M 2870_Biweekly_Report_June_13_1954 2870 Biweekly Report June 13 1954

Memorandum

M-2870

Digital

Computer La'poratory

Massachu~etts

Institute'

of

Technology

Cambridge 39. Massachusetts·

SUBJECT:

PROGRESS

REPORT

&Y

17

mROUGH

JUNE

1'3,

1954

.

To:

Jay

11.

F

ottester

From:

Scientific

and

Engineering

Computation

Gr<?up

1.

MATHEMATICS,

CODING

AND

APPLICATIONS

1.1

Introduction

Page

l"'ef

22

This

report

covers

the

four~ek

period

beginning

on

17

May,

1954.

During

this

period

231

coded programs were

run

on

the

time

allocated

to

the

Scieniif1e

and

Engineering

Computation

(S&EC)

Group.

representing

wor}cthat

has been

carried

on

in

35

problems.

Section

1.2

below

contains

progress

reports

as

submitted

by

the

pro-

grammers

together

with

an

indication

of

the

programming and machine

time

expended

on

each problem.

Abstracts

are

no

longer

included

with

each

problem. However,

as

in

the

past.

a more

detailed

description

will

be

presented

in

the

initial

progress

report

for

each

new

problem.

The

'reader

is

reminded

that

the

laboratory

issues

summary

reports

every

three

months

containing

detailed

descriptions

of

the

problems

studied

during

that

period.

In

particular

terminating

reports

for

completed problems

will

be

found

in

these

summary

reports.

Three problems (1113.

#119,

and

#161)

were completed

during

this

period.

Machine

runs

on

problem 1101

(Optical

Properties

of

Thin

Metal

Films) have been completed and

the

results

are

being

written

up

for

publication,

Programming on

this

problem

will

be

continued

when

Dr.

Loeb

returns

from.

Europe.

Detailed

initial

reports

are

given

below

for

problem #181

(calculation

of

proton-deuteron

scattering).

1187

(response

of

a

fuel-flow

controller),

#188

(relative

water

production

in

an

oil

reservoir),

and 1192

(frequency

and phase spectrum

ana~sis

of

seismograms).

Satisfactor,y

results

have

'been

obtained

for

probl~m

1181 and

the

problem

is

now

terminated.

1.2

Programs and Computer Operation

100.

The

Comprehensive System

of

Service

Routines:

H-2870

Best,

67

hours;

Combelic,

49

hours;

Demurjian, 27.5

hours;

Denman,

23

hours;

Frankovieh,

3P

hours;

Siegel,

52

hours;

WI,

447

minutes

page 2

This

system

(as)

provides

for

the

automatic

selection

and

use

of

utility

programs

during

computer

operation.

The

librar.y

of

utility

pro-

'

grams

is

perma~ently

recorded

on magnetic

tape

and

includes

conversion

programs, binar.y

input

programs, equipment

checking

programs, and

post-

mortem programs.

The

comprehensive system

has

been

operating

successfullY

during

the

past

four

weeks.

Several

diffieulties

in

reading

utility

pro-

grams from magnetic

tape

to

core

m.emory

were

encountered,

but

these

were

explained

by

a

mistake

in

the

program

used

to

re-record

utility

programs

on magnetic

tape.

This

mistake

caused

marginal

recording

on magnetic

tape

•

The

following

changes

are

being

made

in

the

group

11 program:

1.

.Two

entr,y

points

are

being

added whose

use

will

cause

the

program

to

record

the

title

of

the

tape

just

read

in

either

on

the

direct

printer

or

on

the

delayed

printer.

In

addition,

the

date

and

the

time

will

also

be

recorded.

2.

The

logging

programs

are

being

modified

so

that

standard

time

will

normally

be

recorded

in

the

log

instead

of

computer

time.

This

will

be

accomplished by

recording

the

standard

time

on

the

buffer

drum

at

the

beginning

of

the

operating

period.

3.

Provisions

are

being

made

so

that

the

librar,y

of

utility

tapes

used

by

the

.group 11 program

can

be

stored

on

an

arbitrar.y

tape

unit.

This

will

prove

useful

in

checking

out

new

utility

programs

or

modif.ring

old

ones.

The

generalized

post-mortem program

is

being

modified

so

that

the

scope

can

be

used

as

an

output

unit

for

post-mortem

results.

At

present

on~

the

direct

printer

and

the

del~ed

printer

are

available.

The

following

formal

proposals

are

being

issued:

1.

It

has

been

requested

that

a

new

switch

be

installed

in

test

control

which

could

cause

the

computer

to

stop

on

the

instructions

8i

11

or

si

12.

In

this

way

accidental

resetting

of

the

computer

clock

could

be

prevented.

2.

.It

has

been

requested

that

equipment

be

installed

for

storing

the

contents

of

Tsrio'Us computer

lights

after

an

alarm

in

such a way

that

they

can

be

recorded

on

an

output

unit

by a program

contained

in

test

storage.

The

lights

are

presently

being

recorded

by

the

operators,

a

process

that

is

time

consuming and

can

lead

to

transcription

errors.

A

memorandum

describing

the

group 11 program

is

being

written.

staff

M-287

0 page 3

The

program

brief~

desoribed

in

the

biweekly

of

May

16 and

currently

being~trouble

shot

will

provide

a

generalized

means

of

trans-

ferring

binary

information

between

various

external

units.

Specifically,

it

will

be

able

to

read

information

from

the

following

sources:

1.

556

tapes

via

.PETR,

2.

556

tapes

via

MTR,

3.

blooks

(i.e.,

programs

recorded

on

a

MT

unit

in

a form such

that

they

can

be found and

read

in

by

an

input

program) from

HT

units

0.

1.

2.

30

It

may

record

binary

information

in

the

following

ways:

1.

556

tapes

on

MT

unit

3

for

delayed

puno

hing,

2.

556

tapes

on

the

direct

punch

for

emergency

use

only,

3.

blocks

on

NT

units

I,

2.3.

Unit

O.

of

course.

can

never be

recorded

on

directly

but

only

through

the

reoording

circuits

of

unit

1.

Any

556

tapes

which

are

punched

will

have words

arranged

in

order

of

ascending

drum

addresses

and

extensive

use

of

ditto

blocks

will

be

made

to

shorten

the

tapes.

Some

of

the

uses

of

this

program

may

be:

1.

copying

blocks

from

one

tape

unit

to

another

just

to

make

sure

that

there

is

always a good

copy

of

the

magnetic

tape

programs.,

2.

modif,-ing a

block

on

magnetic

tape

by

reading

in

only

the

mod

tape,

3.

obtaining

a

556

tape

containing

the

same

information

as

a

block

on

MT.

