PFC/RR-82-28
DOE/ET-51013-57
UC20
PREDESIGN OF A NEW OHT POWER SUPPLY
FOR VERSATOR II
JAIME S. RAMOS*
PLASMA FUSION CENTER
MASSACHUSETTS INSTITUTE OF TECHNOLOGY
CAMBRIDGE, MA
02139
SEPTEMBER 1982
*Visiting scientist from Instituto Nacional de Investigaciones Nucleares
Benjamin Franklin 161
Mexico, DF 06140
ACKNOWLEDGMENTS
1. To the MIT Plasma Fusion Center, for making this visit possible.
2. To the Instituto Nacional de Investigaciones Nucleares, Mexico,
who also made this visit possible.
3. To Dr. Stanley Luckhardt, for his guidance on the matter of this
report.
4. To Mr. Rajeev Rohatgi, for his plotting subroutines.
CONTENTS
Page
0.
Introduction
0
1.1
Current buildup
1
1.2
Energy calculations
2
1.3
Other equipment
3
2.
Current commutation
5
2.1
Circuit operation
5
2.2
Analysis
5
3.
Start up phase
8
3.1
Operation and requirements
8
3.2
Circuit analysis and design
8
Results
11
Graphs
llb,c
3.3
Breakdown conditions
12
4.
Double swing phase
14
Graphs
14b,c,d,e,f,g,h,i,j ,k
Shut-off Phase
15
Graphs
15bc,d.
Conclusions
16
5.
6.
- 0 -
0. Introduction.
A basic design of a new OHT power supply for
VERSATOR II, is advanced.
A requirement on the design is to permit the operation
of VERSATOR II in two modes: the double swing, in which a
longer plasma pulse should be obtained; and the primary
shut-off, when the plasma is not driven anymore by the OHT,
but by RF waves
(PF current drive mode).
The report was organized as follows:
In Section I the establishment of a current bias is
examined.
In Section 2 the commutation phase is analyzed, restricted
to the use of a saturable inductor.
Section 3 deals with the start-up phase, which breakdowns the gas, creating the plasma.
Section 4 is devoted to the double swing drive, and....
Section 5 to the OHT shut-off.
In Section 6 a list of major components is written down.
ji
E
mo
,,C:
0
Oo
E
->-a:<
0
m
i
2u
W(D
rn W(
C,)
oJz
0 -j
on
0
Cl)
0(l)
a
D
z
Z
L
z
0
a.4>
0
4J
-d
OD
0
I.
u13
0
C)
N
-4
C)
<I-
0
'-4
('I
N
CD
HIJr
OD
0U70
'I)
w
'-4
a)
N
-'.4
(Z4
- 11.
Current Buildup.
This is the the first phase of the OH operation, with
the aim of establishing a magnetic bias in the system, from
which the swing of the OH current can follow.
The maximum
current reached during this phase, Im, must be as large as
compatible with the current capacity of the OHT primary in
order to maximize the available external volt-seconds (MIm
for plasma consumption.
For Versator II, M, the mutual induc-
tance between the primary and plasma equals 8pH, and Im=1 5 kA.
Therefore, volts-seconds = 0.12, a 6% of these available in
the Alcator C machine.
One typical circuit for establishing the magnetic bias
is shown in Fig. 2.
Initially V1 is fired, and IOH starts to
increase until reaching the maximum of 15kA,
crowbarring the OH.
bottle
when V2 is
fired,
The current builds up through the vacuum
(VB), which remains in the conducting state (closed),
and through L sat
These two components are necessary for com-
mutating the OH current through the path AB, later on.
In
practice, VB and Lsat may be three vacuum bottles and saturable inductors in the series parallel connection shown in Fig.
3, according to the current handling capacity of the VB's.
According to the information given by the Alcator Group a
single inductor varies its self-inductance from
3000pH
-
in
the low current regime, <350A - to 20pH - in the high current
regime, >350A.
Its internal resistance is constant and ~3mQ.
Therefore, for the circuit shown in Fig. 2, R sat=lmQ, and
Lsat=lOpH (high current regime >lkA) or Lsat=lOOpH (low current regime <lkA).
- 2 -
2.
Energy Calculations.
To raise the current to the lkA level where the inductors saturate little energy is needed, numerically,
1/2(71OPH) (lkA)2 =355J.
A larger amount of energy is needed
to reach the 15kA; for this case, l/2(620pH)(l5kA) 2 =69750J
This energy comes from Wj in Bank 1.
are necessary.
energy is lost in heat in this process.
