron 32 is still ionized when ~he main current switches
back to the thyratron 32, so that the commutation is smooth
and repeatable. For lower frequencies, when the half-period
of the capacitor discharge current is substantially greater
than the deioni~a-tion time of the thyratron 32, the thyratron
will always become deionized during the half-cycles when the
shunting diode ~3 is carrying the main capacitor discharge .. :
.
current. Thus, the thyratron 32 will always have to be re-
ionized when the capacitor discharge current reverses so that ...
it is switched back to the thyratron 32.
When the frequency of the oscillatory capacitor dis- .:.
charge current is such that its half-period is approximately
equal to the deionization time of the thyratron 32, an ambigu-
ous situation can axise in which the thyratron is sometimes ...
deionized and sometimes not deionized, during the half-cycles : :
when the shunting diode D3 is carrying the main capacitor :
discharge current. This ambiguous situation produces ambi- .. . ..
guities and irregularities in the waveform and the total
duration o~ the spark current, so that the spectroscopic - ~ .
results may not be fully repeatable. The ambiguities in
the total duration of the spark current also produce ambi-
guities in the voltage to which the capacitor is subse-
quently recharged, because the recharging does not commence
until the spark current dies out. Thus, any ambiguity in
: :
-23-
z~
the total duration of the spark current produces a correspond-
ing ambiguity in the time during which the capacitor is sub-
sequently recharged, assuming that the timing o~ the sparks
remains unchanged. The ambiguities in the charging time pro-
duce ambiguities in the charging voltage, which in turn pro-
duce ambiguities in the energy dissipated in the nex-t spark, `
the total amount of light produced by the spark, the maximum
amplitude of the spark current, the maximum amplitude of the
light, the waveform of the spark current, the total duration
of the spark current and the detailed structure of the spec-
trum produced by the spark.
Accordingly, it is highly desirable, particularly in
the absence of the additional inductance element L4, to avoid `
using any combination of capacitance and inductance in the --
capacitor discharge circuit such as to produce an oscillatory
half-period which is close to the deionization time of the
thyratron 32~
In experiments using a hydrogen thyratron, Type 5C22,
it has been found that the half-period of the oscillatory
capacitor discharge current should be either less than 5
microseconds or more than 7 microseconds, to avoid ambigu-
ities as to whether the thyratron is or is not deionized
during the half-cycles when the main capacitor discharge
current is carried by the shunting diode D3. Thus, the
range of half-periods from 5 to 7 microseconds should be
avoided for this thyratron, particularly if the inductance
element L4 is not used. It will be understood that differ-
ent thyratrons will have different deionization times.
Moreover, the deionization time is subject to variations
due to changes in the temperature of the thyratron and the
pressure of the ionizable gas or vapor within the thyratron.
In general, a range of halE-periods which is close to the
-24-
`.'~ "
deionization time should b~ avoide~.
The amblgui~y as to whether or not the thyratron 32
is deionized can cause a significant ambiguity, on the order
of 5 to 10~, in the total elapsed time of the spark current.
The total elapsed time of the spark current is shortened ,~
when the thyratron is deionized during the half~cycles when
the shunting diode D3 is conductive, because the reionization ;~
of the thyratron dissipates additional energy. Such loss of
energy has an additional dampening effect upon the oscillatory -~
capacitor discharge current.
The ambiguity of 5 to 10% in the total elapsed time
o~ the oseillatory spark current can produce an ambiguity on -~
the order of 10 to 20% in the voltage to which the capacitor
C is subsequently recharged. This ambiguity in the capacitor
voltage can cause significant ambiguities in the to~al light
produeed by the spark, the maximum light intensity, and the
detailed strueture of the speetrum produeed by sueh light.
Fig. 4 shows the fur~her modification of inserting
:... .:
another inductance element L5 in series with the shunting - -
diode D3. The provision of the inductance element L5 has
the advantage o producing additional improvement in the
eommutation between the thyratron 32 and the shunting diode
D3. The induetanee element L5 is preferably adjusted to an
induetanee value whieh is large enough to maintain a small
eurrent in the shunting diode D3 when the thyratron 32 is
earryin~ the main eapaeitor diseharge eurrent. Thus, the
shunting diode D3 is kept eonduetive so that it is ready
to earry the main eapaeitor diseharge eurrent when it
passes through zero and reverses so that the eurrent is
switehed away from the thyratron 32 to the diode D3.
When the diode D3 is earrying the main eapacitor
diseharge current, the current also flows through the
-25-
- '.
3~
inductance element L5, so that a magnekic field is built up
around the inductance element L5. When the main discharge
current reverses and is switched to the thyratron 32, the
magnetic field around L5 collapses gradually and induces a
voltage which causes current to flow around the closed loop
comprising the diode D3, the inductance element L5, the
inductance element L4 and the thyratron 32. ~
The provision of the inductance element L5 has the ;-
advantage of obviating the production of any commutation
spikes when the diode D3 starts to carry the main capacitor
discharge current. Thus, the radio frequency inter-ference
which might otherwise be caused by such spikes is obviated.
