Fermi lab UPC No. 134

PROPERI'IFS OF SUPERCONDUCTING MAGNEI'S*

A. V. Tollestrup

Fermi National Accelerator Laboratory

.July 2 1 1 i 9 8 O

Talk given at March Brookhaven National

laboratory Workshop on Superconducting

Accelerator Magnets

*operated by Universities Hesearch A . .ssociation under Contract with the Unih..d. States Deparb:ncnt of Energ"".f

'" - 1 -

Preparing for this talk has teen an interesting experience. I have had

a chance to go back and review many of our early data books, and it has be­

cx:xre clear that in the last 5 years, while v.-e have advanced franmaking 1 ft.

magnets to making 22 ft. magnets, there are still many questions conceming

superconducting magnet technology that remain tmanswered. In sane areas such

as systems, quality control, vacuun technology, and cryostat oonstruction, v.-e

have made great strides. In other areas such as tmderstanding why nagnets

train, we have made very little progress. The data I will use in this talk

will rrostly be taken from Ferrnilab sources, and it .involves the ~k of

many people that have been in the program since its inception. Fran what I

have seen and heard, similar data has been accunulated by \'K>rkers at other

laboratories. Now let's leak at the data:

!"'"'· ,.. Training

--

Perhaps one of the rrost unique features of su,Perconducting magnets is

the fac.t that they "train. " Projection l shows the trajning curves of 4 dif­

ferent magnets o By training I ~ Jrean the phenanena that clS the current is

ramped up in a new magnet, it will reach a point at which a section of the

supcroonduc.tor goes normal. At this point, sane means nust be found to re­

duce the current to zero or the oonductor will rrelt. The next tirre the cur­

rent is run up, the magnet will go to a slightly higher field. A gocd ex­

arrple of this is shown for Magnet ElO-lA. A magnet such as this ~ld not

be suitable for use in an accelerator. The other 3 magnets shown in the pro­

jection have sorrewhat better training characteristics. An exanple of a very

8 0

w.

c ~

,,, j!.

z

g 0 :c

I-

'

0 N

0 CH

0 ~

0 01

0 Ol

0 ......

0 o:>

0 CD

0

N

0

-2-

KIL

O

AM

PS

0 D"

--.~

. .

rl1

~

I a a

a

• 1>

----•

&

Ila

------

-.""

·;

;?'

fT1 - 0 l

~

KIL

O

GA

US

S

~

0

fT1 0 I N

. ,.

I . -

-f

_;o

o_~

_,,

rrt

!!: z

._,

(I)

G)

l'TJ

0 ~c

rn ::

:o CJ

) <

JT

1 C

J)

- 3 -

~11 behaved ma.gnet is s}x)wn in Projection 2, where it is clear that the mag­

net is essentially fully trained after the first quench. The falling points

• after the second quench is because the magnet was tested at a high B and

should be disregarded for this part of our discussion. This particular mag-

net PAB-59 reached a current of 5, 150 amps, 'thlch represents the short sample

limit of the conductor used in the construction of the magnet. If a magnet

quenches before it has reached the short sarrple limit, it necessarily follows

that sare portion of the conductor w:mt nonr.al during the rarrp. This section

of the talk will concern itself with the differences be~ magnets like

ElO-lA and PAB-59.

Projection 3 shows a histogram of the number of que.11ches necessary to

reach 95 percent of I for a series of ma.gnets that ·were constructed early max

in the Fermilab program. It is seen frcm this histogram that it is possible

to build magnets that behave in a fashion suitable for use in an accelerator.

Projection 4 displays an additional aspect of this data, namely, it shows a

histogram of the distribution of peak current in the rr.agnet relative to the

short sample limit of the wire of which the magnet 'VI.as .constructed. Projec-

tion 5 shows the histogram of the actual maxi.mum quench current rear-..lled for

the indi vi.dual magnets.

It is clear that this data represents a series of magnets that trained

with relatively few quenches to a currei.1t that approached very close to the

short sample limit of the wire.

I would nCM like to address the question of what causes a rragnet to

quench before it reaches the short sample limit and M"ly does a rragnet train,

~-4

-

•Ol

Ul

v ( S

d~'1011>f)

l.N3~~n::> H

::>N3nO

0 r-...

0 <O

0 IO

0 v 0 rt)

0 N

0

•' '

0:

w

co N

:s

§ :.:> z

·r-1 .µ

u

:c Q

)

0 ·n

2 •"'l

w

_,, :.:> C

l

-5-

NU

MB

ER

O

F M

AG

NE

TS

f\

)

I I

I I

I I

-

f\) -

z c:

OJ S} (J

J -

0 .,, ~

0 c: ~ -

"'O

111

~ 2

LJ.

('

) (!>

::t

: ()

rn

rr

I-

'· C

f) 0

\ 0

-::i

d

w

w

(J\

~

O't

0 -

H

:s: l>

x .....

,, - -

~

0 0

~ 'T

l

{I)

L.J

. :I

: ~

0 rT

:u

I-

'· -i

g ~

U>

l>

Z! .,, r- rrt

<D

0 <D

I\)

lO

O'>

(LI

(X)

0 0 0 I\)

-6-

NUM

BER

OF

MA

GN

ETS

N

O'>

,---1

4

I

en .... 3 r-- LIJ

I z (!) <(

~ LL 2 0

0: w CD ~ ::> z

5.0 . 5.1 5.2 5.3 5.4

MAXIMUM QUENCH CURRENT ( KILCAMPS )

Projection 5

( ( (

-,

- 8 -

i.e. , why does the current at which it quenches increase each tirne it is ~

quenched? It is obvious that \'le would like to id.P..ntify the mechanism for

quenching and the nechanism for training.

In Projection 6, I show a number of quantities that I have calculated

for a FNAL magnet. The horizontal scale represents tenperature, and the

vertical scale represents energy. The curve labeled cable enthalpy repre­

sents the total heat content of our 23 conductor wire as a function of tem­

~ature per centineter of length of cable. It is seen that at s° K the en­

ergy content of the cable is less than a millijoule per centilreter of length.

The slope of this curve at the o:perating temperature of 4.2° K is about .2 mJ

p:?r degree K. The critical temperature for niobhnn titanium is around 10° K,

and hence, it would take about 4 mJ of energy per centir.eter of C'..able in order

to raise the ternp?..rature to the pojnt wr..e:r:e it \\UUld no longer be supercon­

ducting. Ho\\~ve..r, when the cable is in a rLE.gnetic field and carrying a cur­

rent, a much smaller cha.'1ge in temperature will drive the cable norrra.l. For

instance, if the cable is at 90 percent of the short sarrq;>le l:i.mi t, a change

of a tenth of a degree K will change the cable fran superconducting to nor­

mal. The I 2r loss in the cable is then high enough to drive a quench wave

down the conductor. We will discuss this phenonena in rrore detail later.

Thus, when~ see a newly constructed magnet quenching at 10\'.er currents

than the short sample lim:i t, ~ are forced to look at sources of heat avail­

able for driving the wire nonnal and sources of co0ling for absorbing such

mat. There are two p:>tential heat sinks available. The first is the speci­

fic heat of the ma:trix and any liquid helium contained around the wire. The

Fennilab cable has a certain anount of open space available for helium

100

10

w _. -CD >-<( Cl.. 0 ...I

<{ E :r.: () I-,,........

"~ (J) ..._,.

w _J ::> 0 J

1.0 l'l •o

- 9 -

45 mj /cm

HEAT OF VAPORIZATION

OF ENTRAPPED He

<<<< >>>>> 1/2 Pf E

ELASTIC ENERGY OF

SUPPORT ~ .. 1ATR!X

<l.4.. ~v

~~ ~~ 0~ v~

He HEAT CAP SLOPE - 0.2 mj / ° K Cp (.1°K)

= .65 mj

2 4

FOR 95 °/o S. SAMPLE LIMIT 8T=.05°K --

6 a 10 12

T °K

Projection 6

- 10 -

penetration. This amonnts to alxmt 10 J;Xrrcent of the cross section of the

wire. I have shc'Mn on the curve the heat capacity of the captured liquid

relitnn for a 1/10° change in t~rature (this assuues no boiling). It is

eg:ual to .65 mJ per centimeter of length of the wire. The ultimate heat capa-

city fran this captured helitnn would be represented by the heat of vaporiza-

tion, and that is shown by the at:rO"w at the top of the graph, and it is 45

mJ per centineter of length. It is thus clear that the helium represents the

major heat sink in the magnet. 'Ib give sane idea of the source of energy

available for initiating a quench, I have shown 1/2 P 2 /E, which is the elastic

energy stored in the matrix. Again, I have norrralized this to the volume of

a centimeter length of the cable. The Youngs modulus used is E = 10 6

and is

representative of the Fennilab design. P2

is of the order of several thou-

sand pounds per square inch. It is clear that the energy stored elastically

is much bigger than tllB ene.t'9Y needed to drive· the wire normal when it is

carrying a current close to the short sample limit. However, I was surprised

to see that this energy is not enorm::>usly large canpared to the energy re-

quired to initiate a quench nor is it very large carq;iared to the heat sink

available in the helium.

It should be m::mtioned at this point that there are tw:> prevalent the-

cries concerning the source of the energy to initiate a quench below the

short sample limit. The first inv"Olves frictional notion of the conductor

under Lorentz forces, and the seoond theory postulates that the support matrix

cracks nnder the strain of the IDrentz forces, and the strain energy is ab-

sorbed by the conductor and drives it nonnal. That is why I sh~ on Pro­

jection 6 the elastic energy of the support matrix. I would like to stress

-

-.__,,

- 11 -

at this tine that \'.e do not know enough about our magnets to decide which

rrechanism is the one that causes quenching, and it is this point that needs

to be elucidated by much ~re research into their behavior.

I ·will now show sane information that was accumulated while developing

our magnet, and this evidence will show that, indeed, there were irotions of

the conductor in the magnet, and these were saretilres rather large.

First look at the change in diaireter of a magnet as it is magnetized.

Projection 7 shONS the radial distortion along the diameter perpendicular to

the field as a ftmction of magnet current. This change in diarreter is pro­

p.Jrtional to the square of the current and is the behavior that would be ex­

pected from elastic displacement of the support structure by Lorentz forces.

These rreasurenents were made by inventing a caliper that was read out by use

of a strain gauge. Projection 8 shows the change in outer diameter, and the

difference betv-.reen these two changes probably represents c:..urrpaction of the

winding tmder the large forces. Projection 9 shCMS the histogram of this

outer radial notion for ti.'1e sane series of magnets that we have used before.

In each case, the rrotion is very nearly elastic, i.e., the notion for the cur­

rent increasing and the current decreasing is essentially the same and is pro­

p.Jrtional to the square of the current. The value shown in Projection 9 is

for 4 kA, which represents a field of about 35 kG. The arnpli tude of the no­

tion that we have been discussing fits well with the calculated predictions

for the elastic deflection of the collars under the magnetic forces. It

should be said in passing that this type of notion, which in a synchrotron

can be repeated many tines, can cause fatigue failure in the ·collars. Tests

8.0

7.0

6.0

- s.o (/)

:?2 -,._ \!J 4.0

2.0

1.0

0

- 12 -

RADIAL DISTORTION

MEASURED BY INSIDE GAUGE

Er

PAB 59

780220 INNER RING

2 3

I (KA)

Projection 7 ·

Er= .1617 12

4 5

8.0

6.0

5.0

-(f) _J

/""'"'' ~ 4.0 -,_ \JJ

3.0

2.0

1.0

0

- 13 -

RADIAL DISTORTION

MESURED BY OUTSIDE GAUGE

PAB 59 780220

OUTER

2

RING

3

I (KA)

Projection 8

Er =.1317 12

4 5

r )

15

en ..... 10 UJ z (!) <t :E

I LL. '<1' 0 .....

I 0: LLJ CD :ff: :J 5 z

1.8

2.3 MILS

~+--1 ±.!

2.0 2.2 2.4

)

2.6

AVERAGE = 2 .3 MILS <r = 0.1 MILS

2.8 3.0 3.2

OUTER RAD,AL MOTION (AT 4 KA) [MILS]

Projection 9

' )

- 15 -

an the oollars in the FNAL nugnets origina1ly showed that they ~uld have

failed after about 1 million cycles. The present collars soow a fatigue

failure tine of over 100 million cycles. This difference was achieved by

operating the collars at a lower stress level and relieving the points at

high stress by rreans of rounding the sharp comers.