Some
The efficiency of
this process is defined as
LI
T
2
m = exp(-2ywt )/(l-y 2 )
2
(1)
m
2
Equation (1) is a very good approximation for y 2 =R 2 C /4L10-3 ,
w is the radian frequency = 1//LC,
time.
and tm=1.57/w is the rise-
In the present case L=620pH, and R=12mQ.
For Bank 1
it is advisable to use lmF, 4kV, 8kJ, unit capacitors, in order to maintain uniformity with the 700kJ toroidal bank already in use.
For ten capacitors in a parallel connection,
Cl=lOmF, WI=8OkJ, and n= 9 2%, and therefore 73.6kJ are available for conversion into magnetic energy, and 6.4kJ are lost
heating the conductors during each shot.
Taking into account
a 10% supply margin, Bank 1 can be specified as follows:
BANK #1.
UNIT:
lmF, 4kV, 8kJ
TOTAL OF 11 CAPACITOR IN PARALLEL CONNECTION
llmF, 4kV, 88kJ
n = 92%
-
3.
3 -
Other Equipment
The A-s rating of Vl must be larger than QI=ClVI=44CI
plus the charge passing through it during the crowbar action,
(assume l5kA passing during 2ms before current commutation,
in normal operation) 30C, plus the small amount leaking
through it during commutation.
bers, Vl must be rated at 80A-s.
Adding together the above numFor V2, 36 A-s are required.
Both V1 and V2 must hold -4kV reversed voltage.
Vl:
Tube GL/37207A
V2:
Same Tube.
Therefore:
200A-s
One may consider larger GL8205 tubes for V1 and V2, in the
case of failure (commutation not obtained).
- 4 -
Vac
cum
B ottle
4A
Comm
t=tl
ation
ROH
LOH
Bank
+A+
V2
Fired at t2
Figure 2. Circuit configuration during the current
buildup in OHT phase
Figure 3. Connection scheme of saturable inductors and
vacuum bottles
- 5 2.
Current Commutation
2.1.
Circuit Operation
By the time when the OH current is maximum, Bank 1 has
been crowbarred (at t 2 ), and being a burning arc in VB
state),
V3 is fired at t=t3
(open
in order to provide a diverting
path AB for the OH current, and to drive through zero the curto
rent in VB, the only way to get VB/do its job. Figure 4 shows
the circuit at t 3 .
Bank 2 must be polarized as shown because
it must oppose the current in the vacuum bottle, driving the
current through zero.
When this happens, Lsat increases its
value, the period of oscillation increases and di/dt in the
vacuum bottle decreases; and so current is commutated.
An al-
ternative operation may be the use of a SCR instead of the
saturable inductor, as PRETEXT is operated.*
C 3 >C
2
In Fig. 4,
, and so C 3 can be neglected in the analysis.
Also
shown is a spark gap and the pulse shaping network to be described below.
2.2.
Analysis
The circuit model is shown in Fig. 5.
are IOH
SAT
, Ic=0, and Vc=-V 2 .
Initial conditions
The meaning of the above
symbols are implicit in the Figure.
*S.A. Ekstrand et al, "A Simplified OH Circuit for Tokamaks.
Sept. 1981.
DOE/ET/53042-5.
Fusion Research Center.
FRCR #232, Fusion Research
-
6 -
Neglecting all resistances:
IOH(t) = Im + (V2 /WLOH) sin W(t-t 3 )
ISAT (t)
= Im
IC(t) =
(V 2 /WL)
where
-
(V2/Lsat)
W2 =WOH 2+
sin
sin
w(t-t3 )
W(t-t 3 )
Wsat2 = (CL)
(2)
(3)
(4)
1
and L is the inductance of Lsat in parallel with LOH'
The design of Bank 2 follows from Equation 3. In order to make
ISAT
0'
7r/8,
then from (3):
V 2 > wLsat Im. If the null point is reached at ,(t-t
1/2 C2 V 2 2 =
Lsat (1+Lsat/LOH) m2
3 )=
(6)
Waveforms are shown in Figure 6. After current commutation the
voltage on C 2 reverses because C 2 is flooded by the bias current,
then V4 is fired. In case of V4 failure Banks 2 and 3 can be protected with spark gaps.
Equation 6 gives 2250J. Consider W =3kJ,
choose V 2 =4kV, then C =375pF, and the commutation time T/4=W/LC
~300ps
(low current regime). If V 2 =5kV, C 2 =240yF and T/4~2501ys,
and the commutation time is not shortened very much; therefore the
4kV bank is preferable.
From Equation 2 one calculates the contribution of Bank 2 to the
OH current, it being small,approximately 500A.