Moreovar, the provision of the inductance element L5 results
in a further improvement in the smoothness and repeatabilit~
of the commutation between the diode D3 and the thyratron 32, ~-
despite wide variations in the frequency of the oscillatory
capacitor discharge current.
The inductance elements L4 and L5 are preferably
variable so that they can be adjusted to suit a variety of
operating conditions. However, fixed values of inductance
may also be employed.
The effect of the spark timing upon the capacitor
charging voltage has already been described with refexence
to Fig. 5. Figs. 6-9 illustrate the manner in which the
capacitor charging voltage and current are affected by the
spark timing and various other factors, including variations
in the capacitance of the capacitor C, the effective charg-
ing inductance LT and the effective charging resistance Rl.
Fig. 6 again shows the half-sinusoidal waveform 52
of the rectified alternating voltage from the transformer
secondary TlS and the rectifier Dl. Fig. 6 also shows -~
several sets of graphs representing the waveform 54 of the
-26-
. ,:
" '~
P~9~6
capacitor charging voltage for various conditions. 'rhe
graphs of Fig. 6A were produced with the storage capacitor
c adjusted to a value of 0.005 microfarad. Four different
families of graphs were produced for four different spark
igni-tion times: 0, -2.0 milliseconds, +1.0 millisecond,
and +2.0 milliseconds, where o represents the time at which ~ -
the half-sinusoidal waveform 52 is at its peak value. For
each spark ignition time, the effective charging inductance
LT was adjusted to five different values: 1, 5, 10, 15 and
25 millihenrys. It will be seen that the peak voltage to
which the capacitor C is charged increases with the ih-
creasing inductance. For the higher values of inductance,
the capacitor C is charged to a voltage which is substanti-
ally greater than the peak of the supply voltage 52. The
timing of the spark ignition has a very sifnificant effect
upon the peak voltage to which the capacitor C is charged.
When the half-sinusoidal waveform 52 is high during the
capacitor charging interval, the capacitor is charged to a
relatively high voltage, and vice versa. When the capacitor ~;
C has been charged to its peak value of voltage, the diode
recbier Dl prevents the discharge of the capacitor through
the charging circuit.
The graphs of Figs. 6B and C were produced with
different values of capacitance: 0.010 microfarad and 0.030
microarad, respectively. The same four families of graphs
were produced by varying the spark ignition times and the
charging inductance LT, as explained in connection with
Fig. 6A. It will be seen that increasing the value of the
capacitor C had the effect of decreasing the peak voltage
to which the capacitor was charged, in virtually all cases.
The time required to charge the capacitor to its peak value
was increased, as the value of the capacitor was increased.
., : . .
-27~
,' ~, ~'
3~
It will be se~n from Fig. 6 that, for a spark repe-
tition rate of 120 sparks per second, the peak capacitor
voltage can be adjusted over a wide range by changing the
timing of each spark with respec-t to the half-sinusoidal
waveform 52. Much greater repetition rates can be achieved,
if desired, as discussed in connection with Fig. 5.
Fig. 7 illustrates the capacitor charging voltage
waveform 54 for various values of the charging resistance Rl.
Fig. 7 also illustrates the wav~form of the capacitor charging
current 76. The spark timing was the same for all of the
graphs in Fig. 7.
The graphs of Fig. 7A were produced with the capacitor
C adjusted to a capacitance value of 0.003 microfarad. A
family of graphs was produced with eight different values of
the charging resistor Rl, lOK, 2OK t 3OK, 40K, 5OK, 6OK, 7OK
and 80K ohms. The spark ignition time was zero in all cases. ~-
It will be seen that the peak voltage to which the capacitor C
is charged is reduced by increasing the charging resistance.
Moreover, the total charging time is increased. The charging
current, as represented by the waveform 76, is also reduced
as the charging rasistance is increased.
The graphs of Figs. 7B, C and D were produced with
three different values of capacitance: 0.015 microfarad,
0.030 microfarad, and 0.080 microfarad. In all cases, the
spark ignition time was kept at zero. Families of graphs -
were produced for the eight different values of the charging
resistance Rl, as explained in connection with Fig. 7A. It
Will be seen that increasin~ the value of the capacitor Cl
had the effect of decreasing the peak voltage to which the
30 capacitor was charged. The total charging time was increased. - -
Increasing the value of the capacitor C had the effect of -~
increasing the capacitor charging current, as represented
-28-
. . : ' . ,
,
by the wave~orm 7~. In all cases, the capacitor charging volt~
age and current were reduced by increasing the charging re-
sistance Rl.