4.

shortening

(i.e.,

optimizing)

556

tapes,

5.

inserting

or

lengthening

a

block

on

MT

without

reading

in

more

than

one

556

tape,

6.

in

eombina

tioD

with

an

input

program

capable

of

reading

in

blocks

from any

MT

unit,

testing

extensions

and

modifications

of

the

as

II

system.

Also

entirely

new systems

could

be

recorded

on

any

MT

unit

for

testing.

Best

101 C.

Optieal

Properties,

of

Thin Metal Films: Loeb,.50

hours;

Riehmond,

25

hours;

WI,

223 minutes

The

present

stage

of

research

in

electronic

digital

computation

applied

to

the

optical

and

electrical

properties

of

thin

metal

films

was

completed

during

the

past

month.

The

final

production

runs

have

taken

place.

and

the

will

be devoted

to

the

analysis

and

publication

of

the

results.

106 C.

MIT

Seismic

Project:

Briscoe,

36

hours;

Simpson, 50

hours;

Walsh,

60

hours;

WWI,

362

minutes

Huch

of

our computer time during

the

past

4 weeks has been devoted

to

further

analyses

of

earthquake

records

along

the

lines

indicated

in

reports.

In

the

future,

progress

on

this

aspect

of

our

work

will

be

reported

under a

new

problem number (1192).

The

recent

work

of

this

type,·

consisting

of

frequency

studies,

has

indicated

that

the

so-oalled

"T

phase" encountered

on

these

records

possesses

the

dispersive

characteristics

of

Rayleigh waves

at

the

ground

water

interface,

but

further

verification

of

this

theory

is

needed. A

similar

stuqy

is

contemplated

for

the

minor tremors

known

as

mieroseisms.

Progress

on

other

projects

under

way

have

suffered

some

delays

due

partiall:jr

to

the

school

term ending and

loss

of

personnel.

In

addition

there

has been

difficulty

in

computing

largest

eigenvalues

oinon-symmetric

M-287

0 page 4

matrices.

The

WWI

subroutine

was

tried

unsuccessfu~

and

then

modified

for

further

trial.

When

this

failed

to

converge

satisfactorily

a

slightlY

different

iteration

procedure

was

programmed and has been checked

out.

Modifications

in

oorrelation

procedures mentioned

previously

are

being

included

in

a

new

gene~alized

correlation

program which

is

presently

being

written.

109 c.

Fighter

Gunsight

Calibration:

WWI.

27

minutes

A

horizontal

plane

pursuit

course

program

has

been

written

which

computes

prediction

times

for

a

two-~gyro

gunsight

and a

three-gyro

gunsight.

A

slant

plane

pursuit

course

program has been

written

which

com-

putes

prediction

times

for

a two-gyro

gunsight

and a

three-gyro

gunsight

and

correction

time

ratios

for

the

latter

gunsight.

A

test

problem has

been

run

using

these

two

programs

with

com-

parable

initial

conditions

for

each.

The

results

are

being analYzed and

will

be

included

in

the

final

report

on

this

problem.

It

is

planned

to

make

a number

of

runs

using

the

slant

plane

pursuit

course

program

with

the

airplane

and

target

initially

in

a

slant

plane.

113 C.

Shear

Wall Analogy, Simultaneous

Linear

EQBations: Sytiney, 15

hours;

WI.

11 minutes

The

final_

r~port

on

this

problem

is

being

written.

All

production

runs

are

completed.

119 C.

Spherical

Wave

Propagation:

Ralston,

5

hours;

WWI,

74

minutes

Magnetic

tape

troubles

have

prevented

the

obtaining

of

an7

results

for

the

revised

problem mentioned

in

the

However,

sufficient

results

have been

obtained

for

the

initial

problem proposed

and

the

problem

may

now

be

considered

to

be

terminated.

120

D.

The

Aerothennopressor:

Gavril,

10

hours;

WI,

223 minutes

During

the

.past month, computations

of

aerothermopressor

performance

for

various

Mach

Number

variations

have

continued.

In

these

calculations

the

Mach

Number

is

arbitrarily

prescribed

as

a

function

of

gas

stream

.

humidity,

thereby

defining

the

process.

The

data

thus·rar

accumulated

is

being

studied

to

determine what

Mach

Ntu,nber

variation

(or,

effectively,

what

geometrical

shape

of

duct)

leads

to

best

performance,

that

is,

maximum

pressure

at

exhaust.

A

new

calculation

procedure has been

written

and

successfully

tested

which

to

a

great

extent

allows

Whirlwind

itself

to

determine

the

Mach

Number

variation

for

best

performance 0 This

routine

has

the

advantage

of

reducing

the

uncertainty

and

arbitrary

nature

of

the

procedure

suggested

above and

reduces

the

probl~m

of

determining

best

performance

to

a

routine

M-2870

page 5

WWI

calculation

in

the

least

possible

computer

time.

The

method

of

calculation

is

essentiall.y

a

Ritz

type

of

numerical

solution

to

the

variational

problem

of

determining

the

Mach

Number

var-

iation

. .

2..

3 4

M -

MO

= 01(X-11) + 02(X-Xl)

+03(X-X

1) +04(X-X1)

which

renders

the

final

pressure

p£ a

maximum.

The boundar,y

condition

is

the

requirement

that

the

final

Bach Number,

Hf

t

be

fixed.

For a

given

set

of

values

of

the

parameters

Cit

the

value

of

Pf

may

be

calculated.

Four

additional

sets

of

calculations

lead

to

values

of

the

partial

deriv-

atives

~

Pr/~

01 t

-oPf/

a02 •

C>~f/(}03

'

fJP

f

/'dC

4

and hence

to

the

gradient,

Q"p =

«()

Pr/()01)'l + (?J.Ptla02)j +

••••

The

parameters

01Aare

then

simultaneously

adjusted

so

that

a

prescribed

increase

in

Pr

(~)

occurs

for

the

least

change

in

Mach

Number

variation.

This

is

accomplished

b,y

determining

the

increment

in

0i

by

the

equation

~d

aOi

p

\~p\2

This

calculation

procedure

shows

great

promise

for

providing

a

routine

method

for

aerothermopressor

design

contained

on one 556

WWI

tape.

The

time

required

to

effect

one

systematic

alteration

in

the

Mach

Number

variation

is

seen

to

be

5

times

that

of

a

single

run,

the

latter

requiring

on

the

order

of

8

minutes.

The

gross

time,

however,

is

con-

siderablY

less

than

that

required

for

the

larger

number

of

arbitrarilY

specified

eases

which would

in

general

be necessar,y.