-
7 -
IVB
1+
-rn~I
15kA
V3V3q ~ir
5k;
1
A
ed a tt3
oh
U
Fired at t4
Lsat
spark
gap
Bank 2
ulse shaping
network
LOH
Bank 3
I
Figure 4.
V
I
Equivalent circuit for commutation phase
-
ALL
I
-
4
sat
In
I
'C
transition
C
sat
Figure 5
+- I M
0
2
V2
.
LOH
Model for analysis
of Figure 4.
OH
Im
0
time
Figure 6. Waveforms of currents
through the SAT and OH inductors
-
8 -
3.- Start up phase.
3.1 Operation and requirements.
At time t4 when V4 is fired the bias current is diverted
into the resistor, suffering a sudden drop which induces the
electric field E necessary for breaking the gas. A good analysis
of the physical mechanisms responsible of this process is that
of Papoular*. For present day VERSATOR II parameters, the time of
effective apliccation of the E field, defined as T/8 (of the fast
bank) is < lms, and the start voltage is = 30V. The following design
must be able to reproduce those parameters, as well as determining
the resistor, minimizing the Volt-second consumption, and determining the leftover OH current and voltage in C 3
3
3.2 Circuit analysis and design.
The circuit operating in this phase can be reduced to that of
Figure 7. For t>t , the exact expression of the OH current is:
IOH (t) = I
-
exp(-Iit/2) [ coshwt -
(p/2w)sinhwt ]
(V3/wLOH) exp (-i t/2) sinhwt
where P=R/LOH
, R=Rst+ROH ' w0 2 =/(C3LOH)
(7)
.2
=
22
For the case of strong damping, wo<<p/2, and (7) reduces to
R. Papoular.
The Genesis of Toroidal Discharges.
Nucl. Fus. 16,1,p37, (1976)
-
IOH t)
=
9 -
I m[exp(-it) +(LOH/R)/(RC3)
V3/R (exp(-t/RC3)
-
-
-(exp(-
pt)-exp(-t/RC3))]
exp(-jt))
(8)
As can be seen there are two decay modes: a fast one with time
constant LOH/R , and a slouvne with time constant RCs. Equation
(8) can be further reduced to
IOH (t)
= Im exp(-lUt)
,
for t>t 4
(9)
This is our working expression. The induced voltage is
vst(t) = MdIOH(t)/dt = -MIm exp(-t/T/st
(10)
Tt
where Tst = 1/P = LOH/R
(11)
In the design we will consider a maximum voltage of 30V
and a minimum voltage of 1bV. The latter value should be compatible with Papoular's analysis (see below) and maybe within the
operation regime of Versator II once the field error is reduced
to ~
2G.
The volt-second consumption (VS) must be minimized in order
to leave unwasted flux for a longer plasma pulse. Defining the
percentile Volt-second comsumption as VS = M(Im
VS(t) = 1 - exp(-t/T st)
Setting the values of vst and VS(t),
finds
Tst
,
OH (t))/MIM , then
(12)
from (12) and (10) one
the decay time, then from (11) the resistance, and from
(12) the time tf marking the end of the start up phase. Equation
(9)
gives the left-over OH current for the next phase; the energy wasted
Q in the resistor can be calculated from the usual formula:
I
-
Q =
RI1
Jo st OH
2
dt = R t2
st mn
10 -
sT ( 1 - exp(-2t/T ) )/2
st
st
(13)
Finally the resistor should be specified, taking into
account its increase of temperature.
Q = cMAT = cVoAl
AT
(14)
where c is the specific heat, v, is the mass density, A is the
cross sectional area, 1 is the length of the resistor, and AT is
the temperature increase
Together with the usual resistance formula R = pl/A
Consider stainless steel
(15)
( 0.1% C, 8% Ni, 18% Cr, 73.9% Fe),
then
p = 90xl0 8-1-m, v0 = 7,800 kg/m 3 , c = 460 J/kgC. Choosing a rectangular plate of 1/16" thickness and 2" width (a rectangular geometry
has more dissipation area than a cylindrical one),
and
then from (14)
(15):
AT(C) = 3.85x10- 5 Q(J)/Rst(o)
,
1(m) = 89.55 R stW
Results are summarized in Table I. For a constant flux con sumption the voltage is reduced if the resistance is reduced and
the duration of the pulse increased. For a constant voltage the flux
co sumption is reduced by decreasing the resistance and the time
duration.