Fig. 7 clearly indicates that the cnarging o~ the
capacitor tends to be an oscillatory process, in that the ~ -
charging current 76 rises to a peak and then decreases to
zero. However, the diode rectifier Dl prevents the charging
current from reversing, and prevents the capacitor C from
being discharged through the charging circuit. The peak of
the voltage waveform 54 is reached when the current waveform
76 drops to zero after passing through its peak.
Figs. 8 and 9 illustrate the effects of varying the
values o~ the capacitor C and the charging inductance LT.
.. . . .
The graphs of Figs. 8 and 9 show the capacitor charging volt- : -
age waveform 54 and current waveform 76. The graphs of
Fig. 8 were produced for a capacitance value of 0.003 micro-
farad, a charging resistance value of 20 K ohms, and a peak
,: ,
rectified transformer voltage of lOtO00 volts. The spark
ignition time was 0, corresponding to the peak of the recti-
20 fied transformer voltage. The graphs were produced for nine -
different values of the chargin~ inductance LT: 1, 3, 5, 10,
20, 50, 100, 200 and 500 millihenrys.
It will be seen from Fig. ~ that increasing the
charging inductance decreases the peak charging current to
the capacitor C. The total charging time is increased. As
the charging inductance is increased, the peak capacitor
charging voltage is increased at first, but then is de-
creased, due to the act that the half-sinusoidal trans-
ormer voltage is decreasing during the increased charging
time. In each case, the capacitor charging current 76 rises
to a peak and then decreases to zero. The maximum capacitor
voltage is achieved when the charging current falls to zero.
-29-
. - . .
The diode rectif:ier Dl prevents the clischarge of the capa-
citor through the charging circuit, and prevents the reversal
of the charging current after it drops to zero.
The graphs of the capacitor voltage 54 and the capa-
citor charging current 76 of Fig. 9 were produced for an
increased value of capacitance: 0.010 microfarad. The
values of the charging resistance and the peak transformer
voltage were the same as for Fig. 8. The graphs 54 and 76
were produced for seven different values of the charging
inductance LT: 1, 3, 5, 10, 20, 50 and 100 millihenrys.
The peak values of the `charging current 76 were increased,
due to the increased capacitance. The peak values of the
capacitor voltage were decreased. Moreover, the total
charging times were increased, due to the increased capa-
citance. The peak charging voltage increased at first, and
then started to decrease, due to the decrease in the half-
sinusoidal transformer voltage with increasing charging
time.
The spark sources of Figs. 1 and 3 utilize a single -
phase high voltage transformer Tl. The construction of
Fig. 1 utilizes a full-wave bridge rectifier Dl, while the
construction of Fig. 3 employs a half-wave rectifier 70.
Other variations are possible. For example, a three-phase ;`
transformer array may be employed in conjunction with a
thre~-phase diode bridge utilizing twelve diode stacks.
~ s previously indicated, the timing control means
~ of Figs. 1 and 2 may be constructed to time the spark
tri~gering pulses so that the capacitor C is charged to the ;i
same voltage for all of the sparks in a spark train. It is
also possible to charge the capacitor C to different volt-
ages for different sparks, in accordance with any desired
pro~ram or scheme.
~30~
~. . .
'. ' .,~
~?~9~
As previously mentioned in connection with Figs. 1
and 2, the frequency of the oscillatory capacitor discharge
current can be varied over a wide range by changing the value
of the capacitor C and by adjustin~ the inductances of the
inductance coils Ll and L2. At relatively high frequencies,
the half-period of the oscillatory capacitor discharge cur-
rent will be less than the deionization time of the thyratron
32, so that the thyratron 32 will remain ionized and conduct-
ive during the half-cycles when the shunting diode D3 is
carrying the capacitor discharge current. At relatively low
frequencies, the half-period of the oscillatory capacitor -
discharge current will be substantially greater than the
deionization time of the thyratron 32, so that the thyratron
will become dionized and nonconductive during the half-cycles
when the shunting diode D3 is carrying the capacitor dis-
charge current. Thus/ the thyratron 32 will have to be re- ;
fired when the capacitor discharge current reverses so that
it must again be carried by the ~hyratron. The extra energy ;~
required to refire the thyratron has the effect of reducing
. ~ .
the total length of each spark discharge, so that the re-
charging of the capacitor C is initiated earlier than is the
case when the thyratron remains ionized. This earlier initi-
ation of the capacitor recharging process increases the
voltage to which the capacitor is recharged, above the volt-
a~e that is produced when the thyratron does not become de-
ionized. Thus, the calibration of the computer employed in
the timing control means needs to be different for the two
cases, when the thyrakron is deionized or is not deionized
during the half-cycles when the shunting diode D3 is con-
ductive. This ambiguous calibration can be avoided by intro-
duclng a sufficiently large inductance L4 in series with the
thyratron 32, so as to keep a small current flowing in the
~ .:
-31-
~`' " ''
; .: . , . ~ :
9~
thyratron during the half-cycles when the shuntlng diode D3
is conductive.