So far we have shown data on radial motion of the conductor. In addi­

tion to this radial notion, there is an azimuthal notion that causes cc.crpac­

tion of the ooil. This notion is predcminently governed by the elastic mxl­

ulus of the ooil ma.tr.ix itself. Since the coil rratr.ix is not nearly as stiff

as the collar, this notion can be considerably greater. Projection 10 shows

the azimuthal notion of the second oonductor fran the end of the coil rela­

tive to the collar. Again, in this particular rragnet, the rrotion is well fit

,,..--... by a parabola and is prop:>I.tional to r 2• Also, there is very little hystere­

sis in this notion as the arrows indicate on tb.e curve. The instrument used

to make these rreasurements has a small correction due to the elliptical de­

fo:rnation of the collar, and this is shown in the dotted curve underneath the

azimuthal motion and should be subtracted from t.r .. e top curve. This azimuthal

ccmpaction was very difficult for us to learn how to a:mtrol. In sorre of the

early magnets, the £ shown \\Ould be :rrore than 30 fills or about 1/2 a conduc­

tor width. Such large notion was no longer fit by a simple mcd.ulus, and the

nonlinear relationship gave rise to a very complicated behavior of the de­

flection• Furthermore t the Curve for increasing Current and decreasing cur­

rent oould be very different, the notion showing large amotmts of hysteresis.

8.0

7,0

6.0

5.0 -

4.0

3.0

2.0

1.0

0

PAB 59 780220

- 16 -

AZIMUTHAL MOTION

AWAY FROM KEYS

2 E = • 019 + .27 I

-----

/ /

/

/ / /~AMOUNT DUE TO ELLIPTIC

..,..,.... / DEFORMATION

2 4 5

I (KA)

Projection 10

- 17 -

Projection 11 shows an attempt to display the effect of the amplitude

of this notion on the training behavior of the magnet. The vertical axis dis­

plays the a~rage am:nmt of this notion versus the number of qllf"..nches to

fully train a magnet which is s..hCMn on the horizontal scale. 'Ib make the

rreaning of the points rrore clear, I would like to explain that for instance

the first point plotted at one quench consisted of averaging £8 for 14 mag­

nets which ~re fully trained in one que...'1ch. The size of the error bars

shONn are the dev~io..ri of this average value in the set of 14 magnets. As

long as this notion is no nore than about 5 mils, it is elastic and is ~11

fit by a parabolic notion. When it becorres greater t.han 5 mils, in general

it rreans that the clarrping IIEchanism is failing, and the collar is no longer

restraining the wire. Thus, the points at 15 quenches showing notions of

20-25 mils represrn1t unclamp=d conductors. This curve vX>uld seem to indicate

that as long as the conductor is moving elastically, the size of the notion

is not closely related to the number of quenches required to train the magnet,

but that onc:x: the conductor is no lcnger clanpad, the training .is seriously

affected.

I ·would ne."Ct like to sho.v scree data t.ha.t indicates that a fully trained

magnet really dces quench at tb..e expected point in the winding. Projection

12 shows a view of one of the coils, both as a projection leaking at the top

and a cross section looking at the end. Po.int No. 2 is the high field when

the magnet is inserted in.side of an jron yoke. Th:~ peak B at Point 2 is about

10 percent higher than the B at the center of the magnet. On the other hand,

when the magnet is tested without an iron yoke, the high field point moves

- 18 - .

\ . ....., ~ Ee ~·

w ,_...... \ I 2Z SERIES 0:: en

3= -I ,, :E 81 MAGNETS TOTAL

_J '---' /\ <( z . :I: Q t-

::::> t-

~ 0 :E

N I \ <(

I \ 25

(alp AVERAGE MOTION 1 20 0

! a 15

a 10

5

5 10 15 20 25

# QUENCHES

TO FULLY TRAINED MAGNET

Projection 11

- 19 -

311BORE

Projection 12

I- HIGH FIELD 2- POINTS

J=.Jo Sin 8

- 20 -

to the inside tw:n at the end of the inner shell of the rnagnet, sho\rm as

Po:int 1. The field at this r:oint is about 20-22 peramt higher than the

field on the axis of the coil. Projection 13 shows the relevant data for a

22 ft. rnagnet tested without the iron yoke. The straight line represents

the maximum field r:o:int :in the .coil as a function of the current. The curved

line represents the wire characteristics as detennined in the short sample

test for the cable that was used in this magnet. The po:ints along the magnet

load line represent the individual quenches that took place as the magnet

was trained. It is seen that the magnet gets to about 96 percent of its short

sarrple limit. This type of behavior is typical for the magnets that we have

l:e?rl talking about. I ~uld now like to derronstrate that for one of the 1 ft.

m::x:lel nagnets the quench did take place in the fully trained rnagnet at the

expected high field point. In order to understand these measurements, we

will have to investigate the snbject of quench· waves.

Projection 14 shows what happens to a conductor when a section of it

goes nonral. At the bottan of this projection we see a diagram of the tern-

perature along the axis of the wire when one section of it has gone no:rrna.l

and is carrying a high current. The r2r loss in this part of the wire when

it is carrying 5, 000 amps is about 25 watts per centimeter of length. This

heat generation is so high that the helium in contact with the wire is not

capable of keep:ing it in the superconducting state. Consequently, more of

the wire is driven into the no:rrna.l condition,· and a quench wave propagates

out from this region with a velocity v • The characteristics of tl1ese quench q

waves were studied in the ex.perimmt shCMil schematically at the top of the

projection. A hairpin about 12 in .. long of 23 conductor cable was :i..rmersed

I .,..c N

I

(

6.00

- 5.00 Cl) a.. ~ <r 9 ~ 4.00 -f­z l&J

~ 3.00 :::> (.)

...... w z (.!) 2.00 <t ~

1.00

·,~ ('\~ c'<~ ~-....;

~ ~ .;-:...0 ,v ~ «.~ .. '

'.-<..' -~,C>+ ... ~'

~ o~ ~ c; RFB-168

WIRE CHARACTERISTIC

2054

0 , I I

0 10.00 20.00 30.00 40.00 50.00 60.00 70.00

MAGNETIC FIELD (KILOGAUSS

( ·>:r·<?jection 13 (

:::J Heater

A

8

- ?.2 -

QUENCH WAVE MEASUREMENTS

~v

8=0 A

8 ---------·---- w ID 0 c

E

D

Region of B

NORMAL

0.: c a..

D

E 1--...

Probe response VS t

WWW

R~o T--~ >(A.2°K

12

R = 25 Watts /cm of wire at 5000 Amps

Projection 14

,·

- 23 -

in liquid helium. It had little dip::>le contacts labeled A, B, C, D, and E

located along its length. The bottcm part of the hairpin was embedded in a

high field r~ion, and the value of B in this region could be controlled

separately. The hairpin carried a certain current I0

, and a quench was ini­

tiated in the conductor by means of a heater that could be pulsed. As the

quench wave propagated down the wire, it "-Duld becane normal, and the dipole

probes connected to a chart reoorder shaved the voltage drop across a centi­

meter length of the conductor as a function of time. This is diagra.rmed at

the top right of the projection. By measuring the time of the appearance of

the wave at various probes and knowing the distance l::etween them, one oould

calculate the velocity of the quench wave and study t.'llls dependence upon B

and I 0 •

Projection 15 shows the results of ~arre of these measurements. The r-are

conductor .is sha.m as ~11 as the conductor that '\\BS insulated wit..11 mylar in

a fashion similar to the conductor u&.--=>d in the magnets. It is seen that

quench wa.ves do not propagate l:.ela-1 a current of about 1, 000 amps a11d that

t.Yieir velocity, once they start, is several meters per second and is roughly

linearly d2pendent upon the cun.·ent reing carried in the conductor. We also

oompare in this projection a single strand 'Where we have multiplied by 23 the

current in order to put it on the same horizontal scale. It is seen that the

oooling of a single bare strand is somewhat .better than that of the cable as

a whole. Projection 16 shows the oohavior of the quench velocity in a single

strand as a function of the current through it and as a function of the mag­

netic field in which the conductor is .irrrrersed.

- 24 -

• BARE 23 STRANO

e BARBER POLE SANDWICH

4.0 A DOUBLE MYLAR COVERING

(..) w en

" 3.0 en 0:: w I-w ~

I->-u 0 2.0 ...I w >

1.0

lOOO 2000 3000 4000

I IN CABLE AMPS

Projection 15

24

22

20

18 (.) LtJ

~ 16 CJ)

a: w t-w :2

14-

>-I-

/""""- (.)

0 _.J w > :I: 0 z ttJ ::>

"

0 10

- 25 -

• 02711

SINGLE STRAND !GC

SELF QUENCH= 356A

EFFECT OF

APPLIED FIELD

I ..

20

KILOGAUSS

Projection 16

l 8

30

- 26 -

Figure 17 shows the results of using quench waves to locate the p:::isition

of a quench in a fully trained 1 ft. ma.gnet. Multiple sets of dipole probes

w=re located along the conductor on eac1'1 side of the points labeled X and Y.

These probes were connected to a chart recorder. The ma.gnet was then pulsed

slowly until it reached its high field limit and quenched. The high field

:r;oint is represented by a cross about 1/2 ·way between the points X and Y.

There ~re en.ough probes located on the windi.ng so that one could determine

both the velocity of the quench wave and the point where it started. The

histogram at tl1e bottom of the drawing shows the location of the point where

the quench started. It is see..."1 that within· a few millimeters the coil was

quenching at its high field J;X)int. This experilrent derronstrates that a fully

trained magnet upon reaching its short sample limit, quenches at the p:::>int

predicted by theory.

There are other ways of derronstrating that a magnet is operating at its

short sample limit. The data in Projection 18 \\as obtained by Bill Sampson

at Brookhaven. This was a magnet that exhibited a fair arrount of training.

For these measurements, the temperature was varied. Since one knows the be­

havior of the short sample limit of the conductor as a function of tempera-.

ture and one knows where the high field point is located in the ma.gnet, one

can make a predicted curve of current versus temperature. The predicted

cw:ve is shCMn as a solid line and the measured points as circles slightly

l:elow this line. The feature that is interesting to observe is that at a

temperature of about 5.5°, the quench current no longer increases; it levels

off. It was concluded that this magnet was no longer quenching at its short

7

6

(/)

w 5 :c (.)

z 4 w :::> 0 3 LL 0 2 :ti:

0

x 6

- 27 -

QUENCH ORIGINS

x

5

I I I

4 3

POSITION (CM)

Projection 17

2

E 1-12 no Iron

I I I

BOTTOM COIL

y

4000

3000

2000-

1000

4

' '

- 28 -

CURRENT vs TEMPERATURE

0005 FORCED FLOW

5T-

'' MAGNET FORCE

' . " > MECH FORCE ·---<le~ -.. ,,;,

'- 4T-

0 ' ' MAGNET FORCE'

< MECH FORCE '

' ' ' 3T-

\_\

BNL

5 6 7

TEMPERATURE

Projection 18

' a, a, ' ' '

8

- 29 ~

sarrple limit due to the fact that the magnetic forces ~ greater than the

nechanical forces of .constraint and that the conductor was noving. It is

thought that the conductor motion was causing premature quenching of the magnet.

Figure 19 shCMS sate data obtained at LBL on sore E.SCAR type magnets.

This curve plots the variation in the quenc..11 field as a f mction of the ramp

rate. We will discuss this type of curve in rrore detail in a few minutes,

but the ramp rate induces additional heating into the magnet through eddy

current effects and causes the maxinrum B observed to decrease as a function

0 of B. The data obtained in this case show the result of increased clamping

of the conductors. The highest curve for the tightest clamping was at the

short sample limit of the WLt:'e for t-.he magnet. Again, we have a situation

where notion of the condt~ctor seems to cause premature quenching of the magnet.

I would like to surrmarize tllis situation briefly at this point. I've

shovm data on a large series of Ilkl.gnets that did not have serious training

and v.ient close to the short sample lirni t. These :magnets all displayed elas-

tic notion with little hysteresis. Once this notion be<:::Oires large doo to

lack of clamping, and there is resultant frictional motion of the co!iductor,

the.11 rragnets may take many quenches to train and scxretimes may not even reach

the shoit ~le limit. This is indicated in the tl...u examples shown of a

rragnet fran LBL and one from Brookhaven. In addition, at Fermilab we have

observed mcLny tjmes that the magnetic field shape changes during training as

well as the mechanical dim?...nsions of the magnet. Apparent! y, the process of

training results in permanent shifts in the placenent of the conductor. Once

the conductor is in sane type of stable location, the trainirlg effects cease,

and the magnet tends to go to its short sample limit.

/ )

4.6

4.4

4,2

- 4.01 <( _.J en w J-- 3.8 -!

Co CIJO t~~

I

3.6

3.4

3.2

3,0

·----- --a

)

ESD-15 8

INCREASED PRE-STRESS

REDUCED AXIAL CONSTRAINT Sep 1979

) •.,

~

W. S .G·.·

9/25/79

'-" ESD-15 MODIFIED ES CAR

B., CONTINUING TRIANGULAR I PULSE MODE

O TIME-..