For 10V loop voltage the temperature increase, 17.1 C, seems
rather large. It is advisable to use 1/8"x2" plate, instead. Consereduced by a factor of 4
and the length doubled, to a maximum of
quently, AT will be
35 m;
if i5kA aren't available, then more resistance is needed. For
13kA, a resistor 40m long is necessary to induce 30V. 30 strips, each
1.5m long can be taylored from a 3.175cmxl5Ocmxl67.5cm plate, allowing
for a 10% loss of material. This material will be sufficient for
making a 0.25 Ohm resistor. Shut-off operation requires 20% extra
material
(See page 15b).
- 11 TABLE I
RESULTS OF DESIGN OF THE START UP PHASE
I
= i5kA
0.12 V-s available
Induced voltage
Consumption of Volt-seconds
vst(tf) in V
25%
30
20
10
12.5%
0.193
40.0
28.6
0.164
34.3
15.2
3
11.2
5.7
3.5
13.1
3.6
0.86
43.5
17.3
0.47
21.7
14.7
0,126
26.7
27.8
0.106
22.9
14.7
4.5
11.2
8.6
5.2
13.1
5.3
1.3
43.5
11.3
0.7
21.7
9.5
0.058
13.3
25.6
0.048
11.4
13.3
9
11.2
17.1
10.5
13.1
10.7
2.6
43.5
5.2
1.4
21.7
4.3
Rst
Tst(ms)
tf (ms)
(0)
Q(kJ)
IOH (tf)
AT(C)
Ip (tf)
1(m)
GUIDE TO TABLE I
LIST OF SYMBOLS:
R
: Resistor for start up phase (ST)
vst : Induced voltage
T st:
tf
Decay time.
: Duration of ST phase.
Ipi: Plasma current.
Q:
Joule loss in the resistor.
1: Length of stainless steel resistor. 1/16"X2" plate.
AT: Resistor's temperature increment.
-
O H
C
U
R R E M
T
llb -
C A MP )
D A
MP E D
ov
15, QOQA
2F
00. . ISE
OE
j
0. s
10.0
V O L T S
L OO P
5
6.0
0.1360hm
0.EGE-5.*
0. "3
4
G.6e
0.c9
0.O1
0.C7
£.&oZ
*.
1;
-16.9-
I,
Seeow ablcolmn
-0.1
cas
5
See Table I, row 2 column l. case
Voltage at tf is 20V
5
-0.1SE
T
OH
C
U R R E H T
( A MP )
S
D A M P E D
O P
15,OOOA
2F
M E
I
-30.0
)
10
V OL T S
OV
58mOhm
5.0
0. IE
-eE
4
See Table I, Row 3, column 3 case
1s.0
Both of these graphs are Start-up phase
Voltage at t f is lOV
10E
5
T
I
M E
CS
.
-
0 H
C
0. ISE
3
U R R E H T
C
llc
-
AAM P
AL T E R M
V 0 L T S
L 0 0 P
0.133
0.113
0.SE4+++++
0.020
-~
0.040
0.068
e
0.100
0.0ae
.*
t-30
- -. SOE
-wiE
4
s
f t 13,300A
++ fa,300A
2F
J-9.1SE
oO
-as. 0-
10mOhm
5
T
I
M
E
S
These curves illustrate the double-swing. Note the time
scale. In the ttt case less flux was wasted in plasma
formation, than in the +++ case. Consequently,
the I OH
swing is larger for the +tt case, and the plasma current
will also be larger. Note that the voltage does not
change much, from one case to the other.
- 12 3.3 Breakdown conditions.
Papoular presents a model of breakdown in Tokamaks based
on the Townsend avalanche. The competing mechanisms in the
process are ionisation by electrons accelerated by the induced
E field, and loss of electrons to the chamber walls due to diffusion, drift, and field errors. These mechanisms depend on a,
the first Townsend coefficient, and on Te; which are in turn
increasing functions of E/p. Since present data is limited to
low values of E/p, an extrapolation is made to the range E/p
100V/cm-mmHg. Present operation of VERSATOR II is in the range
1,200-5,000 V/cm*mmHg. In the future operation of VERSATOR II,
E/p can be in the range 400-1000, because of the reduction of the
loop voltage.(The field error is expected to diminish to 2G).