The shunting diode D3, which preferably comprises a
stac~ of solid-state diode units D3U, has a charge dissipa-
tion time which is analogous in a general way to the de-
ionization time of the thyratron 32. However, the charge
dissipation time is considerably less than the deionization
time. Thus, -the charge dissipation time is not likely to
cause errors or ambiguities, unless the oscillatory fre-
quency of the capacitor discharge current is extremely high.Moreover, the charge dissipation time of any particular
solid-state device is generally dependent solely on its temp-
erature, which can be controlled by cooling the solid-state
device. Any ambiguity due to charge dissipation time can
be avoided by introducing a sufficient inductance L5 in
series with the diode D3, as described in connection with
Fig. 4.
The deionization time of the thyratron 32 is the
most significant factor which may cause uncontrolled vari-
20 ations in the total duration of each spark discharge. How- ~
ever, any other uncontrolled variations in the capacitor -
discharge circuit should be avoided because such variations
may afect the total duration o each spark discharge. Such
uncontrolled variations may include, for example, changes in
the capacitance, inductance and resistance of the various
circuit components due to temperature changes.
It has been found that the timing of the spark
triggering pulses can be calculated by the computer within
an accuracy of about 10~. The factors involved in the com-
putation are the capacitance, inductance and resistance inthe capacitor charging circuit, and also in the magnitude
of the high alternating voltage developed in the transformer
-32-
`: - ' '
secondary winding TlS. From these ~actors, the computer is
able to calculate the capacitor voltage which will be achieved
with any given timing of the spark, or the timing which must
be employed to achieve any particular capacitor voltage. It
has been ~ound that the answer to either of these questions
can be calculated to an accuracy of about 10%. Final adjust~ .
ments are made empirically. However, these final adjustments
are small. The empirical adjustment is made by actually .~ :
measuring the capacitor voltage which is achieved with a
particular timing. If the capacitor voltage is greater or
less than the desired value, the timing is retarded or ad-
vanced until the desired capacitor voltage is achieved.
A wide range of adjustment can be achieved by the
timing control. A minimum voltage is required to produce
the proper operation of the thyratron 32. It has been found ~ :: :
; that this minimum voltage amounts to 2,000 or 3,000 volts
for hydrogen thyratrons. The upper limit of the voltage is ..
about two times the peak voltage E developed by the trans- . :
former secondary winding TlSo A typical control range is
from about .5E to about 1.5E, or a range of about 3 to 1.
This is an adequately wide range, particularly when trans~
lated into the corresponding range of coulombs stored in the
capacitor C. The total amount of light produced by the spark
is proportional to the coulombs stored. The amount of .
~aterial sampled from the electrodes is also proportional ~.
to the coulombs.
The wide ranging control afforded by the spark timing
makes it possible to construct a servo system in which the
spark source is servoed by the photoelectric analytical ::~ .
signal developed by the spectrometer which utilizes the
light produced by the sparks. If not enough light is pro- ~:
duced, the magnitude of the analytical signal will be below
-33-
q~
normal. The analytical signal may be fed to the computer
with a pxogramming to cause the computer to change the spark
timing so as to increase the amount of light to the desired
value.
By controlling -the spark timing, the computer can
control not only the peak capacitor charging voltage and the
area under the curve o~ the spark discharge current, but also
the peak to valley ratio of the spark discharge current wave-
form. The mass of the material eroded from the electrodes
is proportional to the area under the curve. The peak to
valley ratio changes the detailed structure of the spectra
produced by the light source, because elements of different
atomic mass are affected differently by the changes in the ~;
peak to valley ratiQ. Such ratio is affected by the value ; ~ ;~
of the inductance L2 in series with the spark gap, and also
by the peak current which flows initially through such
inductance, because these two factors determine the energy
which is stored in the magnetic field of the inductance.
Figs. 13 and 14 illustrate another modified con-
struction for the timing control means 4~ of Figs. 1 and 2,such modified construction being designated 154. As shown
in Fig. 13, the timing control means 154 include a trigger-
ing circuit 156 for producing pulses of sufficient magnitude
to trigger the thyratron 32 of Fig. 1 into a conductive state.
The triggering circuit 156 derives its operating power from
a power supply 158. As shown, the power supply 158 comprises
a power transformer 160 having its primary winding connected
to alternating current supply Iines 162. A fuse 164 and a
switch 166 are connected in series with the power lines 162
The power transformer 16G has its secondary winding connected
to the input terminals of a bridge rectifier 168 having posi-
tive a~d negative output terminals or leads 170 and 172. A
. . .