.2 .4 .6 .8 1.0 1.2

s• (TESLA /SEC)

Projection 19

Better Ven1ilated

Feb 1979

1.4 1.6 1.8 2.0 2.2

- 31 -

Next let's oonsider the subject of strain in the matrix and how it ef­

fects the quenching and training of a magnet. It has been known for sare

tine that epoxy in intimate contact with the conductor of a superconducting

magnet will generate a situation in ·which an excessive airount of training is

displayed. Projection 20 shows an extrene case in a magnet constructed at

FNAL. The conductor for this magnet was the sarre as it was in the series of

magnets that showed rather little training. Ec:Mever, it was spray coated with

a ver:y thin film of epo>..y before the magnet was constructed in an attempt to

solve a tum-to-turn short problem that \\e had at that time. This was a 1 ft.

magnet, and magnets in this series eould generally be expected to train in no

rrore than 10 quenches. It is seen that this magnet trained slowly and that

it never reached the short sample limit. The dips in the curve are caused

by ramp rate tests and should be disregarded in thi.s discussion. This experi­

ment was repeated on several identical It'.agnets with similar rE~sults.

Figure 21 shCMS an atterrpt to tmderstand whether or not the cracks in

the supi;:ort matrix of a 22 ft. long magnet were actually associai:ed with its

training. The horizontal axis is the current through the rragnet on any given

cycle. The vertical along each curve shows the noise pulses as detected by

a microphone connected to the magnet structure. For instance, the first

curve shavs noise pulses up to a current of alxmt 2, 500 arrps at which point

the magnet quenched. The next cycle shows that the noise pulses "Here absent

until the current of the first training cycle was exceeded, at which point

noise again appeared. This could re interpreted as the support matrix cracking

and failing as higher and higher force levels in the magnet are reached.

0 c: f11 z (")

::r: =if:

"'Cl a w.

~ rt I-'· g

"' 0

0

N 0

(JJ • 0

~ 0

c.:n 0

(JJ

0 0 0

-32-

AMPS· ~ 0 0 0

I (JJ N A

cC

< RAMP RATE

~RAMP .... OI 0

RATE ..,.

.... OI

0 O'I

(JI

0 0 0

~ 0 =" cC

r1"I

I (JJ

- 33 -

HELLUM MICROPHONE

NOISE EVENTS IN El-6

0 ' ~ ' I _J

I I

~ ~ ':f' : tJiiff1

-----· --·" ·~ 0 1000 2000 3'()00 4000

MAGNET AMPS

Projection 21

- 34 -

Finally, on the 30th quench, the magnet was fully trained, and it is seen that

the noise pulses have completely disappeared. This type of eA~riment is

suggestive that cracks in the support matrix are contributing to the training

process but as a defi.."1itive experirrent, it leaves saoothing to be desired.

There is yet another parameter that can influence the quenching of a

coil. This situation arises when a correction coil is placed inside of a

dipole or quadrup::>le mcl.gnet or when several correction roils are VX)und together

on one m-:mdril. A coil then finds itself im:rersed in the field of a second

coil • Thus, the maximum B of the coil is not detennined by the current through

the coil itself but by the current through the ooil plus the applied external

field. Projection 22 shows the results of measurerrents nade at Fermilab on

a dipole correcticn coil intended. to be used inside of one of the quadrupoles.

These elerrents are used for tr.irrming the orbit and Ca11p8I1Sating for errors

in the min bending dipoles. '.rhe curve across· the top of the figure repre­

sents a short sa...tple limit of the v.i.re. The vertical axis is the current

through the coil, and the horizontal axis is t.l:le magnetic field at the high

field point. The various curves that have shifted parallel to each other re­

presents the effect of an applied extemal field. Projection 23 shows the

tra:in.ing curve of this correction coil. It was first started off with zero

applied external field. and trained up to about 140 amps. An external field

was then applied at Quench No. 20, and it -was necessary to again train the

nagnet. These results will be presented in rrore detail in the correction

coil session. I want to :trention them here because it represents a difficult

training problem that inust be solved whenever the exte:mal field in which a

coil is irmersed is changed. This action may require retraining the coil.

200

150

--H

100

50

- 35 -

TSDt in TQ 9

The Numbers refer to the Quench Number

SHORT SAMPLE ' ~LIMIT

2

B(T)

Projer..::tion 22

r--4

~

8 u. !i) 0 rt I-'· 0 ::l

""' w

0 c: rrl z n :c

z c s: ro iTl ;.o

0

I\)

0

(Joi ·-0

~ 0

Ol 0

())

0

...,, 0

0)

0

..(!:.

0

Ia= 2500 IQ=4:?.50 G

en 0

. 0 c ·· IQ=2500io •

I.Q::42.50° l e • •ra::2500 { 9 •

{· . 4250 : 25oo{8e

2500{• 4250{·.

l e

2500{. • • • 3000 {·

• • • • 3500 {. 0

4000{ ••

4250{ ~ •

3500·

-36-

I (AMPS)

00 0

IQ =O

0 0

IQ :::JOOO 0

1000 u

IQ=O { •

•

N 0

(")

0 ::0 ;o ITI (') -t --0 z (')

0 r -I ;o )::>

z z G>

-

- 37 -

,........ The problem can be enhanced if the relative directions of the fields can re-

verse because of the desired correction. It "WOuld be very useful if \\e could

learn how to construct a coil for this type of ai;plication with a minimum

arrount of training. The fact that the training recamnences when the forces

are redistributed w:>uld suggest that in each case the wire is finding a new

equilibrium position.

So far v.e have beP...n discussing quenches observed when the magnetic field

is changing very slowly. I-bwever, a synchrotron .involves a rapidly pulsing

magnetic field. The pulsing of the field can induce eddy currents in the

coil and cause heating. This heating can cause the coil to quench at a lower

• current than is observed when the I is very slow. We look briefly at the

theory of this process.

Pi.-ujection 24 shows why a pulsing rragnet heats up. First of all, there

at:"e persistent currents that flow ]11 the volurre of tlY~ s~rconductor itself.

This generates a nagnetization M ~r unit volune, and this ffi3.gnetiza.tion is

an open loop. The first C-"'Urve sravJB such a loop of magnetization as a function

of B, and it is observed that the energy loss per cycle is proportional B • max

F.ddy currents can also exist in the matrix. Jn this case, v.e find that . . the EMF is pro};X)rtional B and as the bot tan of the figure shows , t..l-ie Joules

' per cycle in this case are pro!X)rticnal to 1inax times B. Projection 25 shows

• the behavior of the Joules per cycle as a function of B holding B fix and max

• as a function of B hold B fixed. By measuring these curves individually, max

- 38 -

• 8 INDUCED QUENCHES

pulsing a magnet heats it !

I- M

rr-

B

Jj . Cycle-· B max

Eddy currents in matrix

• Emf-8

2 Power - (emf)

R

92 -R

J/Cycle =Pt= a: BB to R

Projection 24

B, I

• B

~c

- 39 -

TOTAL • J /cycle r.J C( B M_AX + b B MAX B

BMAX FIXED

HYSTERESIS ,..,, 8 MAX

ti

8

Projection 25

B MAX

- 40 -

we can separate out the hysteresis loss from the eddy current losses. Pro-

jection 26 shCMS the actual quench de_pendence of two different m:i.gnets as a

• function of B. The 22 ft. magnet shows a much bigger effect of pulsing than

the 1 ft. rnagnet did. This effect is not understcx:x1, but these curves amply

derronstrate the behavior of superconducting magnets under pulsed field rondi-

tions. Both of these magnets -were fully trained.

Our early magnets were constructed of a cable where the 23 strands -were

individually covered with Stabrite, which is a silver tin alloy. The con-

ductors ~.re in rather intima.te contact with each other, and the eddy current

losses in this matrix were quite high. In fact, the losses were so high that

steps had to be taken to reduce them. We investigated coating every other

strand in the cable with an insulating covering. This insulation consisted

of ebclnol, ·which is a copper oxide ooating applied to the outer copf"'.?r jacket

of the strand. Projection 27 shows a ccrnparison between two 22 ft. magnets,

one of which was const..ructed with Stabrite coated cable, and the second used

a cable ·which we call zebra conductor where every alternate strand is coated

with el::x:mol. It is seen that in each case at fixed B, the losses are propor­

tional to B , and that zebra t~ conductor has considP-rably srraller losses. max • Projection 28 sh<:Ms the behavior of these losses at fixed B at varying B.

0

The intercept at zero B represents the hysteresis loss in the conductor and

should be the sane in the two magnets. It is seen that the eddy current losses

in the Stabrite are large, and that the zebra cbnductor virtually eliminates

this source of energy loss. One might expect a considerably different de pen-

dence of the quench current an the ramp rate between these two types of mag-

nets. However, Projection 29 shows essentially the same behavior for these

·,, ) J

TYPE 5 COLLARS

50

CJ) CJ) :::> 40 <! (.!)

g ~ -

30 I \ E 1-24 b 0

11 MAGNET _J

w tJ..

w Q::

I \ 221 MAGNET 0 .;:. £D i-'

::E 20 ::)

::E x <(

:E

10

0--1-~~~~~~---~~~~~~~--~~~~~~--~~~~~~---

1.0 2.0 3.0 4.0

I .. ·'; RAMP RATE (.KILOGAUSS/SECOND)

.·,.

Projection 26

2000·

1000-

-w 500 _J (.)

>-(..)

" -::> =~»00-·1

(f) Cf) 0

200 _, -

100

500

- 42 -

A. C. LOSS .

B max

e

o STABRITE

• ZEBRA

. 5 I. 0 2.0 3.0 4.0 5.0

CURRENT (KA)

Projection 27

-l.&J _J (.)

>­(.)

'-... ...., en en 0 _J -

1000

500

0

- 43 -

RAMP RATE DEPENDENCE OF A. C. LOSS

-r-100 200 300

o STABRITE

• ZEBRA

400

RAMP RATE (A/Sec)

Projection 28

• 4000

-~ 3000 -I-z w 0:: a: ;:) (.)

:r. 0 z w ::::> 2000 0

1000

0

- 44 -

RAMP RATE DEPENDENCE OF QUENCH CURRENT

0

• • 0

0

•o

o STA.BHITE

• ZEBRA

100 200 300 400

RAMP RATE (A/ Sec)

Projection 29

....,.,,

- 45 -

U\lo cases in spite of the considerably different levels of power generated

in the winding during ramping.

Questions to be Ansv.ered

Fran the preceding discussion, it should be clear that the behavior ex­

hibited by superconducting magnets is not \\ell understood. We have sane ~ri­

cal answers that are enabling us to produce successful magnets, but a science

has not yet emerged fra.'11 an art. I list sane questions that must be answered

cy future research (Projections 30 and 31):

1. Where do quenches occur during:

a. Training

• b. High B

It v.uuld be great to have an awaratus that v.uuld localize quenches

spa.cially. As far: as I'm aware, the measurements that I displayed

on the 1 ft. magnet are the only ones that have pinpointed exactly

the source of single quenches in a tra:i.nGd magnet. We v.ere ura.ble

to apply this technique to a magnet during training. The quenches

did not take place at the high field p:>int.

2. ·what is the source of energy to start a quench?

a. Is it wire friction?

b. Is it cracking of the support matrix?

I would like to maJce a ccmnent on the question of the supi:ort matrix.

The elastic energy stored is equal to:

P .E. = 1/2 Eb.x2 = P2 /2E

- 46 -

where P is the pressure and is a function of B for a perfect coil

and l:x is the arrount of elastic strain that nrust be placed in the

coil in order to collar it. If \Ye had perfect coils, it is seen

fran the first version of this equation that the potential energy

'WOuld be decreased by increasing the Youngs m::.xlule of the supp.:>rt

matrix. Hc:Mever, for nonperfect coils, the /iX. is the more pertinent

variable because the collar determines the coil size, and the d.is­

placEmE"..nt Lx is detennined by the accuracy of the construction of

the unoollared c.Di.l. If the coil is not accurately ma.de, Lx rna.y

be much larger than that necessaiy to provide the forces for conduc­

tor constraint. In this case, the second version of the equation

shCMs that the elastic energy is increased by increasing the Youngs

m::dulus.

3. Cooling - how does it affect the quenches and tl1e training'.?

A coil that i.s near 4 .2° K ha.s very high heat conductivicy due to

the copper in the cable. However., the heat capacity is proportional

to T3 and is very small. This means that the relaxation tirres lo­

cally observed will be exceedingly short. The dynamic tra11sfer of

reat from the wire to the cooling m2diurn could be instrunental in

changing the nature of the training. The -work that has been done

an using HeII at one atrrosphere pressure as a cooling fluid at

Sa.clay and at Berkeley is suggestive of interesting effects that can

re observed at these lower t~ratures. It could be that regard­

less of the material that future coils are made out·of, it w:>uld be

advantageous to cool them with HeII.