Choosing E/p = 400 V/cm'mmHg, the loss rate equals 6 = 1,700 sand the ionisation rate is v = av = 36,000 s~.The avalanche will
develop if v> , and breakdown will occur at a critical time
tcr
at which a sufficient number of electrons are present within the
chamber - a 10% degree of ionisation -.
tcr depends on the number
of initial electrons prior to the external application of the E
field, as well as geometrical parameters and toroidal B. Assuming
a degree of preionisation of 5 electrons per million neutrals,
calculation gives tcr = 0.28ms, which is in fact less than the escape
time due to the various loss processes and tf in Table I. According
to the model the minimum E field is 0.0275V/cm. For 10V of loop
voltage, E = 0.04V/cm. In conclusion, the design parameters in
Table I are compatible with Papoular's model of breakdown conditions.
-
13 -
m
OH
LOH
Doubl
Swing
Rst
ROH
Bank 3
FIGURE 7. Equivalent circuit for start up phase, or
continuing phases (steady state, double-swing)
- 14 -
4.
DOUBLE SWING PHASE.
Once the plasma is formed the loop voltage must be reduced,
because the plasma can't take so much heating. Therefore, in Figure
1, V5 is fired,and later-on V6. Once the whole resistor is shorted
V7 can be fired to help V3 and V4 with the Amp-sec. V3 will stop
conducting when the OH voltage drops below V 3+20V. The larger V
3
the sooner the discharge will end -if V7 is not fired. Other possibility is to use a vaccum bottle, instead of V7. The bottle will
stay open -non conducting - until V6 is fired.
From Figure 7, various graphs of OH current, loop voltage, and
C 3 voltage can be obtained. These are shown in Figures 8a-i. Ideally
a l5kA swing in the OH current would produce a plasma current of 170
kA. This is clearly an overestimation, but the real plasma current can
well be in the vicinity of a 100kA. Time scales are shown. Pulse
duration (plasma) can be 50ms or more. A 2F Bank provides longer
plasma pulse and lower loop voltage..
Figures 8a-c show the circuit response when the initial voltage
on C 3
is diminished: the OH current falls less abruptly. Figure 8d
is to be compared with 8c, the difference being a larger resistance in
the circuit. Figures 8f-h form a group, having the same parameters,
except that the resistance is decreased. Figure 8i is a continuation
of 8f, when the resistance is shorted at lOms; accordingly, the loop
voltage is reduced.
In conclusion, the graphs are more useful to show the principle
of operation, rather than to predict exact circuit dynamics, because
of the problem in modelling the plasma. A 2F double-swing Bank seems
preferable than a lF Bank, because of smaller loop voltage, and longer
pulse duration
14b
C U R R EN T
O H
(A
L T E RN
M P )A
1e.e
V O L TS
L O O P
5
0.1SE
3.,
.1*E
-0.5
11,OOOA
450V
1F
10mOhm
6.0
4
T.0S
- -.
SGE
4
-4~
-
-8.0T
- 0. 48E 3
M E
11, OOOA
450V
1F
1OmOhm
C
S
3
-0.20E
.4
I
-O.EOE
3
-0.4GE
3
T
I
N
Figure 8a. ALTERN means "alternating" response
E
c
C
14c
C U R R E N T
O H
C A M P)
A L T E R
5
e.15E
N
10.0
V O L TS
L O O P
8.5
.1SE
6.0]],000A
If
350V
4.0-
IOmOhm
4
0.6E
- .. ~l1 'l"Illa.,
W*.w&
T
--9. SE
4
0*019
:1
II
0.03S
.0"e
I
I
re.1e
I
0.00
0.0"0
0.040
04070
-0-.1.E
T
V 0 L T S
0 H
3
O.Me
3
S. 1GE
3
1],OOOA
350V
IF
]OmOhm
*
0.0
-0.10E
3
-0.S2E
3
-0.30E
3
-e.40E
3
0.030
0.*040
Farads and Ohms,
Amps and Volts are .i-itial
'
E~w.U~U
*
S.,?.ev
0.9S
T
Figure 8b.
6
C.
3
0.I3E
I M E
I
M E
C
S)
refer to Figure 7.
values in circuit
of Figure7
14d
C A M P )
C U R R EN T
O H
0.15E
A L T E R N
6.0
V OL T 6
L O O P
5
I0.50E
8.0
0. S
IF ] OmOhm
4
0.010
0E
-..
6.0
OV
1),OOOA
e.,2,
0.030
0
B.o-
0.6
.7
4
-4.0-
-a.*-8.0
S
-0.16E
IM
T
V O
O
LT
0.40E
3
0.30E
31],000A
* *
1E
OV
IOmOhm
3
3
0.010
Oo
SC )
C.
IF
0.80E
E
-0.10E
3
-0.80E
3
-0.30E
3
-. 40E
3
O
a
*
03
I
0.630
.
0.0"0
I
.tzo
*.Me0
T
Figure 8c.