123 c •

.rEarth

Resistivity

Interpretation:

.WI.

105

minutes

The

program

for

ana~sis

of

the

Slichter

kernel

has

been

tested

with

more

kernel

data

in

an

effort

to

learn

more

about

its

behavior.

In

addition,

a program change

has

been

made

which

allows

an

indefinite

number

of

iterations,

in

contrast

to

the

maximum

of

25

which was allowed

previously,

and which

increases

the

parameter-change

multiplier

ever.y

20

iterations.

Four

different

kernels,

three

theoretical

and one

observed.

were

tried.

Two

of

the

four

calculations

(a

four-layer

kernel

and.

the

observed

kernel)

<termilia.tea

because

of

error

increase

after

about

three

iterations,

indicating

too

large

an

initial

parameter-change

multiplier.

These have

not

yet

been

retested.

Another

of

the

four_ a

theoretical

three

layer

ease,

was

run

for

25

iterations

and seemed

to

be

converging

to

the

known

paramet~r

values.

The

fourth

kernel

tested

was

also

a

theoretical

three

layer

ease.

It

was

run

for

21

minutes,

giving

90

iterations,

and

the

parame~er

change

multiplier

was

increased

ever.y

20

iterations.

The

final

page 6

value

of

the

multiplier

was

0.4,

and

it

could

have been

larger.

The

correct

solution

was

obtained

on

the

40th

iteration,

but

the

calculation

was

continued.

On

the

50th

iteration

the

resistivity

and

thickness

of

the

second

layer

went

negative

and

the

error

increased.

Beyond

this,

the

error

continued

to

decrease

ver.y

slow~,

and

the

second

layer

resistivity

and

thicknA~s

oscillated

about

zero.

The

first

layer

thickness

oscillated

around

1.03,

whereas

the

true

value

was

1.0.0.

The

significance

of

this

result

will

be

discussed

in

the

report.

126 c. Data Reduction: Hamil

ton

t

20

hours;

Ross,

10

hours;

WI.

19

minutes

The

past

period

has

been used

entire~

in

abortive

attempts

to

make

a

test

run

of

the

complete Data Reduction Program

with

the

mistake

diagnosis

routine.

Tape

preparation,

conversion,

tape

loss~and

FETR

and

drum

difficulties

have

precluded

even a

post

mortem.

131."

Special

Problems

(Staff

training,

demonstrations,

etc.):

Kopley,

hours;

Ackl~y,

"16' ho'tlrsd

;

Denman,

26

hours;

tM,

lSI

minutes

Tours:

On

June

1,

twenty-two

people

attended

the

monthly

tour

which

included

Flexowriter

and computer

demonstration,

a

tour

of

Whirl-

wind

installations,

and an

informal

twenty minute

talk.

Kopley

132.

Revision,

Extension

and

Testing

of

Subroutine

Lil?r~!y',J!~!!9~1~

Programs

for

Obtaining Data

for

the

Numericall:¥'

Con~!:,~C?~!-.~~

__

~;!!~~g

Machine;

routine

numerical'

and

logical

operations:

Runyon, 40

hours;

Kopley,4

hours;

WWI.64l1',irilites

':','

;

'"

.

A

subroutine

for

reading

milling

machine

tape

into

the

computer

was

successfully

tested.

This

subroutine

has been

incorporated

into

pro-

grams and

is

being

used

to

check a long

tape.

MOre

errors

have been found and

corrected

in

a program

for

preparing

tape

for

airfoil

templates.

Results

of

the

latest

version

are

yet

to

be

received.

A program

for

preparing

tape

for

a

conic

wing

section

will

be ob-

tained

by

modifying a program

for

a

conic

test

piece,

which

has

prev-

ious~

performed

satisfactorily.

141.

Matrix

Diagonalizatiot:l and

Inversion

Subroutine:

Mahoney, 67.5

hours;

Dempan,

5

hours;

WI,

46 minutes

The

matrix

diagonalization

and

inversion

subroutine

(ISR

MAS)

is

being

rewritten

in

order

that

it

may

operate

faster.

This program

re-

duces a symmetric

matrix

to

diagonal1brm, determines

the

eigen

values

and

finds

the

inverse

of

the

matrixo

All

interpretive

instructions

in

the

program

now

in

the

present

librar.y

version,

except

those

essential

to

arithmetic

operations

performed

on

the

elements

of

the

matrix,

have

been

replaced

by

WWI

instructions.

The

length

of

this

program,

now

being

tested,

will

not

exceed

that

of

the

original

program.

M-2870

page 7

142

D.

A Study

of

Shock

Waves:

Sydney,

60

hours;

WIt

444

minutes

Production

runs

are

continuing

on

the

last

phase

of

this

problem.

Four more hours

of

computer time

are

required

to

complete

the

project.

143

D.

Vibrational

Frequency Spectrum

of

a Copper

Cgstal:

WI,

77

minutes

The

Spectrum

was

recalculated

for

copper assuming a

central-force

action

between atoms.

The

appropriate

central-force

atomic

constants

were

taken

from

the

work

of

Leighton

for

Copper

as

set

forth

in

Reviews

~Modern

Physics,

Janu3r,y 1948.

The

result

exhibits

the

same

general

character

as

the

more

exact

spectrum

calculated

earlier

on

the

Born

Theory

of

Lattice

Dynamics ,a.s to,befex.,pec'ted.

The

details

are,

however t

quite

noticeably

different.

The

results

of

both

theories

(Born and

Leighton)

will

appear

in

the

report.

147

Co

Energy Bands

in

Cqstals:

Howarth,

80

hours;

WI,

575 minutes

Alterations

have been

made

in

the

routine

described

previous~

to

calculate

and

solve

the

secular

equations

which

constitute

the

last

part

of

this

problem.

Elimination

of

errors

in

these

has

proceeded slowlY.

When

completed.

the

computation time

will

effectively

be

reduced by a

factor

of

two.

The

solution

of

the

secular

equations

submitted

by Dr.

F.

Herman

of

R.C.Ao

has been completed

satisfactorily.

149

C.

Digital

Methods

of

Detecting

Signal

froll

Noise: Dinneen,

15

hours;

WI.

51

minutes

A program

was

written

to

test

a

detector

which measures

the

density

of

ones

(signal

returns)

within

a given

interval

as

this

interval

moves

throughout

the

region

under

study.

Whenever a

given

threshold

(corresponding

to

a

density

of

ones)

is

exceeded,

the

start

of

a

signal

region

is

recorded.

When

the

densit.y

of

ones

falls

below

this

region"

the

end

of

the

signal

region

is

recorded.