-34-
.: ' ' '.
~3~
filtering capacitor 174 is connected between khe leads 170
and 172, which may supply direct current at about 220 volts, ~-
or any other suitable voltage.
In the ~riggering circuit 156, the direct voltage
from the power supply 158 is employed to charge a capacitor
176, one side of which is connected to the positive power
supply lead 170 through a current-lLmiting resistor 178.
The other side of the capacitor 176 is connected through
the primary winding 18QP of a step-up pulse transformer 180
to the nega-tive lead 172, which may be grounded. The capa-
citor 176 is charged -through the resistor 178 and the primary ;~
winding 180P.
The pulse transformer 180 has a high voltage secondary
winding 180S, one side of which is grounded to the lead 172.
The other side of the secondary winding 180S is connected
through an inductance element 182 and a resistor 184 to a
l~ad 186 which is adapted to be connected to the control
electrode of the thyratron 32, shown in Fig. 1.
In the triggering circuit 156 of Fig . 13, the capa-
citor 176 is adapted to be discharged by a control element,illustrated as a silicon controlled rectifier (SCR) 188 having
its negati.ve electrode connected to the grounded lead 172.
The positive electrode of the SCR is connected to a lead 190
which forms the junction between the capacitor 176 and the
charging resistor 178. When the SCR 188 becomes conductive,
the capacitor 176 is abruptly discharged through the SCR and
the primary winding 180P. This discharge produces a high
voltage pulse in the secondary winding 180S, such high volt-
age pulse being employed to fire the thyratron 32 into a
30 conductive state. The discharge current of the capacitor ~ -
. ~ .
176 is slightly oscillatory, to insure that the SCR 188 will ~
not latch in a conductive state. ;
-35-
In Fi~ . 13, the -triggering circuit 156 has an input
lead or terminal 192, eonneeted ~hrough small induetanee
elements 194 and 196 to the eontrol eleetrode of the SCR 188.
Pulses of variable timing may be supplied to the input lead
192 by a timing circuit 198 having internal means Eor pro~
ducing pulses at a repetition rate of 120 per seeond, related :
in phase to the 60 Hertz alternating current from the power
supply lines 162. The timing circuit 198 includes means for
~rarying the phase relationship between the pulses and the :
10 60 Hertz alternating eurrent.
The timing circuit 198 of Fig. 13 includes phase
referenee means 200 to refer the phase of the timing output ~: ..
pulses to the phase of the 60 ~ertz alternating current from :: :
... .. .
the supply lines 162. It will be reealled that the rectified
high-voltage pulses employed to eharge the main capacitor C .... :
in the spark discharge eireuit 10 are also referred in phase
to the 60 Hertz alternating eurrent from the supply lines 18, ~- -
to whieh the supply lines 162 are connected. Thus, the phase
of the timing pulses from the timing cireuit 198 is effect-
20 ively re~erred to the phase o:E the rectified high-voltage
pulses employed to charge the capacitor C.
In Fig. 13, the phase referenee means 200 include a
transformer 202 having its primary winding 202P conneeted to .
the alternating eurrent supply lines 162, which in turn are :
eonneeted to the main alternating current supply lines 18. ..
A phase shifting eapacitor 204 is eonneeted in series with
one of the leads to the primary winding 202P. A switch 206
is provided to short-eircuit the capacitor 204 when the phase . .
shi~ting aetion of the capacitor is not needed. :.~ .; .
~ ,
The transformer 202 has a secondary winding 202S with . .
end leads 208A and B and a center tap 208C. A phase shi:Eting ~ .. .:
circuit 210 is conneeted to the seeondary winding 202S. As : .
-36- : :
'" ,:'.'. '
.
' ~.:
shown, the phase shifting network 210 compri.ses a capacitor
212 and a variable resistor 21~ connected in series between
the end leads 208A and B o~ the secondary winding 202S. An
output voltage o~ variable phase is produced between the
center tap 208c and a lea~ 216 connec~ed to the junction
between the capacitor 212 and the variable resistor 214.
By varying the resistor 210, the phase of the 60 Hertz volt-
age between the leads 208C and 216 can be varied through a
phase angle range of nearly 180 de~rees.
The variable phase voltage between the leads 208C
and 216 is rectified by a full-wave bridge rectifier 218
which supplies its rectified and pulsating output to a ~.
potentiometer 220. The negative ou~put terminal of the
bridge rectifier 218 is connected to the grounded lead 172.
The potentiometer 220 has a slider 220S for adjusting the
magnitude of the pulsating direct voltage between the
slider 220S and the ~rounded lead 172. :
The pulsating direct ~oltage from the slider 220S ~ .
is converted into abrupt pulses by a pulse ~orming circuit
utilizing a trigger element 222, which is illu~trated as a
breakdown diode, such as a diode of the four layer type.