I
R
1
006E?0
E
CS
)
14e
C U R R E N T
O H
(A
MP
A L T E R H
s
0.15E
L 0 0 P
10.0
V O L T S
8.0
E5
OV
],OOOA
O.SGE
IF
4
4.0-
20mOhm
a.,-
0~i l~
e
0.030
e
ee
ee
-4.0-09.10
-6.*0
GE
-8.0
-|
-0.1TE
~T I M
-
U O L T S
0.46E
3
0.30E
3
ON
.
OV
]],OOOA
0.88E
-10.0
E
IF
3
20mOhm
0.10E
L
v.wwE.
y
-0.10E
3
-0.50E
3
-0.30E
3
-0.40E
3
0.0"0
0.030
0.040
T
Figure 8d.
0.079
G-0.0
0.05
I
M E
(
S
)
14f
O H
C U R R E H
A L T E R N
c A M P)
5
0.15E
L
0 0 P
V
10.0
OL T S
8.0
.ISE
5
6.0
300v
SI, OOOA
20mOhm
]F
4.0-
0.6
4
-TI.6E
4
-4 . 0-
6
-I.SE
T
V 0 L T 8
0 h
1
14
3
*.I0E
3
G.OE
3
I
0.010
-0.120E
3
-6.30E
3
-0.4SE
3
0.030
0.040
)
20mOhm
0.060
0.050
T
Figure 8e.
-10.0
300V
1F
0.080
C S
C
]LOOOA
9.30E
E
I
M
E
(
S
)
14rj
O H
C U R R E H T
0. ISE
A M P)
C R I T I C
5
L 0 0 P
10.*0
UOL T S
*.E
0
8.e300V
]],OOOA
2F
2.S-
35mOhm
r
qw
.
09.5
4.0"
4
a..-
.es
0.&0.
*.G0S
-6408'.0
'
.06
0.070
-E
-0
j -0.16E
T
I
T
U 0 L. T 9
0 M
M E
0. BOE
3
9. 19E
3
0.010
-0.10E
3
-e.EM
3
-0.30E
3
-6.40E
3
3
I
*00
S)
0
OOO0A
W
W.
WEW-
-
2F
I
0.030
I
0.040
300V
35mOhm
9
I
0.060
IW V.W
0.070
0.060
T
I
E
9
( S)
Figure 8f. CRITIC means that the circuit is critically damped.
At t=10ms it can continue to Figure 8i
(R=lOmOhm).
14h
OH
C U R R EN T
0.1SE
A M P)
A L
T E R N
10.0
V0 L T 8
L 0 0 P
S
8.0
.16E
5
]] ,OOOA
S
300V
6.0-
20mOhm
2F
4.00.5E
4
.00.010
-8.50E
eo
0.030
0.040
*.060
.
V 0 L T 9
0.49
3
0.38E4E
3
9.28E
3
0.10E
3
S * 01
-0.10E
3
-6.80E
3
-0.30E
3
-0.40E
-
-4.6
-
-s.*
IRNE
T
z
-2.0
4
5
-0.15E
m*
0.07
0 h
I
-1.0
)
C.
300V
]],OOOA
6.980
S
I
0..030
I
0.040
I
0.050
0.060
Im-.wow-0.070
3
FIGURE 8g.
T
I
M E
(
S
)
14i
0 H
e.ISE
C U R R E M T
CA M P)
A
5
L T E R h
10.0
V O L T S
L OO P
8.0
.10E
]],000A
5
6.0
] OmOhm
2F
*.
300V
4.0-
SOE
a.,I
0.10 e.
0.030
0*040
0
COeSS
0.070
-4.0-
-,.16E
5
S-0.0)
T-0-0
I M E
E ( ~s)
V 0 L T S
0 h
C0
0. 40E 3
300V
1,OOOA
3hm
2FI
- .30E
3
9.29E
3
0. 19E
3
ee
0 &Ampe
-0.
E
76.
2E
3
-0.30E
3
-0.40E
3
.. ea.
0.030
0.040
,.ee.
s.e
0.070
3
T
Figure 8h.
I
M E
C S
)
14j
O
H
U
C
0.ISE
S
0.1OE
5
R R E N T
C A
M
A
P)
L
L
T E R N
OO
P
2, 500A
2F
V
O L
T
S
325V
6.0-
lOmOhm
4.0-
4
0.50E
0.030
0.020
.010
0.
0.070
0.060
0.050
-2..
-0.SOE
-
4
-4.0-
-0.1OE
5
-0.1SE
S
V 0 L
-
T
S
0 11
T
I
-I0.0S)
M E
C.