No

results

have

been

obtained.

The

post-mortem

is

being

studied

to

determine

the

error.

155

D.

Synoptic

Climatology: '

MUler,

60

hours;

Blood.

80

hours;

Demurjian, 5 hours;

WI,

727

minutes

Multiple

correlation

coefficients

have been computed

for

91

grid

points

signifying

the

24 hour

predictability

of

surface

pressure

using

past

surface

patterns

as

predictors.

Various combinations

of

grid

sizes

have been

investigated

and,

in

all,

the

results

have proved

extreme~

interest:iing.

It

was

necessar,y.

in

carr,ying

out

the

above

analysis,

to

write

a

program which would

solve

an

n x n

matrix

by

means

of

the

Crout

method

for

n

up

to

33.

It

is

felt

that

this

program could

easily

be

exte~ded

to

an n x n

matrix

where n

is

limited

only

by

the

auxiliar.y

storage

units

or

roughlY n = 150.

Mi-2870

page 8

Future

plans

are

directed

toward

constructing

a

four

dimensional

model

of

the

atmosphere

(x,

y,

z,

t)

to

predict

and

specify

surface

temperature

and

rainfall.

Miller

This'problem

will

attempt

to'specify

and

predict

daily

surface

temperatures

at

five

selected-

stations

using

circulation

patterns

over

North America.

Using

14

Z's

daily

to

describe

the

circulation

pattern,

regression

equations

weredsrived

fpr

prediction

and

specification

for

each

of

the

five

stations.

The

eqJations

are

of

the

form:

Correlation

coefficients

and

root

mean

square

errors

are

also

calculated.

The

first

month

tested

showed a

correlation

of

0.77

for

specification

of

daily

8

AM

temperatures.

ApproximatelY

1/30th

of

the

planned problem

is

completed.

Difficulties

encountered mainly

involved'scale

fa.ctoring.

159 D

•.

Water

Use

Ll a

Hydroelectric

System:

Little,

200

hours;

WI,

300 minutes

A program has been

written

and

fairly

well

debugged

to

determine

the

best

way

of

operating

the

given

~dro

system. A program which

uses

this

operation

to

run

the

system

on

the

historical

river

flow has been

written

but

not

debugged.

161 C. Res onse

of

Mass-Plastic

Sydney,

20

hours j

ww.:r::,

stem

to

Transient

Loadin

Production

runs

are

completed

on

this

problem.

The

final

report

is

now

being

written.

162 C.

Determination

of

Phase

Shifts

from Experimental

Cross-Sections:

Campbell,

24

hours;

Denman,

.5 hour;

Wilt

20

minutes

A phase

ghift

analysis

is

being

made

of

the

elastic

scattering

of

protons

by 01•

From

a

partial

waveEKpansion

of

the

differential

eross-section

those

phase

shifts

required

to

provide

a

fit

to

a number

of

experimental

angular

distributions

will

be

determined.

As

part

of

a preliminar.y program

to

obtain

plots

of

the

sum

of

the

squares

of

the

percentage

errors

in

the

cross-section

as

a

function

of

each phase

shiftJone

tape

was

successfully

run

giving

the

error

as

a

function

of

51'

the

P

Wave

phase

shift.

M-2870

163 c.

Ferrite

Phase

Shifters

in

Rectangular

Wave

Guide:

hours;

WWI.

61

minutes

. page 9

Button,

120

The

theoretical

investigation

of

the

nonreoiprocal

phase

shifter

at

lower

frequencies

was

described

in

the

last

report.

Part

I

of

this

p~ogram

provided

design

information

for

two

phase

shifters

using

two

different

ferrite

materials

at

this

lower

frequency.

Further

analysis

of

these

results

has

revealed

an

excellent

method

of

checking

the

theor,y

experimentally.

It

has been found

that

the

primary

characteristic

9

the

differential'

phase

shift,is

of

opposite

sign

for

these

two

phase

shifters.

The

point

of

sign

reversal

may

be found

either

as

a

function

of

a

ferrite

characteristic

(the

saturation

magnetization)

or

of

the

externally

applted

DC

magnetic

field

intensity.

This

sign

reversal

would

not

be

encountered

in

phase

shifters

operating

at

the

original

high

frequency

of

9375

MO.

Formerly,

only

qualitative

experimental

verification

was

possible

since

the

theory

is

forced

to

deal

with

an

infinitelY

long

ferrite-loaded

transmission

line

while

the

experimental

apparatus

is

necessari~

finite

and

involves

transitions

from empty

to

ferrite

loaded

waveguide.

Correct-

ions

to

the

measured

values

had

to

be

estimated.

Furthermore,

the

funda-

mental formulas

for

ferrite

permeability

parameters

were presumed

to

be

inadequate

at

S-band (3000

Me.)

because a

proper

loss

term

required

at

these

low

frequencies

has

not

yet

been

derived.

~erimenta1

measurement

near

the

point

of

sign

reversal

is

expected

to

improve

the

relation

between

theoretical

and

experimental

findings.

The

problem

of

6

otlstructing

lossless~'non-reciprocal'

phase'

shifters

at

frequencies

lower

than

9000 HcpS.

is

now

being

investigated·.'

Data

.obtained from

Part

I

(evaluation

of

the

differential

phase

shift)

and

Part

II

(evaluation

of

the

attenuation

constant

at

9375

Mcps.)

are

leading

to

the

formulation

of

design

conditions

for

these

lower

frequency

devices.

Low

frequency

non-reciprocal

transmission

lines

are

expected

to

be

used

much

more

widely

in

practical

applications

than

the

researc:h

devices

which have

been

developed

at

X-band.

The

difficulties

encountered

in

low

frequency

devices

may

be

described

briefly

as

follows:

They must be

operated

far

enough

away

(on

one

side

or

the

other)

from

the

gyromagnetic resonance peak

to

be

near~

lossless.

This

is

possible

at

X-band because

the

resonance

absorption

peak occurs

at

a

value

of

DC

magnetic

bias

of

about 2500

oersteds.

An

operating

bias

of

several

hundred

oersteds

is

suitable.

However,

this

resonance

value

of

bias

is

proportional

to

operating

frequency.

At

frequencies

below 3000

Mcpso

the

magnetic

bias

would have

to

be

much

too

small

to

provide

a

usable

differential

phase

shift.

It

therefore

becomes

clear

that

these

devices

must be

designed

to

operate

at

values of

magnetic

bias

which

are

larger

than

the

resonance

value.

The Whirlwind programs

are

now

being

used

to

predict

operating

characteristics

at

frequencies

as

low

as

375

Mcpso

About 22 minutes

of

computer time

will

be

required

to

complete

the

investigation.