Either of two capacitors 224 and 226 is charged through
resistors 228 and 230, and then is discharged through the
diode 222 and the primary winding 232P of a pulse trans- :
former 232. The capacitor discharge produces a high volt~
age pulse in the secondary winding 232S of the transformer
232. One side o~ the secondary 232S is grounded to the :
lead 172, while the other side is connected to the input ~::
lead 192 which extends to the control electrode o~ the ; .
SCR 188.
In Fig. 13, the charging resistors 228 and 230 are .
connected in series between the slider 220S and a junction. - `
?~3~3~Z6
lead 234. A two-position switch 236 is connected between the ~ -
junction lead 234 and the capacitors 224 and 226. The capa-
citors 224 and 226 are connected to the fixed contacts of the
switch 236, the movable contact being connected to the junction
lead 234. One side of each of the capacitors 224 and 226 i5 - :
grounded to the lead 172~ The switch 236 makes it possible
to change the time constant of the charging circuit compris-
ing the resistors 228 and 230 and the capacitors 224 and 226.
The resistor 228 is variable to provide an adjustment of the
time constant.
The breakdown diode 222 is connected in series with -
the primary winding 232P between the junction 234 and ground.
In Fig. 13, the switch 236 is shown in its position
in which the capacitor 226 is in the circuit. During each `~
rectified pulse from the slider 220S of -the potentiometer 220,
the capacitor 226 is charged through the resistors 228 and
230. When the capacitor 226 becomes charged to the break-
down voltage of the diode 222 r the diode breaks down abruptly
and becomes conductive, so that the capacitor 226 is dis-
20 charged abruptly through the primary winding 232P. The ~ ;
discharge current pulse produces a high voltage pulse in
the secondary 232S. Such high voltage pulse fires the SCR
188, so that it becomes conductive.
The capacitor 176 is then discharged through the SCR ;
188 and the primary winding 180P. The discharge current
pulse produces a high voltage pulse in the secondary 180S.
Such high voltage pulse fires the thyratron 32 of Fig. 1, so
that the main capacitor C is discharged across the spark gap G.
The potentiometer 220 and the variable resistor 228
::
30 can readily be adjusted so that one triggering pulse will be -
produced for each rectified pulse from the bridge rectifier
218. For a 60 ~ertz alternating current supply, there will
-38~ `~
'' ' ": ' '
;,:
be 120 rectified pulses per second. Accordingly, the timing
circuit 198 will produce 120 ~iming pulses per second. The
timing of these pulses can be adjusted, relative to the phase
of the rectifiea pulses, by adjusting the potentiometer 220
and the variable resistor 228. For example, the timing can
be adjusted so ~hat each timing pulse will occur at the peak
of the corresponding rectified pulse~ Other timing adjust-
ments can also be employed.
After the capacitor 224 or 226 has been discharged
through the breakdown diode 222, the diode again becomes non-
conductive, so that the capacitor will again be charged during --
the next rectiiied voltage pulse. Thus, the ~iming circuit
198 normally produces 120 timing pulses per second when oper-
ating under internal control. The phasing of these pulses
can be adjusted by varying the resistor 214~ which provides
an adjustment of the phase angle of the alternating voltage
through nearly 180 degrees. An additional change in phase
can be produced by opening the switch 206 so as to introduce
the capacitor 204 into the primary circuit of the transformer
202. Thus, the timing pulses produced by the timing circuit
198 can be adjusted into any desired phase relationship with
the rectified high voltage pulses employed to charge the main
spark discharge capacitor C.
The timing circuit 198 also includes a provision for
external control. I'hus, in Fig. 13, an input lead 240 is
aonnected to the junction 234 through a diode 242 and a
switch 244. When external control is desired, the switch
244 is closed. The potentiometer 220 is also generally ad-
justed to zero, or the transformer primary 202P is discon-
nected from the alternating current supply lines 162.
~ Fig. 14 ilIustrates an external control timingcircuit 250 for supplying two or more timing pulses to the
-` -39-
'.".. " .
.
3~i
external con~rol lead 240 during each o~ the 120 recti~ied
pulses. Provision is made for individually timing each of
the timing pulses during each rectified pulse. The external .. ..
timing pulses break down the diode 222 in Fig. 13 and produce
current pulses in the primary winding 232P, sO that the SCR
188 is ~ired. This, in turn, fires the thyratron 3Z of
~ig. 1.
In Fig. 14, the external timing control circuit 250
includes an output amplifier 242 to produce output pulses of
a sufficient magnitude. Amplifiers of various types may be
employed. The illustrated amplifier 252 employs three trans- ..
istors 254, 256 and 258 in a "totem pole" circuit which also
utilizes a biasing resistor 260 and an input resistor 262.