- 47 -

QUESTIONS ON TRAINING

I - EPOXY /MYLAR

NEED TO UNDERSTAND ROLE OF THE ENVIRONMENT

2- QUENCH FINDER

WHERE ARE QUENCHES DURING

a) TRAINING

b) B0

3- SOURCE OF ENERGY TO START QUENCH

i) WIRE FRICTION

ii) SUPPORT MATRIX

DX= P /E p2

P. E. = 2E = I 8 2 -- E X 2

P. ONLY FUNCTION OF B FOR PERFECT COIL

p ..... 8x FOR ERRORS IN COIL

4 - HOW DOES COOLING EFFECT QUENCHES 7

k hi, C ..... T 3 RELAXATION TIME SHORT

5- SUPERFLUID He ?

6- EFFECT EXTERNEL FIELDS (CORRECTION COIL)

Projection 30

- 48 -

7- WHY DOES EBONOL / STABRITE WORK

8- MORE UNDERSTANDING OF RAMP RATE EFFECT

Projection 31

- 49 -

4. What is the effect of the support niatrix and what role does mylar

around the cable play in the training processes for a niagnet?

Epoxy has been observed to be very bad in certain cases. In other

cases, it seems to be good. We need to understand the ef feet of

plastics that are in contact ·with t..11.e superconductor.

5. The ef feet of external fields ui::on the training of coils.

6. We need research on controlling eddy current losse~ within the

winding. At Penni.lab, WE! have care up with an em,E:erical solution

involving el::onol and Stabrite. OUr investigation of niagnets oon­

structed of pure el::onol insulated cable indicated that the current

sharing among the stt'ands was inhibited by this much insulation, and

sane of the all-e.bonol rna.gnets perfo:rned very poorly, reaching per­

haps only 80 perc~t of the short sample limit. On the other hand,

sane of the rnagnet.s perfm:m:ed very ~11. Why does ebonol saretimes

work and sometimes not? It was observed also during this series of

tests that a layer of Kapton down th~ center of the Rutherford cable

reduced the hysteresis losses. The hysteresis losses in these mea­

surements include not only the real hysteresis of the oondu.ctor, but

also the effect of any frictional forces inside of the magnet or

conductor. Perhaps Kapton was reducing these frictional losses.

Additional research on this subject oould lead to a significant

savings in refrigeration.

- 50 -

7. Fmally, a subject on which I have not spent any ti.ire and which is

very important to understand in rrore detail is the question of radia-

tion from the nachine inducmg quenches in the winding. Sarre data

has been accumulated on the subject at Brookhaven and at Fenn.ilab.

Magnet Quality and Construction Techniques

Projection 32 and 33 show the equations that determine. the nagnetic field

m te:rms of the current distribution in the winrung for a current shell. '1.'he

vector potential has only a z canponent, and ir1side radius r = a r.as terms

that are proportional to rr1. The coef f icic:.nts ~ are determined by t11e cur-·

rent distribution. M = 1 is the dipole field and involves the integral of the

two curves shown along the right hand side of the figure. The quadrupole is

EY..JUal t:o zero by sy.rrrcetry. The sextupole can be made to vcmish by stowing

the winding at the 60° point. The octapole is zero by symmetry, and in the

case of the Fermilab magnet, the dectapole can be ma.de to equal zero by.bal-

ancing the inside coil angle versus ilie outside coil angle. Projection 34

shcws the quadrupole field at the top and the skew quadrupole field at. ti1!e

rot tom. These fields will be generated by errors in the winding that cause

a lack of syrnrietcy. Projection 35 shows the results of an exact calculation

by Stan SI1CM7don for the J Bell through the Fenn.i..lab rragnets. Out to 1 in. ,

the field integral is constant to arout one part in 104 fo~ a perfectly can-

structed rragnet. H~ver, it is very important to recognize that the magnets

that ~ are dealing with are different than the ones that have been used in

the past. The field in each rragnet is uniquely detennined by the current

distribution in that magnet and not by the shape of an iron oonductor that

- 51 -

SINGLE CURRENT SHELL

r< a

Az= ~ Am (~f Sin mB

Az = ~ Am(~f Sin mfJ

Am = ~~~o ;:71" K2 (8)

m = I: DIPOLE

m:2:QUAD.

a:Q BY SYMMETRY

m = 3 : SEXTUPOLE

:: O IF 80 = 30°

m = 4: OCTAPOLE

= 0 BY SYMMETRY

m = 5 ~ DECAPOLE

Sin m8 dB

FIRST TERM NOT EQUAL TO ZERO

Projection 32

Tf

- 52 -

TWO CURRENT SHELLS Kz

82

e, .,, I _ _J

TOP COIL

current current into

out of paper paper

NOW HAVE TWO ANGELS e 0 e I MAKES IT POSSIBLE TO· ELIMINATE

m = 5 DECAPOLE

m = 6 0 BY SYMETRY

m =7 1st NON-ZERO TERM

Projection 33

8 = 12° I

8= 48° 0

COIL

...,,

I "'

2 7T 8

.;. 53 --

y

-

/

/

Projection 34

x

NORMAL QUAD

SKE\V SQUAD

- 54 -

EXAMPLE OF 5' MODEL CALCULATED BY S. SNOWDON

J Bdl. THRU MAGNET

X f Bdl. NORMALIZED TO I AT X = 0

INCHES 0 1.00000

.3 1.00001

.4 l. 00002

.5 1:00003

,6 1.00004

.7 I, 00005

.8 1.00002 -·

.9 .99992

1.0 .99965

I. I .99906

I. 2 .99789 ---

I. 3 .99564

1.4 .99134

1.5 .98267

Projection 35

- 55 -

can be replicated easily by means of precision stampings. A little calcula­

tion shCMS that in general the conductors nrust be positioned to an accuracy

of the order of 1/1000 of an inch. It should also be clear that the accuracy

of the envelope is detennined by the collaring structure that supports the

coil. We are thus faced with a new problem in rnagnet construction, narrely,

v.ie can get a given field provided v.ie can h:Jld the placement of the conductor

to a sufficiently small tolerance. The first impulse is to try t."1e solution

shown in Projection 36. I call this approach the Old W::>rld Craftsrnanship

approach. Each coil is constructed with ve.ry high precision and very careful

control. Maxwell's equations then guarantee that we will have the proper

field. HCMever, I feel that this is an analog ccrnputer for solving t.hese

~tions, and it does not represent a viable solution for mass prcxlucing

superconducting rnEtgnets for. use in high energy machines. Rather, ve need to

apply nod.em control techniques to this problem. Figure 37 shews a block

diagram of the new type of approach. In this case, there is a factory whose

basic function is to take in all of the raw materials and produce a finished

coil in a fairly reproducible fashion. We then measure the output of this

factory using a wa.nn bore magnet testing tech'1ique and feed the errors back

into the factory in order to cancel out the noise. This is the system that

has been developed at Fel'.ITii.lab. At present the feedback lcop is not closed

on an individual ma.gnet. The test is made at the point when the coil is col­

lared when about 1/3 of the cost has been invested in the rnagnet. If a coil

is out of tolerance, te roan te:rperature test will indicate the trouble, and

the collars can re renoved and the coil recollared. H~ver, at present we

are trying ver:y hard to close the feedback lcop with zero delay. The theory

of this is explained in Projection 38.

- 56 -

OLD VJORLD

CRAFTSMANSHIP

DO i = I, N

RESULTS

I TO

TEST

JJ. ANALOG COMPUTER

FOR .SOLVING ~

'\/. B=O

Projection 36

,,...........

- 57 -

SCRAP HEAP /

oops!

ERRORS ------.

11 FACTORY 11

ECRYOSTATJ I /DAY

WIRE

\A/ARM BORE

INSULATION

SUPPORT

STRUCTURE

. .------..

MUST NOT H.4VE

TOO MUCH

11 Rms NOISE

11 c-. -=i 11 GOOD M1~GNET II

CORRECTION SIGNAL

ANALOG COMPUTER FOR 11 SOLVING GOOD MAGNET 11

PROBLEM

Projection 37

TEST

- 58 -

I \<X)Uld like to corrrnent a little bit on the philosophy that has been

evolved for coil construction. Rather than controlling with great accuracy

the shape of a coil, we have put a great deal of effort into building a na­

chine that will make coils in a fairly reproducible manner. It does not mat­

ter if the coil is exactly round or if it is elliptical. Any nearby nagnet

shape will have an expansion in field ha.rm:mics of the type shavn in Projec­

tion 32. The high harrronics will be mainly determined by the comers of the

coil blocks. The magnet constructor has very little control over these things.

The l~r ha.monies are detennined by slow variations in the coil compaction.

Figure 38 sho.vs a perfect coil as a rectangular block and an actual conductor

distribution that might be obtained out of such a production machine. The

short term wiggles in this curve only C.'Ouple into the high ha.monies in a

rather insignificant fashion. However, the gross variation oouples into the

lower 5 field ha.nnonics and causes these coefficients to fluctuate. In fact,

experience has taught us that the quadrupole, the sextupole, and octapole

terns are the place where one has t..11.e main problems with field quality. How­

ever, the integrals that deter.mine these coefficients are not only dete:rmined

by the distribution of current over the block but also by the angle at which

it is terminated. If one could measure these lower ha.rm::mics and then shim

the individual coil blocks to have individually detennined angles, one could

correct the lc:Mer ha.monies of the coil for any inaccuracy within the b:rly

of the winding. The Fennilab coils have just this feature. First of all,

we use the roan te.rnp::rature measurement to emperically determine the distri­

bution of wire in the lxrly of the coil, and, secondly, we can control the

- 59 -

correct J Kz cos n 8 d8

by ad,j. upper lim1tsf

ELEMENTS NEC. FOR CONTROL LOOP

I- MEASUREMENT

2- CONTROL POINT. i.e. SHIMS

Projection 38

- 60 -

field by inserting shinlS as is shown in Projection 39 at the keys of the col­

lar. There are 8 points where shims can be inserted and, consequently, there

are 8 hanronics that can be corrected in the coil. At present, we are con­

structing a temr;:orary collaring apparatus that will allow us to measure the

normal and skew, quad and sexb.lfole manents of the winding. This process

will only take an hour or so and then individual shims ca1 be placed in the

rragnet at the points indicated in order to individually correct a coil. When

this system is in ope.ration, it is anticipated that we will control the field

to a few parts in 10 4•

There is a systematic source of field errors other than just errors in

the winding. As m::mtior1ed previously, the forces in a coil are very large.

'l""hese forces cause conductor notion a11d so a coil, even with perfectly placed

oonductors, would have a field that is dependent upon currE>.nt. Projection 40

shCMs the forces on individual conductors in one of tlle Fennilab nagnets.

In addition to the forces shown in cross section, there is an axial force of

16,000 lbs. which lengthens the coil by about .07 in. during pulsing.

Projection 41 shaws the azimuthal force as a f U1ction of conductor nu:n­

rer for the inner and outer winding of our magnets. It is seen that the

force is alnost linear with conductor number .. The force is largest at the

high field near the top key of the inner winding. Let us now calculate the

effect of this force on the conductor. To do this to the accuracy required

at this PJint, does not involve using a big computer and a "finite element"

stress a'r'lalysis program. We can get a good feeling for the problem by the

following simple m::x:lel.

- 61 -

,__SHIM

D COIL

ADJUST SHIMS TO GET PROPER FIELD

IN PRODUCTION iv1AGNET

Projection 39

42

- 62 -

,,, /

\ ~ / "/

I / \ ./

45 kg

----

80

:JI: __ _...,..71

2950*

NET/QUARTER

FORCES AND FIELDS IN .A

SERIES E MAGNET

Axiol Force=(}~~ ( 12

Ll2

)

= 16000 :fl:

Projection 40

- 63 -

Projection 42 shows the winding as a series of springs. The one end is

at the key and is fixed, and the other end is the centerline or the midplane

of the magnet winding and is fixed by sy.rmetry. The forces shown as Pi are

linear azimuthal forces shown in the last figure, and the spring constant

represents the canbined elastic constant of the cable plus the insulation

matrix. We can linearize the problem and treat the springs as though they

had a fixed constant k. We will treat Pi as proportional t.o i, the conductor

number. The difference equation can be solved, and it gives a displacerrent

lixi as a cubic function of the conductor nurnter. The max.linum displacarent

occurs a little bit past the center of the winding.

However, the main i;x:>int for solving this problem is to find out the pre-

load necessary in the spring in order to keep the end of the roil near the

key in contact with the key when the coil is magnetized. 'l'his force for our

magnet turns out to be about 1500 per linear inch of length of conductor.

The resulting notion for a Youngs rrodulus of the matrix of about 106

lbs.fan

is about 2-1/2 mils maximum and is shown at the top of Projection 43.