After

this

Whirlwind

data

has been

evaluated,

the

use

of

Whirlwind

time

for

obtaining

solutions

for

the

phase

shifter

problem

will

not

be

required

for

several

months o

Theoretical

work

will

be suspended temporaril,y

to

await

further

experimental

verification

and development.

M-2870

page

10

166

D.

Construction

and

Testing

of

a

Delta-Wi~g

Flutter

Model:

Grantz,

110

hours;

Porter,

4 hours;

WI,

234

minutes

Substantial

progress

was

made

during

this

period.

The

basic

program

for

the

first

cycle

of

the

Error

Iteration

Procedure has

been

successfully

tested.

Two

additional

tapes,

representing

modifications

to

the

basic

program

to

enable

it

to

handle

the

succeeding

cycles

with

input

from magnetic

drum

storage

rather

than

from a

separate

tape

fed

into

high

speed

storage,

have been completed and

satisfa'ctorily

tested.

The

original

tapes

have been

altered

for

use

with

CS

n and

rearranged

and

combined

so

that

the

complete program

is

represented

on

four

tapes.

The

first

tape

is

a Flexo

tape,

designed

for

direct

read-in,

which

contains

the

input

to

the

p~sical

s.ystem'represented

by

the

three-

bay

lattice

structure;

that

is,

the

given

influence

coefficients,

or

de-

flection

pattern

that

the

model

is

to

conform

with,

and

the

first

estimates

of

the

torsional

and bending

stiffnesses

of

the

component

members

of

the

lattice.

This

is

a

compafatively

short

tape

which can be

easily

and

quickly

modified

should

theneed

arise.

The

second

tape

is

in

556

form and

contains

most

of

the

geometry

of

the

system;

that

is,

it

describes

..

the

three-bay

lattice

in

terms

convenient

for

the

subsequent

manipulations.

The

third

tape

is

also

in

556

form and corresponds

to

the

computation

of

the

~C1S;

it

computes

the

influence

coefficients

corresponding

to

the

first

trial

stiffnesses.

compares them

with

the

given

values,

and

prints

out

the

errors.

The

last

tape,

in

556

form, corresponds

to

the

formulation

of

the

~XIS;

if

the

trial

influence

ooeffioients

are

not

close

enough

to

the

given

ones,

corrections

must

be

made

to

the

trial

stiffnesses.

This

tape

also

incorporates

a

recent

mcdificationto

the

Crout

method

matrix

inversion

routine

which

solves

the

final

set

of

equations

to

yield

the

~xts--the

now

present

check column

gives

a quan-

titative

indication

of

how

large

the'

accumulated

errors

are

by

observing

how

far

a second

set

of

solutions

to

the

equations

deviates

from a column

of

ones.

The

entire

program, which

will

be

run

on

the

computer

some

time.lib.

the

very-

near

future,

works

as

.follows:

for

the

first

cycle

all

four

tapes

are

read

into

the

computer

with

the

PETR

and each succeeding

cycle

will

require

the

reading

in

of

only

the

last

two

tapes

0

The

preliminary

output

from each

cycle

is

stored

on

tne

drum

right

over

the

corresponding

output

from

the

therefore

itis

unnecessary

to

repeat

the

reading

in

of

the

initial

input

which

is

on

the

first

two

tapes--only

the

tapes

corresponding

to

the

actual

computations

(the

last

two) need

be

repeated

for

each

cycle.

In

the

future

more

efficient

use

of

the

drum

is

plannedo

Testing

of

the

completed program has

also

y;elded

useful

infor-

mation concerning

the

p~s1cal

assumptions which form

the

preliminar.y

input

to

the

system.

It

has been deduced from examination

of

the

test

results

that

when

an

unknown

must be

eliminated,

it

is

much

more

satis-

factor,y

from

the

mathematical

point

of

view

to

assume a

given

cross-

section

for

a member,

thus

defining

the

ratio

ot

torsional

stiffnesses

of

two members. Although

both

types

of

assumption

are

reasonable

from

the

physical

point

of

view,

it

has

been

found

that

the

latter

type

results

in

a

very

poorly

conditioned

final

matrix

which

causes

the

H-2870 page

11

values

for

the

~

X'

s

to

become

unnaturally

large

due

to

error

f!l

ccumulation.

It

is

expected

that

thelll-conditioning

of

the

final

matrix

will

be

..

remedied;

in

the

meantime

the

check column

for

the

solution

of

the

final

set

of

equations

will

be a welcome

addition

to

the

program

in

this

respect.

A

test

run

was

made

with

accordingly

modified

input

and

the

6.

C' s

were observed

to

have

slightly

better

values

than

previouslYo

The

real

test,

of

course,

is

what

effect

on

the

~

X'

s

the

revised

estimates

have

and

it

is

expected

that

this

will

be

ascertained

in

the

ver.y

near

future.

168 c.

Indicial

Downwash

Behind a Two-Dimensional Wing: Hobbs,

160

hours;

WI,

43 minutes

The

solution

for

the

indicial

downwash,

that

is,

for

Q =

0,

has been completed

for

four

points

behind

the

wing.

The

rema¥ning

problem

consists

of

utilizing

the

indicial

downwash

to

obtain

by

numerical

integration

the

downwash

for

values

of

Q

other

than

zero.

V

169

utilizin

Design:

Thirty-two

SWitching

functions

on 5

v~ables

with

an average

of

about

ten

terms have been reduced

by

the

existing

reduction

program.

The

average

computation time

per

problem

is

approximately 1 second

or

about

1000

times

as

fast

as

a

rapid

worker.

However,

the

results

reveal

that

the

last

of

the

rules

that

have

been

u~ed

~s

not

sufficientlY

general.

Examination

indicates

that

it

can

be

rewritten

in

a

much

more

general

form.

In

addition,

the

generality,

by removing unnecessar.y

restrictions,

permits

a

shorter

program.

The

new

program has been

written

but

the

results

of

the

tests

have

not

yet

been

received.

Methods

for

obtaining

the

most minimal form

or

one

that

is

as

simple

as

any

other

have been

studied

but,

as

yet,

none seems encouraging

enough

for

programmingo

172

B.

Corbat6,

During

the

last

four

weeks,

all

the

integrals

necessar.y

to

make

tight-binding

calculations

of

Graphite

and Diamond have been computed.

In

addition

Dr.

Howarth's (Problem #147)

routine

for

solving

secular

equations

of

the

non-~

type

has

been

adapted

for

use

with

a program which

gene~ates,

from

the

essential

integral

values"the

matrix

elements

arising

in

the

Graphite

secular

equation.