Timing pulses of relatively small magnitude can be supplied
to the input lead 264 of the amplifier 252.
In Fig. 14, the individual timing pulses are pro~
duced by individual trigger circuits 270. Two such trigger ...
circuits 270A and B are illustrated, but any desired number -.
of trigger circuits may be provided, corresponding to the ...
number of timing pulses which are desired for each of the
120 rectified alternating current pulses. Each trigger
circuit 270 can be adjusted to control the exact timing of
its output pulse, with respect to a common input pulse, ~:
whichis phase locked to the 60 ~ertz alternating current : . :.
:
supply, and thus to the rectified high voltage pulses .
supplied to the main spark capacitor C of Fi~. 1. The . :
common input pulse is produced by a phase locking circuit . ::
272 in Fig. 15 and is supplied to the trigger circuits 270 :. :~
by a common lead 274. .:;.
The trigger circuits 270 have individual output ~
leads 276 which are coupled to the input lead 264 by a .~ .
multitude NOR gate 278. Any suitable number of such gates
-40-
3~;
may be employed to provide enou~h inputs ~or all o~ the
trigger circuits 2700
The phase lockin~ circuit 272 o~ Fig. 1~ includes
reference means 280 for referring the input of the circuit
to the phase of the alternating current supply, such refer-
ence means being shown partly in Fig. 13 and partly in Elig.14.
Thus, in Fig. 13, such reference means include an additional
transformer 282 having its primary winding connected to the
alternating current supply lines 162, which in turn are con
10 nected to the main supply lines 18. The transformer 282 has
its secondary winding connected to a full-wave bridge recti~
fier 284 with output leads 286 and 288. The negative output :~ .
lead 286 is grounded to the lead 172, while the positive lead
288 is connected to a voltage divider, comprising resistors .
290 and 292 connected in series between the leaa 288 and ~ ~ :
ground. A lead 294 is connected to the junction between the
resistors 290 and 292. The lead 294 also appears in Fig. 14 : :
and is operative to supply rectified alternating pulses to .
the phase locking circuit 272. ..
In Fig. 14, the phase reference means 280 may include
a follower amplifier 296 having its input connected to the
lead 294. The output of the amplifier 296 is connected to
pulse generating means 298 for producing brief pulses of
uniform amplitude, corresponding to zero points of the recti-
fied pulses from the amplifier 296. .
The illustrated pulse generator 298 utilizes a mono-
stable multivibrator 300, preferably in the form of an inte-
grated circuit. The B input of the monostable 300 is con- .... ..
nected to the output of the amplifier 296. The duration of
each pulse produced by the monostable 300 is determined by a
timing capacitor 302 and a timing resistor 304, connected to
the monostable, the resistor 304 being connected between the :.
-41- ~ ;
. ,
monostable and a power supply line 306. sypass capacitors 308
and 310 are connected between the supply line 306 and ground.
The phase locking circuit 272 functions as a zero
crossing detector, to produce a brief pulse corresponding to
each zero crossing of the alternating current supply. Such
pulse is supplied to the output lead 274 which is connected
to the input of each of the trigger circuits 270.
All of the trigger circuits 270 may be the same,
except for adjustment, so that it will suffice to describe
the first trigger circuit 270A. In general, the trigger
circuit 270A comprises a timing or delay circuit 320 and a
pulse duration circuit 322. The timing or delay circuit 320
.
determines the timing of each output pulse produced by the
trigger circuit 270A, relative to the reference provided by
the corresponding input pulse. The duration circuit 322 -
determines the duration of each output pulse.
In Fig. 14, the timing or delay circuit 320 utilizes
a monostable 324 having a timing capacitor 326 and a tlming
resistor 328, the resistor being connected between the mono- `
20 stable and the power supply lead 306. The resistor 328 is
variable so that the duration of the monostable output pulse -~
can be adjusted. Bypass capacitors 330 and 332 are connected
to a supply lead 306.
A two-position switch 334 is connected between the
input pulse supply lead 274 and the B input of the monostable
324, the movable contact of the switch 334 being connected to
the B input. One fixed c~ntact is connected to the lead 274,
while the other fixed contact is grounded, so that the trigger
circuit 270A can be inactivated by operating the switch 334
so that the B input will be grounded.
The duration of the output pulse from the monostable
324 determines the delay in the production of the output
-42-
9~;
pulse from the trigger circuit 270A. The duration circuit 322
determines the duration o~ the output pulse from the trigger
circuit 2 7 OA .
As shown in Fig. 14, the duration circuit 322 utilizes
another monostable 340 having its B input connectea to the
output of the monostable 324. Thus, the monostable 340 deve-
lops an output pulse corresponding to the trailing edge of the
output pulse from the monostable 324. The monostable 340 has
a timing capacitor 342 and a timing resistor 344 which deter-
mine the duration of the pulse produced by the monostable 340,the resistor 344 being connected between the monostable and
the power line 306. Bypass capacitors 346 and 348 are con-
nected between the power line 306 and ground. The pulse out-
put line 276 of the trigger circuit 270A is connected to the
Q output of the monostable 340.