In addition to this elastic notion of the winding, there is also an

elastic notion of the coil collars. Th.is is shCMn in the bottcm of Projec-

tion 43. The circular collars distort into a slightly elliptical shape. The

diagram shown is IIn.lch exaggerated. The p::>int at 45 does not change in ra-

dius but rroves in azimuth. The p:>ints at the pole and the equator do not

nnve in azimuth but nove in a radial direction. We thus have altogether three

notions of the wire that canbined to make field errors. These notions all

·al 2 · th. ·ua are prop::>rtion to B in eir magnit e. The amplitude of them at 40 kG

- 6tl -

100 F9 INNER

SHELL

.t::: 0 ~

"'-.. en

..0 _J

F 8

50 OUTER

tN

10 20 30 40

CONDUCTOR NUMBER N

Projection 41

- 65 -

8x i+I

i = N

KEY

p. I

Linearize : Pj =O:i

p. I I-

Spring hos lineGr con::>tcmt k

8xi+l -2 8xi + 8xi·-I ::: rt..;

Solve

= - CC. ( i-1) [(i+l)i ·· (N+!)N] 6k .

MAX Xj at Lm = ~-,-/-3-(N_2_+ ___ N_+_I )---iP gives X max

GET Prelood from spring at end

F = k 0 X N - I = 1/3 a:( N - 2) N ~ 1500 #

Projection 42

i = I

- 66 -

------ --- 2. 4 mils.

ELASTIC MOTION OF

COIL

KEY

10 20 30

""COLLAR MOTION

0 X 1 Er 1

Eg make FIELD errors

Sf! - 12 B

Projection 43

40 conductor Number

radial, EB= 0

-- 67 -

is shCMn in Projection 44. c8 and er are due to the oollar notion. The

direction of these curves can be understcod as follcws: ce compacts the

winding rcore. to the equator and, hence, makes the field stronger as one goes

tcMard the winding. er rroves the winding away fran the ax.is and, hence, causes

the field to fall off. The detailed mechanics of the collars oonnect £ r

and Ee. If the collar was a unifonn ring, Er would be twice Ee. The sum of

the above tenns is shown in the bottom of a figure and for a corriparison, the

arrount of sextUFOle necessary for chrorroticity oorrection, if it were distri­

buted uniformly throughout the diJ?Oles, is shown as a .dotted curve. This is

only shCMn to give one sane feeling for the magnitude of the error fields

involved.

The :point I would like to make here is that there are many reasons that

r, oorrection ooils are necessary in a superconducting machine. The persistent

curre..nts must be corrected, arid their TIB.jor component is a sextur:ole term.

In addition, as we have just se=>...n, conductor notion causes a distortion of

the winding during pulsing, and finally, the residu..'11 of the iron yoke nu.1st

be corrected, and a clrromoticity tenn must be added. They must also provide

oorrection fields over the whole range of magnet operation. This indicates

that the correction circuit problem needs much rrore attention in a supercon­

ducting rna.cliine than it has received in past machines.

I would nCM like to address a problem that has caused and is causing an

eno:rm:ms arrount of difficulty. This has to do with the subject of preload.

Whatever scheme is used for confining the coil, thE:: mechanical forces must

exceed the magnetic forces or the coil will rrove away fran the support fonn.

Once this happens, the quality of the magnetic field deteriorates rapidly.

- 68 -

50 Mt,GNETIC STRAIN ERRORS ,..., 12

40 ...,,,,,

30

20 8; ..

'•''

10 ..

1.0 .;.10 R

-20

.-30 Cf) Cf)

-40 ::::> Er= . 005" <{

<..!)

~ -50

Ee= . 00311

-60 . O; = .·002n f -·-~l

I= 5000 Amps

8 "' 40 KG / 0 /

No IRON / /

10 / 0

SEXTAPOLE -4 -2 - .,/

8 = 3XIO in / ;r

,..., AMOUT REQUIRED FOR /

6 CHROMATICITY ,,R>/

/

4 CORRECTION ft SUM / OF

ft 2

.,, ABOVE TERMS .d'

.,.()",... --0 --.5 1.0

R INCHES

Projection 44

- 69 ~

In addition, we have seen in the first part of t.he discussion til.at the situ­

ation can lead to excessive training or the coil not reaching its short sam­

ple limit. ~owever, there is a nasty reality that must be faced here. The

coil when it is cold shrinks away fran the collars. Thus, the stress in the

coil at room temperature is much bigger than it is when it is cold. Projec­

tion 45 she.MS a typical situation that we have encountered at Penni.lab. The

coil size is shQ'i..Vl1 along the horizontal axis, and the force applied is shown

alo.'1g the vertical axis. The open loops represent the stress strain diagram

for the coil matrix. Whf>.n the coil is cooled, it shrinks relative to the

collar and winds up at the point indicated on the drawing as "preload." This

preload must be greater than the magnetic forces or the conductor will mJv-e

aJNay from the confining collar. Thus, one e:.m see that the rrore coil shrinks

relath'B to the collar, the. rrore it must l:::e compressE..."Ci at rocm tem:t;:erature

in order that the proper preload ·will exist wh€11 it is cold. The det:iils are

shown ill Projection 46. There do.:?sn't have to be a solution to this problem

if the coil shrinks too much relative to the collar. It could easily happen

that the ffi3.trix crushes before a sufficient pre.load at room temperature could

be applied. We do not know haw to control the m_cchanical properties of the

insulation matrix, and research neio~ds to re done on this subject.

Projection 47 and 48 shew sorre of the areas that \>Je need to investigate.

gatigue Stability_

1. Coil support. We do not know how to build a satisfactory SUf>port

structure for a 10 Tesla coil. Much work needs to be done on this

subject.

2. The mechanical properties of the insulation and its aging under

many magnetic cycles is not understood.

>­•. J :r: I··-

3000

2500

::> ::?; - 2000 N <(

J: u z ......... 1500

* 1000

- 70 -

COIL

Thermal contraction

1-

500 Pre!oad

0 15 30 45

E =GAP

Projection 45

60 . 75

X 10-3

INCHES

- 71 --

PRELOAD PROBLEM

I - Coil· shrinks

2 •Collar shrinks

If I > 2 Lose preload

If 2 > I Gain preload

FORCE

DETECT Preload = 0

COLLARS SS Gloss epoxy coils

SIZE

b2

, b4

change with 1

Projection 46

COLD

- 72 -

3. The cryostat is also subjected to ooth thennal stress and magnetic

stress. Work needs to be done on this subject.

4. It is not knCMn whether stainless steel at high stress, low tempera­

tures, and high fields is stable.

Matrix

5. At Fe:rmilab the support. matrix has sane of the properties of styro­

foam, na:rrely, it crushes inelastically after a certain force is

applied to it. It behaves elastically around the penranently de­

fonned p:>sition. This leads to easy assembly of the coils. This

solution is probably not applicable at the 10 Tesla level.

6. One can linagine strength=>_rLing the support matrix by lirrpregnating

the epoxy with aluminum oxide or glass. We tried this as an exr;:eri­

irent once, and it leads to an exceedingly stiff roil. Glass may be

necessary if one goes to Nb3Sn conductor.

7. In addition to supplying support, the matrix has to provide adequate

insulation and guarantee that there are no turn-to-tum shorts.

This represents a wide area that needs study.

Cryostat

8. Cryostats need nore research on them. The mechanical support that

the cryostat must provide is such that the magnet vertical axis needs

to be stable to a mrad, and the X and Y position of the quadrupoles

stable to 20 mils or better. This, in spite of rather large rrechani­

cal defonnations. The Fennilab magnets shrink about 3/ 4 in. between

rcan temperature and liquid helium temperatures. The behavior of

superinsulation in a large cryogenic system needs to be studied.

Hew stable is the infrared reflection coefficient of superinsulation?

- 73 -

9. How does one counter the thermal stresses in a cryostat?

10. And hCM does one find leaks in a large system?

11. HCM. does one rapidly wann up a system and change a rnagnet?

12. And the question of high current leads needs to be studied :rrore.

Quality Control

13. Finally, the question of quality control needs to be studied in much

:rrore detail. ~ have made trerrendous strides in this field, but we

have consistently underestimated the care that must :te exerciserl to

build a sur:erconducting rnagnet.

Sys~

Hew does a large system of magnets behave? At Fei"Inilab, we have three

areas where the behavior of strings of suµ;rconducting magnets is being studied.

The first is the program at Bl2 which is above ground and ~nvolves a str.ing

of 16 m.'lgnets and their associated quadi::ur-oles. It rep:cese.nts as closely as

p:issible the conditions t.'1at will be found in a. string of rna.911ets in the tun­

nel. Quench protection techniques are being stud.ied. Failure :rocx:~es of sys·­

t.Ems of magnets will also eventually be studiecl. in this area. So far the

string has been pulsed to about 3,000 amps and quenched safely. 'I'he pressure

rise in the cryostats under quench conditions and the bt-:havior of the cooling

system for the magnet string is also being in\estigatc<l.

There are also tw::> strings of lffignets installed in the tunnel. One has

:teen ccx:>led for rrore than a year, and the second one is just now stal."ting to

CCX)l da-m. What problems can arise in a system of magnets?

First of all, each one of our :magnets stores about 350 kJ. If a magnet

quenches, its terminals are shorted by an SCR, and an internal heater is

fired in order to drive the magnet nornal as fast as FQss.ible. The energy

- 74 -

RESEARCH

FATIGUE STABILITY

I T'. t F - · 82

- J.fOn supp or · -

2 - Insulation

3 - Cryostat (Thermal cycles) (Magnet Forces)

4- Stainless Stee I

a) Hi stress

b) Low Temperature

c) Hi Fields

MATRIX

IOT

6X FORCE

\\ &. II a- Styro1oom Easy essembly

Self correction

b- Al 2 o3 + epoxy

Glass

What about N b3 Sn

c- Electrical insuloiion

CRYOSTAT

I- Mechanical Support

EJ- I mr _._. '· ,.,

. II

x, y - . 02

2 Termal Stresses

Do es i~ extrapolate To high field

Projection 47

- 75 -

3- Leak testing !

· 4- Hi Current lead.

Manufacture

QUALITY CONTROL TECHNIQUES

Projection 48

- 76 -

fran the nonquenchin:g magnets is extracted and dissipated in an external

resistor. Brookhaven and Fennilah have evolved two different scherres for

protecting their magnets. In each case, the full energy within a magnet is

dissipated in the conductor.

To urrl.erstand hCM this ~rks, consider Projection 49. Here v.:e show a

piece of the conductor going normal and generating a quantity of energy given 2

by I rdt. If v.:e neglect the helium o.::>oling, this heat has to result in a

ternf)3rature rise to the material which is equal to cdT. We can rewrite the

equation as sham in the second line. The heat capacity and the resistance

as a function of temperature are known, hence, the right side of the equation

can be nurrerically integrated a.'ld is only a function of the final temperature.

Thus, for a given final temperature, there is a fixed [ I2dt the conductor

can tolerate. We have rra.de deta.iled 1roasurerrer1ts of this property of the

oonductor and find that at high currents the helium cooling can indeed be

neglected, and the curve shavm. in Projection 50 for t..l-ie temperature as a func­

tion of J r2dt is quite accurate. One can check this by exposing the con­

ductor to a given Jr2dt that should take it to the melting point of either

Stabri.te or Mylar. These tv.u substances give a calibration point at about

600° K.

The magnet protection system must always keep the J r 2at below sane value

set by the designer. Since the conductor has to absorb all of the energy in

the magnet, one can lower the average by driving all of the conductor in the

magnet nornal at once. BNL does this by causing the quench to propagate by

thermal conduction over a large number of turns in the magnet. At Fermi.lab,

we ai;::ply an active quench protection system, and v.:e actually activate a

-..- .....

I

,r--·.

- 77 -

--·--------------------------

I2

Rd t = Cd T if He disregarded.

t

f I 2d t =

t=o

For Short Times

T rv t 1/3

For Long Ti me

12 T MAX ,.... e

C(T) R(T)

dT

t known

Project.ion 49

900

800

700

600

~ 500 a.. :E LIJ

~400 LIJ ...J en U) 300 (J)

0 a..

~ 200 ~

100

(

0 1.00

QUENCH HEATING VS f I 2dt

f I 2 dt

2.00 3.00 4.00 5.00

HEATING SURGE (AMP) 2 SEC

( Projection 50

6.00 7.00 8.00

XI0 6

(

--.J o:i

-r-

- 79

mater in contact with the turns in order to spread the quench over a large

region of the magnet. Once enough conductor turns no:rna.l, the resistance

beccmes large, and the tbre constant to discharge becx:xnes short. Thus, the

J r 2dt is limited. The Fennilab ma.gnet protection scherre holds the J I 2dt

6 2 to about 5 x 10 amps /sec. It should be noted that as Projection 49 shows,

Tmax varies like an exponential of J I 2dt, so this quantity must be carefully

controlled.