This arrangement

is

now

being

used

to

produe~

final

results

for

Graphite

o

The

subroutines

for

calculating

integral

values

have been

revised

in

order

to

reduce machine time by

30%.

Pending

final

tests

of

these

revisions,

the

tapes

will

be

consolidated

for

the

convenience

of

possible

future

use.

H~2870

page

12

176

B. Connector

Provision

in

Automatic Telephone Exchanges:

Porter,

3

hours;

WI,

148

minutes

Since

the

last

report,

two

further

solutions

have

been

obtained

for

5 =

100,

and

selected

values

of

P.Results

for

a

total

of

8

values

of

P

are

now

available,

and

these

are

sufficient

for

a continuous

set

of

curves

for

S =

100,

over

the

desired

range

of

P.

The

first

of

seven

runs

for

S = 200 has

also

been

run,

and

satisfactor.y

results

obtained.

Allowance

for

a

typical

distribution

of

calling

rates

produces

the

following

equations:

s

...l2-.. •

l-p

( ) (

..JL

.w+k-l)

S-w-k+l

~

1-

l-p

0

5-1

w=O

3a

L p = 1

k

3b

k=O

S-l

S-k-l

L

~

(S-k-w)~Pk-_

(1-

...:2..

1_ •

k+w

)

P 8-1

k=N

w=O

4

These

equations

have been

solved

for

S=25, and p=O.02(0.Ol)

0.20.

Results

differ

so

little

from

those

given

by

the

original

set

of

equations

that

it

is

clear

that

the

effect

of

the

distribution

of

calling

rates

may

be

safely

neglected

o

\.

179 c.

Transient

Temperature

of

a Box-Type

Beam:

Schmit, 8

hours

t

'WI,

22 minutes

An

investigation

of

the

effect

of

variable

speoific

heat

estab-

lished,

for

this

particular

case,

that

this

effect

is

not

important.

The

large

increases

in

contact

admittance

produced by

the

steel

screws

in

the

alUminmmstructure were

estimated.

The

effeotive

absorptivity

of

the

exposed

surface

of

the

box

beam

was

back

calculated

from

test

data

(measured

temperature

and

total

heat

energy).

It

was

found

that

the

effective

absorptivity

was

considerab~

less

than

values

established

experimentallY

for

the

same

surface

prior

to

the

field

test.

The

179-86-7

tape

was

prepared.

The

heat

input

information

was

based

on

the

effective

a~sorptivity.

the

total

energy

delivered

as

measured by a <calorimeter, and on normalized

heat

flux

time

histories

obtained

by

radiometer

measurements.

The

high

contact

a

elmi

ttanoe

values

due -to

the

steel

screws were

also

incorporated

in

this

modification.·

The

179-86-1

tape

ran

successfullY

in

22

minutes.

H-2B70 page 13

An

error

in

the

basic

difference

equations

which had

not

previouslY

been

apparent.

was

substantial~

magnified by

the

increased

contact

admittance

terms.

The

p~sical

equivalent

of

this

error

was

a

heat

source

in

the

flange

near

the

web-flange

jointo

The

179-86-8

tape

was

prepared

eliminating

this

erroro

The

entire

problem

was

reviewed and

the

179-86-8

tape

prepared

independent

of

all

tape

is

now

being

tested.

IBO.Bo

Crosscorrelation

of

Blast

Furnace

Input-Output

Date: iIWI,"

213

minutes

During

the

past

four

weeks

the

summer

session

crosscorrelation

program

was

rewritten

for

CS

II

and combined

with

the

Ross

Fourier

Trans-

form program 0

When

optimizedp

the

combined program

will

permit

the

eval-

uation

of

the

cross

or

auto-correlation

of

a

function

and

the

Fourier

Transform

of

the

resulting

correlation

function

with

a

minimum

of

operating

and

tape

preparation

efforto

Computer time

is

also

minimized

since

onlY

one

operating

period

is

required

for

the

complete

solution

of

a.

given

correlation

transfonnation

problem.

It

is

expected

that

the

problem

will

be

terminated

within

the

week

period.

181

Co

Perturbed

Coulomb

Wave

Functions:

Paul,

48

hours;

'WIt

40

minutes

(Computer time from

April

19-May

16)

N~clear

scattering

in

general

remains

the

most

direct

w~

to

get

information

about

nuclear

forces.

Of

particular

interest

is

the

comparison between

neutron-deuteron

scattering

and

proton-deuteron

scattering

because

it

may

be assumed"that

the

forces

involved

in

the

two

cases

differ

I

only

by

the

addition

of

an

electrostatic

interaction

between

the

ch~~ged

proton

and

deuteron.

The

validity

of

this

assumption

requires

that

the

specifically

nuclear

force

between two

protons

be

the

same

as

that

between two

neutrons,

and'its

proof

or

disproof

is

of

basic

interesto

An

analysis

of

the

nuclear-deuteron

scattering

data

requires

knowledge

of

the

scattering

that

would

result

with

onlY

an

electrostatic

interaction

between

proton

and

deuteron

o

In

the

case

of

two

point

charges

t

this

is

given

by

the

so-called

Coulomb

wave

functions,

tabulated

in

same

detail

by,

the

U.SoBureau

of

Standards.

For

proton-deuteron

scattering.

the

charge

distribution

of

the

deuteron

cannot

be

considered

at

a

point

and

the

corresponding

functions

must

eome

directlY

from

the

linear

seeond-

order

differential

equation

by

a numerical

solution.

With

these

functions

at

hand&

one

can

use

the

neutron-deuteron

scattering

data

to

predict

proton-deuteron

scatteringo

Sufficient

data

of

both

kinds have been

taken

in

this

laboratory

to

make

a

significant

comparisono

A

test

of

this

so-called

charge-symmetr.r

of

nuclear

forces

may

be

obtained

by comparing

p-D

and n

...

D

scattering

0

To

do

this

& one compares

the

experimental

p-D

phase

shift

&

with

that

calculated

fram

CP'Cb)'"=

G-'

(b)+

cof~'F'

(b)

'f)

(I;,)

G-

Cb)..c.of-8 F

(b)

'.~

M-2870

pige

14

where G(r) and

F(r)

are

the

regular

and

irregular

solutions

of

the

p-D

wave

equation

-k

/-e.

0...

R.~

]c(J=o

k=wave no.

R~rotonic

phor

radius

a=Scattering

length

The

numerical

solution

of

this

differential

equation

has been done on

WWI

by

the

following procedure.

The

solution

is

started

by a power

series

Regular

Sol.

i

bo=O

b=].