In operation, the phase locking circuit 272 produces
a brief pulse corresponding to each zero crossing of the ,
alternating current from the supply lines 18. The output
pulses from the phase locking circuit 272 are supplied to
all of the trigger circuits 270. There may be any desired
number of the trigger circuits 270. Each trigger circuit
27 0 produces a single output pulse in response to each input
pulse~ The timing oE each output pulse is adjustable by
varying the resistor 328. The trigger circuits 270 are
adjusted so that the respective output pulses are produced -
sequentially at -the desired time intervals during each half
cycle of the alternating current supply. The timing of each
output pulse determines the magnitude of the corresponding
spark discharge. It is generally desirable that all of the
spark discharges be substantially the same in magnitude.
The timing of each output pulse from the trigger circuits
270 can be adjusted to bring this about.
.: - .
-43- ~ ~
- ",' ' ' ": '
'.,.
This adjustment can be brought about by connecting
the input of an oscilloscope, or some other voltage indicator,
to the spark discharge circuit ~f Fig. 1, so as to display
the capacitor charging voltage of the main capacitor C or the
spark current across the spark gap G. The timing resistors
328 can then be adjusted suf~iciently for the successive
trigger circuits 270, so that the successive spark discharges
are all the same in magnitude. There is no serious diffi-
culty in carrying out this series of adjustments.
10Alternatively, the timing of each spark triggering
pulse can be calculated, as previously discussed, with or ;~
without the aid of a computer, on the basis of the various
parameters of the spark discharge circuit of Fig. 1. Such
parameters include the capacitance of the main capacitor C,
the charging inductance represented by LT, the resistance
represented by Rl, the alternating voltage developed by the
transformer Tl, the capacitor voltage to be maintained, and
the duration of each spark discharge.
For those who wish to employ a computer to make
these calculations, a computer program is available to the
public at the computer center of the University of Wisconsin
in Madison, Wisconsin. Access to this computer program can
be obtained by anyone having a standard teletypewriter
computer terminal. Such computer program is designated OLD
LCR2. This computer program makes it possible for the
computer to compute the timing of the spark triggering
pulses in order to produce any particular voltage across
the main charging capacitor C fo~ each spark discharge.
Conversely, it is possible to compute the capacitor charging
voltage for each spark discharge, if the tLming is given.
Those skilled in the art will be able to assign
appropriate values and type designations to the various ~
~44~ ;
: ~ . . ,
3~
components of the electrical circuits disclosed herein.
However~ it may be helpful to offer the following component
value and type designations which have been employed suc-
cessfully in actual tests. It will be understood that
these component values and type designations may be varied
widely to suit various operating conditions.
C A P A C I T O R S
CAPACITOR VALUE IN MICROFARADS
OR PICOFARADS (Pf)
10 174 8.
176 0.002
204 0.22
212 1.
224 0.015
226 0.015
308 3
310 50 Pf-
302 100. Pf.
326 0.015 ~-
20 330 3
332 50 Pf- "
342 0.015
346 3.
348
,'''~ '~ ''' '
' ' :
. :'.
'. ,
.. ;: .:.'
: '. . :
-45-
3~
R E S I S_T O R S
RESIS~ORS VALUE IN OHMS
178 2.7 k
184 100.
214 10. k
220 10. k
228 300. k
230 10. k -
262 18. k
10 264 1.5 k ~ -~
290 1000. `
292 1000. :
324 10. k
328 10. k
344 1. k
I N D U C T O R S
INDUCTOR TYPE
182 Ferrite bead.
194 Ferrite bead. .
20 196 Ferrite bead.
:':
.'
,,:;
-46-
~~ a3~;~6
SOLID STAT~ COMPONENTS
COMPONENT TYPE
_ .
188 2N4333 Motorola SCR.
222 lN5160 Motorola 4-layer diode.
254 2N3904 Motorola transistor.
256 2N3904 Motorola transistor.
258 2N3906 Motorola transistor. -
278 SN7423 Dual Quad-input NOR gate.
296 ~A742C
300 74121N Monostable multivibrator.
324 74121N Monostable multivibrator. --~-
340 74121N Monostable multivibrator. ~ -
.'' ' '
. ' - . .:
. :'~ '; ., :'
"" ' "~ '
.. ..
..', ,: ' ,. . .
, ," .
'; '~ '' :'' ..':.'
.. . . .
.
~ - ,~ ."
; ' :"'
..
:. - .
., .:
-47
:. . ,: - . -
. :: '-,:
, '