When a magnet quenches, abnomal voltages appear across the turns. In

order for the magnet to survive is the voltage bet'ween turns must not becane

large enough to cause breakdown. Helium is not a very <Jood insulator in its

gaseous fonn. Projection 51 shows sane emperical breakdCMn data for the

Fermi.lab type conductor as measured at various temperatures in helium. Addi-

tional data that has been ~asured is shONn in Projection 52. It is safest

to disregard t.he properties of helium as an insulator and insure t..hat the

conductor is adequately insulated by SOITE material such an nwlar L.11 order to

guarantee that there will not be shorts when the magnet quenches.

Another problem that will be encmmtered in large systems is the radia-

tion que.n.ching of the magnets by any beam that manages to hit them. Sorre

rreasurerrents have been made on Fermi.lab magnets, and these appear in the

Tevatron Design Report. I show two examples in Projection 52. Two cases

\.\ere studied:

1. A fast beam loss, such as that ~uld be encountered in slow extrac-

tion.

2. A slow beam loss, as would be encountered in slow extraction of the

beam, for example for use in the Meson Lab.

-· 8

0

-

z 3: 0

Cf)

0 _

J

~

0 <::

(.) w

0::

-' CD

<(

z z

IJ_

0.:: :::::)

z I-0

Cl>

..... :c

z l.J...

a::: 0

:::::) .....

. ·---------· .

co U>

I[) c;;t

f()

/\ >f N

80.L 0.1

N8

n.l

0 0 f()

0 I[) (\J

0 0 N

0 -

LO

0

0 0 0 LO

0

~

0

r--i L

i)

§ ·r-1 .p

~ ·n

9

"""""'

--

- 81 -

Hav well we can control these loss mechanisms will control how close

v;e can operate the machine to the short sample limit. A great deal of work

:reeds to be done in this area yet.

Let m= surrrnarize here sate of the systems prablerns that we are going to

have to cope with. I have already mentioned the quench protection systems

are passive at BNL and active at Penni.lab. I think a question remains of how

active the system sh::>uld be. In the Fennilab case, a microprocessor is mmi­

toring the magnet behavior. The next level of sophistication atove protecting

the magnets would involve nonitoring the cryogenic part of the system and

minimizing the perturbation to the refrigeration.

The cooling system is unique because a synchrotron is spread out over

a large area and is very complicated with m:my parallel paths. The system

needs to be brought under ccroputer control. There are sane special problE'.rn.s

that arise because wne.11 the syste.711 is pulsed, a large heat load is added

through the eddy current and hysteresis loss in the magnets. In the Fe:rmildb

case, this results in a large evolution cf gas v.hich must be handled by the

refrigeration system.

We desrx=rately need new lea1< detection systems and ways to isolate leaks

in the complicated cryostu.ts t.11.at house the magnets. We need to study how

to get fast access to the ITB.gnets in order to change t.}iem if one is dam:tged.

We need to understand for the future the behavior of superfluid helillll in

large systems.

Pciwer supplies will need additional develoµnent. The long spills that

are possible with the superconducting system can change the proton economy.

"""'82 -

Long spill tines n~an that the secondary beams will becorre mJ.jor users of

protons. In the past, the cmmting rate has set a limit to the number of

protons that could be utilized in these bearn.s. Long spills involve very

stable IXJWer supply systems.

Correction coils will dem:md much more conplicated programs and more

sophisticated power systems in order to correct the field dynamically. In

the pa.st., the rnam burden of good field has been placed upon the di.pole mag­

nets, and this good field has b2en determined by iron stampings. This is no

longer Lrue, and indred, in the future it could becorre very expensive to place

the rrain burden for good field on the dipoles. The correction coils must be

o:msidered as part of the econcmic equ.a.tion. It may well be less expensive

to concc-mtrate nore on the correction circuit and less on the quality of the

cli~les.

Finally, as I have rr.entioned, we need much rrore work on understanding·

the interact.ion of radiation on the magnet system.

Conclusion

I would like to conclude this talk by surrrnarizing shortly our situation.

We have not made very much scientific progress in understanding training and

the quenching phenane.na that take place in our magnets. In order to take the

next step and go to a 10 Tesla magnet, much work needs to be done. The sup-­

p:>rt matrix must re urrlerstood in detfill, and the black rnagic turned into a

science. We need to understand the action of HeII an the cooling system, and

in order to go to higher field magnets, we rnay -well have to develop new tech­

niques to form Nb Sn in situ. We have made great strides in understanding the

role of quality rontrol in the construction of magnets and understanding the

-- 83 -

m-:mufacturing process that will eoonanically construct useful magne€s. We ·· · ·

are on the verge of obtaining much inforrration about systems of magnets and

hcM to control them, and I feel we are making great strides in this field.

It seems inconceivable to me that only five years ago -Y.e were still building

1 ft. rrodel magnet5 at Fennilab. HOY."ever, it is still true that many inter-

esting and vexing questions still remain.

Many of my colleagues at Fennilab ~e involved in oollecting the infor-

mation I have discussed. Their names can re found in the appended FNAL Biblio-

graphy. In addition to their articles, there are a series of 100 or rrore in-

temal rrerros or technical notes tmder a series number UPC-1 to UPC-129. This

index is available fran the FNAL Librarian, as are the notes themselves.

Finally, the great influence of Dr. R. R. Wilson can re seen throughout

the FNZ\T.., w::>rk.

... ,.•

-BIBLIOGRAPHY OF FERMII.AB MA'l1ERIA.L

RELATING 'ID SUPERCDNDUCTilJG MAGN.l:,""""I'S

Section 1 - Energy Doubler Division Publications - 1976, 1977, 1978, 1979

Section 2 - .UPC Notes

Section 3 - List of Technical Menos Available at Fei.11tllab (Janua.iy 1979 to

Present)

- Fermilab

ENEffiY DCUBLER DIVISION PUBLICATIONS - 1979

March 24, 1980 Section 1

"Progress Rep::>rt - Fenn:ilab Energy Doubler", A. V. Tollestrup, IEEE Trans. ~-2·, MAG-15, No. 1, 647 (Jan. 1979).

"Fennilab Doubler Magnet Design & Fabrication Techniques", K. Koepke, G. Kalbfleisch, W. Hanson, A. Tollestrup, J. O'Meara, J. Saarivirta, IEEE Trans. Magn., .MAG-15, No. 1, 658 (Jan. 1979).

"The Support & C:ryostat System for Doubler Magnets", G. Biallas, J.E. Finks, Jr., B. P. Strauss, M. Kuchnir, W. B. Hanson, E. Kneip, H. Hinterberger, D. DeWitt, R. Pooers, IEEE T-.cans. Ma.on., MAG-15, 131 (Jan. 1979).

"Rcom Temperature Field Measurerrents of Superconducting Magnets", R. E. Peters, L. Harris, J. M. Saarivirta, A. V. Tollestrup, IEEE Trans. Magn., MAG-15, 134 (Jan. 1979) •

"A Superconducting Synchrotron Paver Supply & Quench Protection Scherr>.3", R. Stiening, R. Flora, R. Lauckner, G. Tool, IEEE Trans. Magn., .MAG-liJ, No. 1, 670 (Jan. 1979). --·--

"Sur;erconducting Correction Elem?.nts", M. Leininger & F. Kircher, IEEE 'l'rans. Nucl.~Sci., NS-26, No. 3, 3931 (June 1979).

"Power for the Fe.uPilab Tevatron Helium Liquefier", J. Hoover, J. R. Orr, J. Iqk, A. Visser, IEEE 'I!.?l_lS._Nucl. Sci., rJS-2~, No." 3, 3938 (June 1979).

"How to Mass Produce Reliable Cryostats for Large Particle Accelerators", B. P. Strauss, G. Biallas, R. Powers, IBEE Trans. Nucl. Sci., NS-26, No. 3, 3941 (June 1979) • -

"Process Control System for the Tevatron Liquefier", H. R. Barton, Jr., et. al., IEEE Trans. !'Juel. Sci., NS-2_£., No. 3, 4099 (June 1979).

"Central Liquefier for the Fermilab Tevatron", M. Price, R. Rihel, R. J. Walker, et. al., IEEE Trans. Nucl. Sci., NS-26, No. 3 {June 1979).

"Doubler-Teva.tron µP Quench Protection Systema, R. Flora & G. Ta>l, IEEE Trans. Nucl. Sci., NS-26, No. 3 {June 1979).

"B:xun Detector Assembly for the Fenuilab Energy Doubler", E. Higgins, Jr., T. Nicol, IEEE Trans. Nucl. Sci., NS-26, No. 3, 3426 (Jtme 1979) •

- 2 -

"The Multiple Magnet Test Program", R. Flora, M.. Kuchnir, et. al., IEEE Trans. Nucl. Sci., NS-26, No. 3, 3888 (June 1979). ......lflf

"Pressure DeveloJ;ID?.nt During E:nergy Doubler Quenches", M. Kuchnir & K. Koepke, IEEE Trans. Nucl. Sci., NS-26, 4045 (June 1979).

"Recent Measurement Results of E:nergy Doubler Magnets", M. Wake, D. Gross, M. KunE.da, D. Blatchley, A. V. 'lbllest:rup, IEEE Trans. Nucl. Sci., NS-26, No. 3, 3894 (June 1979).

£"' .... ...,,, £" .·-' \_., &:..'\,r· J Fermilab

- 3 -

,-.. t:. ,..J

-

... lt..,j

ENEffiY OOUBLER MA.GNEr DIVISION PUBLIO.TIOOS - 1978

11Ferm:ilab Experience on Large Superconductor Purchases and Its Implication to the Fusion Program", B. P. Strauss, R.H. Rernsl::ottom, and R.H. Flora, Proceedings of the 7th Symp::>sium on Engineering Problems of Fusion Research, IEEE No. 77CH1267-4-NPS (Jan. 1978).

"Production Test of Energy Doubler Magnets", R. Yamada, M. E. Price, and D. A. Gross, Adv. ~· ~·, Vol. 23 (1978).

r-:~~

r· 1 r ;.: \ , L ··- .. ! . .. J l.~~.J

- 4 -

Fer1nilab Decanber 28, 1977

1977 PUBLICATIONS LIST - ENERGY DOUBLER GROUP

"The Technology of Producing Reliable Superconducting Dipoles at Fermi.lab", W. B. Fowler, P. V. Livdahl, A. v. Tollestrup, B. P. Strauss, R. E. Peters, M. Kuchnir, R. H. Flora, P. Li.Iron, C. Rcx1.e, K. Koepke et. al., IEEE Trans . .Magp., MAG-13, No. 1 (Jan. 1977).

"Results of t.he Fennilab Wire Production Program", B. P. Strauss et. al., IEEE Trans. Magn., MAG-13, No. 1 (Jan. 1977).

"Measurerrents of Magnet Quench Levels Induced by Proton Fearn Spray", C. Rode et. al.*, ~ Trans. M?-9.!:!·, MAG...;13, ~'lo. 1 (Jan. 1977) •

"Danage to Stabilizing Materials by 400 GeV Protons", P. A. Sanger, B. P. Strauss et. al., IEEE 'l'rans. Ma~., MAG-13, No. 1 (Jan. 1977).

"Coil Ext.E'.nsion, Defonuation & Canpression During Excitation in Su}?9rconducting .Accelerator Dipole Magnets", A. V. Tollestrup, R. E. Peters, K. Koepke, R.H. Flora, IEEE Trans. Nucl. S-ci., NS-24, No. 3 (June 1977).

"Operation o.f Multiple Su_P2rconducting Energy Double..r Magnets in Series", G. Kalbfleisch, P. Lim~n, C. Rode, TF;"F!E Trans. Nucl. Sci., NS-24, No. 3 (J'une 1977) •

"Energy Doubler/Saver Safety Leads", M. Kuchnir & G. Biallas, 1£.:F:f: TrCl!l~· Nucl. Sci., NA-24, No. 3 (June 1977).

"Status of the Fermilab Energy Doubler/Saver Project", Energy Doubler Staff, P-.cesente<l by P. V. Livdahl, IEEE 'l'rans. Nucl. Sci., NS-24_, No. 3 (June 1977).

"~.ia~etization Effects in ~J?2rc'OJ1ducting Dipole Magnets", R. E. Peters et. al. , IEEE Trans. Nucl. Sci., NS-24, No. 3 (June 1977). .

"Energy Doubler Refrigeration System", C. Rode, D. Richied, IEEE Trans. Nucl. Sci., NS-24, No. 3 (June 1977) •

"Radiation Darrage Limitations fer t.1-ie Fennilab Energy Doubler/Saver", P. A. Sanger, IEEE Trans. Nucl. Sci., NS-24, No. 3 (June 1977).

"Op3ration of the Fennilab Accelerator as a Proton Storage Ring", A. V. Tollestrup et. al.*, IEEE Trans. Nucl. Sci.; NS-24, No. 3 (June 1977).