3
0.40E
.
500A
0.30E 32,
2F
0.20E
3
e.1OE
3
0.020
-0.1OE
3
-0.20E
3
-0.30E
3
-0.40E
3
0.030
0.040
325V
10MOhm
0.050
0.060
T
Figure 8i. Continuation of 8f.
I M
0.070
E
(
S
)
V OT
*.4GE
S
O
C.
3
0. 3W
3
0.0EW
3
-2F
0.00
-0.30E
3
-0. 4E
3
OV
15, OOOA
10mOhm
L%.~
ed~a
e.ese
0.03e 0.040
5.075
-e.3eE 3
I N
T
V 0 L. T 8
0 h
S
0
15, OOOA
G*4E 3
e.3M
(
E
Dv
CRITIC CASE
3
-. ME 3
ttt
iF,
50mOhm
+++
2F,
35mOhm
t
0.114+
dbdb
0.08m 0.030
-..
iSE
DWte
--
-0.30E
0.40
0."80
O.S*
0.076
3
3
3
-I
T
ZF
Eis
Figure 8j. For calculating A-s rating of ignitrons.
)
-
15 -
5. Shut-off phase
In this operation the plasma is created as usual, then it
is maintained for about 15-20ms, to allow for heating and thermalization; at about 20ms the primary current is gently driven to
zero, without inducing large loop voltages; afterwards the loop
voltage is zero, and the plasma is not driven anymore by the OHT.
The OHT has been shut-off;
the plasma current and heating will be
maintained by RF waves.
Result
of simulations are shown in Figures 9a-b. They
bring out different performances obtained by charging the
double-swing capacitor to 100V. Other parameters that can be varied
are resistance and bias current.
After plasma formation, the loop voltage must be kept
below 5V, and should be about lV, when the primary current is
zero.
-
e*
(S9
)
a
w
*NOIIV~IadO JaO-LflHS 'V6 aanblTa
II
S
3SVOI-
2-
ir
30S 1
-1
- suoZ qnoqv
qBu~pue 'aspqd
wG* sIo
J;o-4nqls aq
ceo We
230*3
ST sT4JI
T"
I-
$I
see a~~
e
-
r ft~
p.
e0g.-
-
r
9.9
ur4oIo L
AO
S..
s~IL
f
u
AZ
V0L0 I
41a
d0A
4 3d W
Sv )
(
I
V0
9 39,100
S
C of
3SVS
Ho
I M 3 v ano
wv)
I
a
39S*.-
F
-else-
30T
*P92T; ST SA U914m
*02- 'su
4
bu-rpug pup 'uoT4RUxxO; VuzslTd ;10o;
A8Z BUToflpUT '9S~qd dn-4aes Oq4 ST ST4ql
* ST-
I
S307*S
e~s
s .I~f
G*Og
1 dA
00
urtto ZE*
AO
d
-
qST
V00S'
S
w v)
k
3ST*,
3 a vfno
m
H o
-
OH
C
e. 5E
5
0. 1E
5
5.
4
0.4
U R R EN T
A
15c M P)
D A
MP E D
L O O P
10.01
V O L T S
5.0
100V
0.3 2Ohm
6,500A
2F
~5.S]
Si
.S
-. "3
5.054
8.SE
0.006
5.007
*ets
qAO_es
-15.0-
;A
7
-.
j
This is the start-up phase, inducing
28V for plasma formation, and ending
-ae.. at lms when V5 is fired.
SE
easE
Capacitor initially charged
S
T
O H
C
S.I5E
5
U R R EN T
( A
MP )
D A
I
N
E
C
S
-30.0 )
MP E D
L O O P
4070A
2F
0. IE
J~
V O L T S
5.2
100V
70mOhm
5
4
This phase ends at 6ms (real time)
Ah- Tr
O H
C
.ISE
5
VF
U
R
R
E N T
A L T E R N
C A M P )
L
1, 640A
2F
0. iE
ira
5
OO P
UO L T S
105V
1OmOhm
ShUT-OFF PHASE. Ends at 15ms (real time)
5.-5.
S
i
4
.OO
S.39
.
4'
-is.0
-F.igE
4
Figure 9b.
-
f
- 16 -
6. Conclusions
.
6.1 Capacitor Banks.
Bl is specified on page 2.
B3 can be the. 2F, 350V, 122.5kJ Bank. See page 14.
B2 is specified in Figure 1. Try a lmF, 4kV capacitor
for this task, charged to 2.5kV. If the commutation
time turns
out to be long and losses are important, then C 2 can be lowered.