1

Irregular

Sol.

f b

0"'.1.

...

l bl=D

~.

It

Co

~

(.,x)

cQ.

~(l()=~~-

I-~·~

J

The

regular

and

irregular

solutions

have been, done

successfully

for

O~r"'.3

at

intervals

of

.05,

for

a

,=

.216,

Rl

= 201.47 and.R

==

.447,1

and

k~:;:

.2158"k;:

=.4316" k3"=.6474,

J¢=.8632.

k!

=

1.0790,2k~

=

1.618511

k~=

2.1580.

TlP.s

concludes

the

problem.

183 D.

Blast

Response

of

Aircraft:

.

Sternlight,

40

hours;

Demurjiant

.25 hour;

WWI.

144 minut'es

TWo

more

parameter

tapes

were

run

successfully

during

this

perioq,.

The

automatic

.6

1::

reduction

program

is

giving

trouble.

A.

new

f~ature

was

added

to

allow

for

linear

curvature

when

desired

ili

the

calculation.

Results

are

almost completed

for

time-history

runs 0

When

the

A

't'

reduction

section

is

checked

out,

peak

print

runs

will

begin

on

a ;production

basis.

184

D.

Scattering

Electrons

from

HYdrogen:

WIt

175 minutes

A program has been

written

to

perform

the

two

dimensional

integrations.

It

is

being/tested

on

an

ana~~cal~

evaluable

integrand.

M-2870

page

15

Another program

to

calculate

several

of

the

integrands

involved

is

being

tested.

187 c. Response

of

a Fuel-Flow

Controller:

Merwin, 8

hours;

WWI.

48

minutes

The

following system

of

first

order

nonlinear

ordinary

different-

ial

equations

d~scribing

the

response

of

a

fuel-flow

controller

has been

so~ved

on

WWI

making use

of

CS

II.

It = f

(z)

UJ

1=

const.

GJ

2

is

quadratic

function

of

x40

The

solution

did

not

prove

to

be

p~sica~

realistic

so

a

second problem

is

now

being

solved

using

the

same

equations

with

different

constants.

The

Gill

Method1)

',an

adaptation

of

the

Runge-Kutta t has been

used throughout

the

program.

188

Co

EUac:t

of..

Gravi..t.y

on

ReJ..ative

..

Wat.er.

Producti.on

in,

Oil

Reservoirs

:

Porter!)

14

hoursi

WIt

374 minutes

In

a

large

number

of

oil

reservoirs

water,

which

underlies

the

oil,

is

produced along

with

the

oil.

The

relative

amount

of

water

production

depends

for

a given

geometrieal

configuration

on

the

effect

of

gravitational

forces

relati.ve

to

the

imposed

pressure

forces.

In

order

to

enhance

the

effect

of

gravity

the

common

practice

is

to

plug

the

bottom

part

of

the

well

bore

so

that

fluid

is

produced from near

the

top

of

the

formation;

in

some

instances

the

imposed

pressure

forces

are

held

to

a

minimum.

This problem has been undertaken

in

an

effort

to

gain

quantitative

information

concerning

the

effect

on

relative

water

production

of

partiall7

penetrating

wells,and, reduced

pressure

forces.

The

interesting

feature

of

this

problem

is

the

demarcation

surface

separating

the

oil

and

the

water.

This

is

a

free

surface

in

the

sense

that

its

location

is

not

known

beforehand

but

must be determined

in

th~

course

of

the

computation.

The

pressure

P

in

either

the

oil

or

water

zones obeys

the

following

equation:

( 1 )

M-2870

page 16

where r

is

the

radial

distance

from

the

center

of

the

well

bore and

~

1s

the

vertical

depth below

the

top

of

the

reservoir

o

Since

no

fluid

moves

across

the

upper and lower boundaries

of

the

reservoir

nor

across

the

cylindrical

surface

where

the

well

bore

is

plugged,

simple boundar.y

conditions

in

terms

of

the

normal

derivative

of

Pare

obtained

at

these

boundar1eso At

the

well

bore and

the

outer

rad~us

of

the

formation

the

pressure

is

a predetermined

function

of

deptho At

the,

interface

between

the

oil

and

the

water

the

pressure

in

the

water

zone

must equal

the

pressure

in

the

oil

zone; furthermore,

since

no

fluid

moves

aeross

the

interface,

the

normal

derivatives

of

water

and

oil

pressures

are

simply

The

proposed method

of

solution

is

to

assume a

fixed

interface

position

and

solve

equation

(1).

for

the

water

and

oil

zones

independently.

The

interface

position

and

the

pressure

difference

between

the'well

bore

and.

the

outer

radial

boundary

are

then

adjusted

s 0

that

water

a~d

oil

pressures

at

the

interface

are

equal

in

the

least

squares

sense~

The

procedure

is

then

repeated~with

the

new

interface

position

and

new

bounclar7

pressures

0

In

the

actual

calculation

the

procedure has been

simplified

by

making a

linear

transformation

on

pressure;

a

logarithmic

transformation

has been

made

on r

to

increase

accuracy

near

the

well

boreo

Preliminar.y

results

have

indicated

that

the

logarithmic

transfor-

mation

greatlY

decreases

the

rate

of

convergence

of

the

numerical

solution

of

equation

(1);

furthermore

it

appears

that

equation

(1)

must be

solved

with

a

great

amount

of

preCision

before

the

proposed adjustment

of

the

interface

is

stable

o

For

this

reason an

effort

will

be

made

to

improve

the

procedure

of

adjusting

the

interface

fram

p~sical

concepts and

results

tn

the

:lmmediate

future.

.

192

D.

Frequency and Phase Spectrum AnalYsiS

of

Seismograms: Walsh, 60

hours;

WI,

238' m.iimtes

In

this

section

it

is

desired

to

determine

the

contribution

of

various

specified

frequencies

to

the

reflected

and

refracted

energy

recorded

on

prospecting

records.

It

is

hoped

that

this

knowledge coupled

with

information

of

the

noise

spectrum

will

give

rise

to

signal-to-noise

ratios

which,

will

assist

in

the

design

of

electronic

and mathematical

filters

to

be used

in

putting

these'reflections

and

refractions

into

a

better

relief

in

time.