*Principal authors were not members of the Energy Doubler Group.

- 5 -

Fermi lab December 6, 1976

1976 PUBLICATIONS LIST - ENERGY DCUBLER SECI'ION

"A large Scale Pt.nnped Subcooled Liquid Helium Cooling System", P. VanderArend, s. Stoy, and D. Richied, Advances in Cryo:;enic Engr., ~! (1976).

"Measurement of Thenna.l Conductance", M. Kuchnir, Advances in C~enic Engr., 21, 153 (1976).

"The Fabrication of Multi-Filament NbTi CatlfOsite Wire for Accelerator Appli­cations", B. Strauss, R. REmsl::ottan, and P. Reardon, Applied Polymer Sympo­sium No. 29, (John Wiley & Sons, Inc., 1976) 53-59.

The Energy Doubler -- A Progress Ref.Ort for the Energy Doubler, Saver, Col­lider Project, by the superconductor Group et. al., June 1976, 2nd printing O::tober 1976.

(lL

.ipH- UPC cti11_L Number

D

D

D

L

D

D

D

D

D

D

None

0

2

3

4 (Revised)

5

6

7

8

9

10

11

12

13

ENEHGY DOUBLEH HEPOHTS BY UPC NU l\113EH Section 2

Serial Number

51 .

62

65

15 16

65

6

7

8

9

10

11

12

January 3 1, 19 79

Title

Doubler Design Report (DH.AFT) dated October 21, 1978, by the Fcrmilab Staff

Draft Outline for Energy Doubler Design Report. H. T. Edwards, October 27, 1978

.. Thc:.Abo!'tablc Bca~n. -T. L. Collins. Rev. Nov. 14. 1978

Radiation on Magnet Coils from the \Vire Septum. T. L. Collins, Rev. November 14, 1978

The Pressure Bump Instability in the Fcrmilab Doubler, P. Limon, November 6, 1978

Dipole Acceptance Criteria, H. Edwa;rds and P. Koehler, February, 1979

Cormnents on Prcsst~re Bump Instability in the Energy Doubler, D. A. Edwards, No7ember 6, 1978

(Superceded by UPC No. 32)

Multipole Components of the Dipole Field, S. Ohnuma, October 23, 1978

Field Quality of Dipoles in the Tunnel, S. Ohnuma, September 11, 19 78

Field Quality of Dipoles for the Doubler, S. Ohnun1a, September 11, 1978

Harmonic Components of the n)ubler Dj,1iole Field Field in DC and AC Modes, S. Ohnuma, September 29, 1978

Doubler Dipole File, S. Ohnuma, October 6, 1978

Memo from S. Ohnuma to A. Tollcstrup regarding Harmonic Components, October 12, 1978

Center Coordinates from DC and AC Data, S. Ohnuma, October 27, 1978

St!nt to Dupli­cating

D

D

D

D

D

D

D

D

D

D

D.

D

D

UPC Number

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

Serial Number

41

57

39

81

49

52

40

56

38

37

42

66

58

13

-2-

Title

High-Beta Straight Sections for the Doubler, T. L. Collins, November 14, 1978

(Superceded by UPC No. 22)

Magnet Aperture and .Extraction, Dated December 4, 19 7 8. M. Harrison, REVISED January 17, 19 79

Energy Doubler Circ1.:_nnferenc~ .. -:: _pp_colliding be.a111.s co11sid~rations, December 1, 19 78; L. C. Teng.

Layout of lvIR and ED Functions in Long Straight Sections, Dece1nbcr 1, 1978, L. C. Teng

Beam Transfer - Normal and Reverse Directions, December 4, 1978, S. Ohnuma, REVISED Jan. 11, 1979

Energy Doubler Beam Abort System, December 7, 1978, F. 'ri.lrkot, HEVISED January 1, 1979

Doubler Lattice,November 17, 1.978, D. E. Johnson (Do Not Duplicate - - too long)

Effect of a High Beta Insertion on Resonant Extraction from the Energy Doubler, D. A. Edwards, REVISED December 13, 1978

High Eta for Stacking, T. Collins, December 13, 1978

Preliminary Doubler Low - Insertion, D. E. Johnson, December 13, 1978

Correction and Adjustrn.ent Magnet System Progress Report for Meeting of December 1, 1978, D. E. Edwards

Electron - Proton Instabilities in the Energy Doubler, P. Limon, December 13, 1978

Local ()!:bit Bumps, extracted beam phase space and an initial look at fast :resonant c::...i;raction, M. Harrison. _,,,·. Dcccmbei· 19, 1978, HEVISED January 12, 1979

Multipole Fields in Doubler Dipoles - Review, S. Ohnuma, DN:tnnbcr 21, 19 7 8

,._!nt to illpli- UPC ~4-i no· Number ~.:_::_,........., __ -----D 2'-)

D 30

D 31

D 32

D 33

D :34

D 35

D 36

D 37

D 38

D 39

D 40

D 41

n 42

Serial Number ------1·1

63

17

4

67

55

53

71

24

50

54

61

3

-3-

Title

Effects of Full Quenches anJ \Vacm·-Cool Cycle on Multipole Contents, S. Olmuma, December 26, 1978

Studies on Radiation Shielding of Energy Doubler Magnets (1), H. Edwards, S. Mori, and A. Van Ginneken, December 2 7, 19 7 8

Study of Eddy Current Effects in Energy Doubler Dipole Magnets Introduction, January 2, 1979, I-·L -Shafer : :-~-- · ::- - · . -

Status of l\llagnetic Measurements of Energy Doubler Magnets, F. T. Cole and D. R. Gross, December 20, 1978

Heat Loss by \Vall Resistivity, S. Ohnuma, Decern.ber 28, 1978

Comparison of Half Integer and Third Integer Ex-traction for the Energy Doubler: 1. Basic Processes, D. A. Edwards, December 1978

Energy Doubler Abort Kicker, Q. Kerns, and J. lVfcCarthy, January 8, 1979

Transverse Active Dampers for the Tcvatron, B. Miller, January 17, 1979

Transmission Linc~ Characteristics of Energy Doubler Dipole Strings, R. Shafer, January 10, 1979

Main Ring E)ctraction and Doubler Injection Kicker, J. McCarthy, January 12, 1979

Pulses Magnet for the Doubler, J. McCarthy, January 16, 1979

Supplement to UPC 30, S. Iv'Iori, H. Edwards, and A. Van Ginneken, January 12, 1979. Supplement to UPC--10.

Design of the Energy Doubler by H. T. Edwards i\ppcndix 1 - Design Status of the ED, March 1979

Qu~clrupolcs, G. Kalht1cisch and H.. Peters, Jan. 1979

Sent to Dupli- UPC eating Number ---· -· -- -·----

D

Hold 44

D 45

D 46 .

D 47

D 48

D 49

D 50

D 51

D 52

D 53

D 54

D 55

D 56

·D 57

Serial Nmnber

.5

18

19

20·

21

22

23

25

26

28

30

31

32

33

34

-4-

Ti tlc~

Stmlies of Some Energy Doublur l'viugnet Properti_,:;:;, D. Gross, Janm.ry 15, 1979

Dipole Data Table, B. P. Strauss, JanuG.ry 16, 1979 (Must be updated)

Doubler Dipole and Quadrupole Design Field Specifications, S. C. Snowdon, January 16, 1979

.. lJT"'ff·ol'.·YYI u~~p of 4 TY10h B'O'Y'C l\.'Ta: 0·no-ts i"n J)ntih]e·p !_,. 1.i.-....L.·• ·• •J....L.• .._........ ..._ ..1..l.J..'--'4A .,I. ....._ ..l..f"- s ........ l.J... • J \.,.._. . ...._ J

S. C. · Snowdon, · J ~muary 1 7, 19 79

Energy Doubler Magnet Power Supply and Quench Protection, R. Flora and G. Tool, October 18, 1978

Energy Doubler Po\\rcr Supply Location, G. Tool, January 7, 19 79

Return Bus Quench Protection, H. Flora, J·an. 12, 1979

Energy Doubler Rippie H.cduction and Regulation G. Tool, January 10, 1979

Energy Double1· Refrigeration System, C. Rode, October 18, 1978 (Revised January .. 1979)

Energy Doubler Tunnel Cryogenic Cornponents, C. Rode, January 1979

Energy Doubler Vacuum System, P. Limon, October 20, 1978

Vacuum Requirements for pp Colliding Beams, S. Ecklund, January 11, 1979

B12 Doubler l\.'Iagnet Operating Experience, K. Koepke, January S, 1979 Il12 Awning Program - Phase IV, · R. Orr, J::1n. 6, 1979

A1 Cryogenic J\lagnct Huns, C. Hock, Jamw.ry 17, 1979

Energy nouhlcr Magnet Tnstalb.tion, H. Orr, Oct. 12. 1978

SL~nt to Dnpli.- UPC eating N1tmher ······-- -----

lJ 58

D 59

D 60

D 61

D 62

63

64

b 65

D 66

D 67

D 68

D 69

·p---- 70

Serial Number ------35

36

43

44

45

46

47

48.

59

60

61

68

69

-5-

Title

Doubler Laltio.:: ;:i.ncl Ph:1sc-..:\rnplitude Function;;; T. Collins, June 7, 1978

Lattice for the Doubler, T. Collins, January 17, 1979

Notes on Excitation of Correction and Adjustment Elements, D. Edwards, December 20, 1978

.A. Test of LCP2 inside QB6, _Jan. 5, i979 "J3.· Par2.mete:rs of the I~ong Tun~ Quadrupole,

Jan. 5, 1979 C. Strength of the Long Correcting Coils, Nov. 21,

1978 D. Test of LQT-1, February 1979

QB Quadrupole/Spool Package, T. Nicol, January 8, 1979

Beam Position lVlonitoring, E. Higgins, Jan. 31, 19 79

Tevatron RF - Q. Kerns, and Vv. l\tTiller, October 19, 1978

H.F R2quirements for Tevatron Use with pp Colliding Beams - J. Griffin, Jan. 1979

Good Field Region of the De sign Bend Magnet and Expected Behavior of Extraction, H. Edwards and lVI. Harrison, October 19, 1978

Energy Doubler Multipole Field - Randorn Errors and Slow Extraction, L. C. Teng, October 21, 1978

Sensitivity of an Energy Doubler Dipole to Beam Induced Quenches, B. Cox, P. 0. Mazur, and A. Van Ginncken, November 1978

Aclc.lendum to the Beo.m Pipe Heating Calculations, J. Griffin, January 12, 1979

Progress Report on the Study of the Effects V/hic:h Arc Intensity Dependent in the Energy Doubler, A. G. Huggic:ro, January 12, 1979

Sent to Dupli- UPC eating Nun1ber ------

D 71

D 72

D 73

D 74

D 75

D 76

D 77

D 78

D

D

D

D

D

D

J) 79

80

Serial Number

70

72

73

74

75

76

77

78

79

80

1

2

27

29

Title

Consideration - about the Bellow, A. G. H.uggiero, January 17, 1979

Im1ivi<lual Bunch Longitudinal Instabilities, A. G. Ruggiero, January 5, 1979

A Scenario to Achieve a Luminosity of Approxi-1natcly 5 X 1029 cm - 2sec-1 for pp Collisions in the J?ermilab Energy Doubler, F. Mills and D. E. ~Young, November 11, 1978

Transparencies on pp, F. E. Mills, November 1978

Doubler Requirements for Proton Antiproton Collisions, F. E. Mills, January 17, 1979

Kissing Magnet Design for the pp Collider, D. Ayres, J. A. Bywater, R. E. Diebold, M. IL Foss, R. J. Lari, Nov. 8, 1978, Rev. Dec. 21, 1978

Easy Low Beta for the Main Hing, r.r. Collins, l\1arch 11, 19 76

pp Transparencies, R. E. Diebold, November 1978

Control System Facilities, M. Gormley, Oct. 20, 1978

1000 GeV S\vitchyard Upgrade, H. Dixon, Oct. 16, 1978

Progress Report - Fermilab Energy Doubler, A. Tollestrup, September 19 78

Fermilab Doubler Magnet Design and Fabrication K. Koepke, et al., September 1978

The Central Helium Liquificr, R. \Valker, Nov. 6, 1978

Refrigeration System for the Energy Doubler. P. C. Vandcr Arend, Nov<.!mbcr 27, 1978

Taken from Talk of T. Collins, December 1978.

High l ,umitto!;ii.y pp Colliding Beam Straight Section n. E. Johnson, January 18, 1979

Sent to ,--.. Dupli- UPC

C'.ltin,•: Nun1her -----D 81

D 82

D 83

84

85

86

87

88

89

90

91

Serial Number

-7-

'riilP.