6.2 Ignitrons.
Vl and V2 are specified on page 3.
To calculate the A-s rating of V3 and V4 one can monitor
the change of voltage on C 3 , which is then proportional to A-s.*
For example, Figure 8d gives 200 A-s for C 3 = lF, and Figure 8j
gives 270 A-s for a 1F Bank, and 400A-s for a 2F Bank. Therefore
for V3 and V4 use 1500A-s, GL8205 tubes. V5 can be a 200A-s,
GL37207A tube. For V6 it is good to take into account V7. Once V6
is fired then V7 can be closed if it is a Ross' vacuum bottle; if
V7 is not a vacuum bottle but an ignitron, then it can not be fired
until the reverse cycle starts. Therefore, I recommend the following combination;
with
V6:
200A-s, GL37207A tube
V7: a Ross' vacuum bottle
;
Otherwise, use two 1500A-s tubes.
V8 must be a 1500A-s tube because the crowbar time can
be as long as 60ms.
*
A-s = fIi(t)Idt
= C(v(t)
-
V3 )
- 17 -
Summing up:
- Three 1500A-s GL8205 ignitrons, for V3, V4, and V8.
It is too bad there is not a tube rated between 200 and 1500A-s.
- Four 200A-s Gl37207Aignitrons, for Vl, V2, and V5, and V6.
Four Ross' vacuum bottles, one of them for V7.
6.3 Resistor.
Can be made out of stainless steel strips 5' long, with
a cross section 1/8"x2". Each strip will have a resistance of
8.5mg .
(I took 90pQ-cm for the resistivity); as many strips
can be placed on a frame as needed. Using this plate, the temperatures shown in Table I would be reduced by a factor of 4.
40 strips would make a resistor of 0.340. Neglecting cutting
losses
.plate 5'x7' is required.
6.4 Diodes, charging supplies, and other are not specified.
I
PLASMA FUSION CENTER INTERNAL DISTRIBUTION:
PFC BASE MAILING LIST
Argonne National Laboratory, TIS, Reports Section
Associazione EUIRATOM - CNEN Fusione, Italy, The Librarian
Ballelle-Pacific Northwest Laboratory, Technical Info tenter
Brookhaven National Laboratory, Research Library
Central Research Institute for Physics, Hungary, Preprint Library
Chinese Academy of Sciences, China, The Library
The Flinders University of S.A., Australia, Jones, Prof. I.R.
General Atomic Co., Library
General Atomic Co., Overskei, Dr. D.
International Atomic Energy Agency, Austria,
Israel Atomic Energy Commission, Soreq Nucl. Res. Ctr., Israel
Kernforschungsanlage Julich, FRG, Zentralbibliolhek
Kyushu University. Japan, Library
Lawrence Berkeley Laboralory, Library
Lawrence Livermore Laborary, Technical Info Center
Max-Planck-Institut fur Plasma Physik, FRG, Main Library
Nagoya University, Institute of Plasma Physics, Japan
Oak Ridge National Laboratory, Fusion Energy Div. Library
Oak Ridge National Lab.oratory, Derby, Roger
Physical Research Laboratory, India, Sen, Dr. Abhijit
Princeton University, PPL Library
Rensselaer Polytechnic Institute, Plasma Dynamics Lab.
South African Atomic Energy Board, S. Africa, Ilayzen, Dr. A.
UKAEA, Culham Laboratory, England, Librarian
US Department of Energy, DOE Library
Universite de Montreal, Lab. de Physique des Plasmas, Canada
University of Innsbruck, Inst. of Theoretical Physics, Austria
University of Saskatchewan, Plasma Physics Lab., Canada
Unihersity of Sydney, Wills Plasma Physics Dept., Australia
University. of Texas at Austin, Fusion Res. Ctr., Library
University of Wisconsin,' Nucl. Eng. Dept., UW Fusion Library
INTERNAL MAILINGS
MIT Libraries
Industrial Liaison Office
G. Bekeli, A. Bers, D. Cohn, B. Coppi, R.C. Davidson,
T. Dupree,.S. Foner, J. Freidberg, M.O. Hoenig, M. Kazimi,
L Lidsky, E. Marmar, J. McCune, J. Meyer, D.B. Montgomery,
J. Moses, D. Pappas, R.R. Parker, N.T. Pierce, P. Politzer,
M. Porkolab, 1. Post, I-. Praddaude, D. Rose, J.C. Rose,
R.M. Rose, B.B. Schwartz, L.D. Smullin, R. Temkin, P. Wolff,
T-F. Yang