Bunch-to-Bunch Longitudinal Instabilities, A. G. Ruggiero, January 1979

Energy Doubler Refrigeration Load,, M. Kuchnir,, January 19 79

An Effect of Full Quenches on the I~.,ield Quality of Dipoles Hysteresi_$. ·Dependence of the Dipole -Field Quality at Injection on the Base Current of Ramping S. Ohnuma, February 1, 1979

Notes for Energy Doubler Correction Element Power Supplies, R. Yarema,, January 25., 1979

Interpretation of Inductance Bridge 1\!Ieasurements of EnGrgy Doubler Dipoles, R. Shafer, February 14, 19 79

The Amateur Magnet Builder's H::i.ndbook, A. V. Tollestrup .. February 22, 1979

Extraction III - Fast Resonant Extraction, lVI. Harrison, February 27 .. 1.979

Multikilogauss Fast Kickers, Q. Kerns, February 16 .. 1979

Coil Size Scaling Laws for AC Losses in ED Dipoles, R. Shafer, March 29, 1979

Beam Heating by vVall Resistivity, R. Shafer, March 5, 1979 Revised March 27, 1979

Recommendation to Heplace Existing Inductance Bridges for ED Dipole L and Q Measurements, R. Shafer, February 22, 1979

Proposal for Simple Reliable L and Q Measuring Circuit for ED Dipoles, H. Shafer, February 16, 1979

UPC Number

92

93

94

95

96

97

98

99

100

101

102

103

104

105

106

Serial Number

-8-

Title

The Longitudinal Coupling Impedance /\. s~>oc~.ah.:cl to the Pick-Up Devices for the Energy Doubler, A. G. Ruggiero, January 1979

The Multiple Magnet Test Program, P. Brindza, R. Flora, G. Kalbfleisch, K. Koepke, M. Kuchnir, P. Limon, D. Richied, C. Rocle, IL Stiening, S. Stoy, and G. Tool, March 19 79

Energy Doubl~r Refrigeration System, P. Brindza, D. E. Richied, and C. H. Rode, March 19 79

Recent Measurement Results of Energy Doubler Magnets, M. Wake, D. A. Gross, M. Kumada, D. Blatchley, and A. V. Tollestrup, March 1979

The Great Doubler Shift, T. L. Collins, April 1979

Normal Doubler Lattice, D. E. Johnson, March 23, 19 79

B12 Quench Test Update, K. Koepke, May 10, 19 79

AC Dependence of ED Dipole Transfer Hatio \Vi th ~

and \Vithout Damping Resistor, R. Shafer, 1\'Iay 11, 1979

The DO Electrostatic Septum, M. Harrison, i\'Iay 22, 1979

Sensitivity of Beam Position Detectors for the Tevatron; R. Shafer, May 18, 19 79

The Effects of Power Supply Hipple on the Energy Doubler, R. Shafer, June 4, 1979

An estim.ate of the Vertical Dispersion Function, S. Ohnuma, June 22, 1979

Some Calculations About the Heaters Used in the EDS Dipoles, F. Kircher, June 1978.

Summary of the First Twelve 11200" Series i\lagnets in Hegards to Ramp Hate Dependence of Quench Current, D. Gross and G. Jordc:in, June 22, 19 79

._, LargeAmplitucleOrbits, M. Harrison, June28, 1979

UPC Nmnber

107

108

109

. 110

111

112

113

114

-115

116

117

118

119

120

121

122

Serial Number

- 9 -

'ritle

1 1 Eft'~ct of 10 Proton Bunch on Cavity , I'· Teng, July 20, 1979

Lower Order Dipole Ifo.rmonics, M. Harrison, August 20, 1979

Bend Magnet Quri.drupole Component, M. Harrison, August 20, 19 79

-J)ipole Sl;:ev1 Ha.rmonics, JVL Harrison, August 23, 1979

On Reducing the Coupling Impedance of Step Dis­continuities in Vacuum Chambers, R. Shafer. August 26, 1979

Proposal for Satellite Refrigeration Control Inter.face, H. Edwards, September 11, 1979

Studies on Radiation Shielding of Energy Doubler :rvfagnets (H), H. Edwards, S. Mod, and A. Van Ginneken, June 1979 (Continuation of TJPC-40 and UPC-30)

(Jostlein) 'rhe ED/S Vacuum Control System, November 21, 1979. {H. Justlein and L. Bartelson)

Monte Carlo Studies on Tevatron Extraction, l\ll.. Harrison, October 26, 1979

G4idelines for the Field Quality of Doubler Quadrupoles, S. Ohnuma, Novernber 19. 1979

Comments on the Beam-Beam Interaction in Storage Hings, S. Ohnuma, December 10, 1979

Field Flatness of Doubler Dipoles, S. Ohnuma, December 20, 1979; Addendum December 27, 1979

Vacuum and the Beam Lifetime in the Doubler, II. Mizuno, S. Ohnuma, A. Ruggiero, Dec. 21, 1979.

Energy Doubler Dipole String Electrical Impedance Measurements at B12, IL Shafer, November 26, 1979

Handorn JTield Errors in the Tevatron Dipoles, A. V. Tollcstrup, January 18, 1980

Vacuum Tests on a Sarnplc Lambertson 1\fa.L~n!;t

ll. Jostlein, G. Lawrence, M. May and l\'f. Tiaph::icJ. .Tanu::try 19 80

UPC Number

123

121

125

126

127

128

129

130

131

132

133

Serial Number

-10-

Title

The Qucldrupole Component in Dipoles and tlte Injection JVlismatch, S. Ohnuma, March 6, 1<)80

Tcvlat: A New Program for Computing Lattice · functions for the Energy Saver /Doubler, A. D. Russell, , March 17th, 1980.

Correction of Skew Quadrupole Field in the Doubler, S. Ohnurna, March 26, 1980

On:the-inductance of the Doubler·l\-Iagnet Syste1n, R. Shafer, April 2, 1980

Non-Linear Characteristics of the Doubler Magnet Yoke Inductance, R. Shafer, April 8, 1980

Longitudinal Instabilities Revisited, S. Ohnuma, April 8, 1980; Addendum April 22, 19 80.

Notes on a Study \Veek on Beam Storage in the Tevatron, Aprll 7-11, 1980, F ... .r. Cole, April 14, 1980

Vacuum Proposal for tl~e Injection Straight Section ( EO) ..._,, H. Jostlein, April 29, 1980

Comments on Correction Package Test Procedure D. Edw·ards, June 16th, 19 80

A Reply to the Question from HEPAP "Are the Doubler Magnets Good Enough for the Collider?, S. Ohnurna, June 26, 1980.

On The Longitudinal Coupling Impedance of the Energy Doubler Beam Position Detectors, Hobert E. Shafer, July 15, 1980

-

Ferm ii ab Section 3

LIST OF TECHNiu\L .MfilOS AVAILABLE AT FERMILAB

(January 1979 to Present)

TM-844, "Short Sample Tests and Tevatron Magnet Performance," R. H. Flora, January 22, 1979.

'IM-845, "Application of Superconducting Technology to Big Accelerators," R. Yanada, January 1979.

'IM-849, "Pump Down Tirre of Energy Doubler Magnets," M. Kuchnir and J. Tague, January 1979.

'IM-860B, "Magnet Test Facility - Magnet Protection, Qt...iench Circuit Tie-In", Cryogenic Consultants, January 8, 1979.

'Il'1-863A, "Helium Cooling of the Shells of a Dir,:ole .Magnet in the Mooel 135 Cryostat," Cryogenic Consultants, January 30, 1979.

TM-863B, "H2lium Cooling of. t.~e Shells of a Dipole Magnet in the .Model 135 Cryostat," Cryogenic Consultants, January 30, 1979.

'IM-866, "Quadrupole Torsional Pi.gidity Measurerrents," Mark Leininger and Dan Smith, March 1979.

TM-867, "Doubler-Tevatron µP Quer:ch Protection System," R. H. Flora and G. s. Tool, March 1979.

TM-874, "The Great Doubler Shift," T. L. Collins, April 1979.

'IM-878, "A Four-Quad.rant Magnet •rrim Pa.ver Supply," R. Yarerra, Vay 7, 1979.

'IM-880, "Status of the Fermilab •revatron Project," A. V. Tollestru.p, l'>..pril 1979.

'IM-881A, "Evaluation of Various Options for the Compression and Distribution of Compressed Helitnu Gas to the Satellite Refrigeration System of the Energy Doubler," Cryogenic Consultants, March 1, 1979.

'IM-881B, "Evaluation of Various Options fer the Ccrnpression and Distribution of Corrpressed Helium Gas to the Satellite Refrig2ration System of the Energy Doubler," Cryogenic Consultants, March 1, 1979.

-

.,.

- 2 -

'IM-882, "Turbo Expander for Satellite Refrigerator, 11 Cryogenic Consultants, April 1979. ~

'IM-891, "Support Location for ED Dip:Jle Ma.gnet," W. B. Hanson, August 2, 1979.

'IM-892, "Evaluation of Superconducting Dipoles," L. C. Teng, August 16, 1979.

'lM-897, "Heat Rerroval from Beam Scraping of an Energy Doubler Magnet," Cryo­genic Consultants, May 15, 1979.

'IM-898, "Investigation of Means to Increase Refrigeration capacity of the CH4 Satellite Refrigeration S'yst..eill of the Fennilab Energy Doubler," Cryogenic Consultants, August 30, 1979.

'IM-899, "Design Evaluation of Valve Actuator Cams (Fennilab Reciprocating Design)," Cryogenic Consultants, August 6, 1979.

'IM-900, "Evaluation of Wet and Gas Helium Expan..sion Engines (FermilaJ:> Recip­rocating Design)," Cryogenic Consultants, July 10, 1979.

'IM.-903, "Wet F~sion Engine for Helium Refrigerator Application," Cryogenic Consultants, September 6, 1979.

'IM-907, "Mechanical Drive and Control System for High Efficiency Wet Helium Expansion EngL.i.e, 11 Cryogenic Consultants, Sepb?1nber 18, 1979.

'IM-910, "Field Quality of Doubler Di:r:oles and Its Possible Irnplications, 11

S. Ohnuna, October 15, 1979.

'1M-910A, "Addendum to 'Il1··910, Field Quality of Doubler Di:r;:oles and Its Pos­sible Implications," S. Ohnuna, November 6, 1979.

'IM-916, "Method for Reducing Cryostat Pressure During a Heater Induced Magnet Quench," Cryogenic Consultants, October 25, 1979.

'IM-917, "Satellite Vacuum Vessel Pressure Rating and Relief Port Size," Cryogenic Consultants, October 26, 1979.

'IM-922, 11Harm:>nic Fluctuations in Energy Doubler Dipoles from Azinn.lthal Wire Displacements Inside the Coils," J. 'hblff, November 9, 1979.

'IM-925, "Bellows Failure of Energy Doubler Magnet String," Cryogenic Consultants, November 29, 1979.

'IM-927, 11.Ma.gnetic Measurements on a Main Ring Trim Quadrupole," D. Krause, R •. Mills, C. Moore, F. Turkot, Dece.."Tlber 1979.

'IM-929, "The Alx>rt Dump for the Energy Saver/Doubler and Main Ring," T. E. 1'oohig, December 6, 1979.

- 3 -

'Il1.-935, "Sensitivity and Accuracy of the Measurerrent with the Yoke Coil of the Doubler Dipole," H. Kaiser, January 1980.

TM--943, "Mechanical Distortions in the Energy Doubler Dirx:>les During Cooldown. A Possible Mechanisn to Induce Torques in the Cryostat Canponents," C. Hojvat, February 15, 1980.

'IM-948, "Cryogenic Tests at Al," P. Brindza Cll1d T. Peterson, March 7, 1980.

'lM-949, "Design Considerations for High Efficiency Wet Helium E>:pansion Engine," C:ryogenic Consultants, March 5, 1980.

'IM-951, "CHL Liquid Helium Dewar/Ring Interfaces and Operation," Cryogenic Consultants, March 18, 1980.

'Il'1-956, "Quench Protection Circuit for Superconducting Magnets with Eddy Currents," M. Wake, April 1980.

'Il1-962, "Electrical Protection of a Low-Current Superconducting Magnet, 11

A. Visser, April 25, 1980.

'IM-967, "CENTRAL: A Corrputer Program to do Performance calculations for a Helium Liquefier," H. Barton, May 19, 1980. ·

'IJYl-968, "Evaluation of Crankshaft Designs Pro:r:::osed for the Dry and Wet Helium Expansion Engines, " Cr1·oge.'1.j_c Consultants , May 15 , 198 0.

'Il'1·-971, "l~n Analysis of the Cooldow.~1 of the Central Liquefier, 11 H. Barton, ,June 1980.

'IM-972, "Tftf~ Bucking Coil - An Improverre.nt for Dipole Rocrn TEmiperatu.re Mea­<~1_i .. rc....~ts I •l S, 'WJlff I May 27 I 1980 o