In

this

part--a$

in

the

above--it

is

desired

to

ascertain

the

spectrum

(both

phase and . frequency)

of

the

various

directo

reflected

and

refracted

body

waves

recorded

at

a

number

of

seismic

stations

situated

on

the

continent

of

North Americao

It

is

hoped

that

if

such

information

were

displayed

in

time

that

a

great

assistance

would be rendered s

eismolo-

gists

in

the

detemination

and "picking"

of

these

phaseso Furthermore,

it

is

intended

to

apply

similar

analysis

to

the

surface

wave

and "coda"

section

of

earthquake

records

in

the

hope

of

determining

the

magnitude

and presence

of

dispersion

J which knowledge would

in

turn

be

instrumental

in

tpe

determination

of

the

type

of

continental

and sub-oceanic

layering.

In

addition

to

the

above a frequene.y

analysis

of

microse~sms

and

the

:so-ca11ed,

·T..J>hase

wiil

be

made

0:

. The

intent

here

is

to

obtain

knowledge

page 17

which should be most

instrumental

in

deciding

the

nature

and

origin

of

these

phenomena.

For

the

past

two

weeks computation was

confined

to

the

calculation

of

traveling

spectra

of

the

various

component seismograms

of

aT-Phase

whose

origin

occurred

on

the

North Coast

of

the

Dominican

Republic.

These

spectra

were

plotted

according

to

a

density

plot

rout~e.

The

results

clearly

show

that

this

phase

has

dispersive

characteristics

peculiar

to

Rayleigh waves

traveling

at

a ground

water

interface.

Addi

tiona1

com-

putation

on

other

phases

originating

at

different

places

must

transpire,

however,

before

definite

conclusions

can

be

made

as

to

the

nature

of

this

phase.

In

the

following

weeks

it

is

intended

to

investigate

by

spectral

analysis

microseisms

recorded

at

Weston, Mass.

in

the

hope

that

additional

proofs

may~

made

concerning

the

nature

and

origin

of

these

waves o 'rhus

far

11

analysis

has

shown

them

to·be

Rayleigh waves

of

high

frequency

gen-

erated

by

standing

ocean waves.

The

following

problems used computer time

but

did

not

report.

107 c.

(a)

Autocorrelation

and (b)

Fourier

Transform,

Evaluate

Integrals

108

C.

An

Interpreti

Te

Program.

144

C.

Self-consistent

Molecular

Orbital

147 minutes

44 minutes

62 minutes

167D.

Products

of

Batch

Distillations

with

Holdup

140

minutes

171 C. lmproved Power Spectrum

Estimates

48 minutes

173 Course 6.537

Digital

Computer

Application

Practise

332 minutes

174

C.

Tight

Binding

Calculations

in

Cry-sta1s 5 minutes

175

C.

Impurity

Levels

in

Crystals

108 minutes

190 D. Zeeman and

Stark

E£fect

in

Positronium

9 minutes

1.3

OQerating

Statistics

1.31

Computer

Time

The

£ollowing

indicates

the

distribution

of

WWI

time

allocated

to

the

S&EC

Group.

H-2870

Programs

C onTers

ion

Magnetic

Drum

Test

Magnetic

Tape

Test

Scope

Calibration

PETR

Test

'

Demonstrations (11.31)

Total

Time

Used

Total

Time

Assigned

Usable Timet Percentage

Number

of

Programs

1.4

Academic Program

as

Programming Course

116

hours,

26

minutes

13

hours,;.(

·P.3

minutes

76····in.inutes

119 minutes

62

minutes

27

minutes

2 hours!

26

minutes

1.36

hours,

39

minutes

l70hours

t

52

minutes

aCYl>

563

page 18

Sixteen

students

were

enrolled

in

the

two-week

CS

programming

course which

started

on

June

7.

A

revised

version

of

the

manual c,o.ver-

ing

the

first

week

11

s

work

was

issued

to

each

student.

Whe:reas

the

origilial,manual

reflected

as

I,

the

revised

version

has

been

made

consis-

tent

wit4

the

provisions

available

in

CS

II~

?reliminary

notes

on

part

of

the

second

week

I s curriculum have been

issued

to

the

students.

These

will

eventually

be

incorporated

in

the

manual covering

the

second week's

work.

Movie

TheS&EC

Group's movie "Making

Electrons

Count" has been

com-

pleted

and

will

be

shawn

to

Dlgital

Computer Laborator.y

personnel

in

the

ver,y

near

future.

2.

CCNPUTER

ENGINEERING

(S.H.Dodd)

~ge

19

Computer

reliability

decreased

during

the

past

biweek~

period,

partiallY

because

of

troubles

introduced

b,y

modifications

of

terminal

equipment. Bugs

in

the

recent~

installed

Ferranti

photoelectric

tape

reader

and

failures

of

the

auxiliar.y

drum

were

the

major

contributors

to

the

lost

computer

time.

The

auxiliary-drum

write-group-seleetion

system

is

being

converted

to

electronic

switching,

digit

by

digit.

It

was

intended

to

do

this

without

interfering

with

the

use

of

the

auxiliary

drum

by

WI.

However,

by

some

means

as

yet

undetermined,

the

information

on group

,11

is

occasionally

altered

during

the

switchover

of

the

write

oircuitry.

Sinoe

the

read-in

program

is

on group 11,

this

has caused programmers

some

inconvenience.

2.1

Wi! System

Operation

(A.J.Roberts,L.L.Holmes)

The

second

magnetic-tape-printout

system has been

completely

checked

out

and has been

placed

in

operation.

,A

memorandum

is

being

prepared

that

will,

explain

the

operation

and

flexibtlityof

the

two

systems.

(D.A.Morrison)

Tube

data

continues

to

be

collected

for

the

computer-reliability

study

0

The

da

ta

is

being

arranged

to

alow

consideration

of

blocks

of

equipment

and/or

the

complete computer. .

Measurements

are

being

taken

of

the

in-out

control

matrices'

input

and

output

voltages

to

assure

proper

operation.

These

matrices

are

not

marginal-checked.

2.12

Typewriter

and Paper Tape

(L.H.Norcott)

It

has

been proposed

that

the

remaining

short-carriage

Flexo be

sent

to

the

factory

to

be equipped

with

a 20-inoh

carriage

and a

pinfeed

platen

capable

of

handling

our

wider

forms.

We

are

awaiting

a

quotation

from

the

manufaoturer

before

proceeding

further.

202 Terminal Eguipment

2.21

Magnetic

Drums

(L.D.Healy)

Work

has been begun

on

the

cirouit

additions

and

modifications

neoessar.r

for

the

addition

of

two groups

of

auxiliar,y

storage

to

the

buffer-drum

systemo

H-2879

3.

ADMINISTRATION

AND

PERSONNEL

New

Staft

(J

.C.Proctor)

page

20

Francis

R.

Drugin

is

working

as

a

DDL

Staff

Member.

He

had

been

employed,

until

recently,

as

a Technical Engineer

tor

IBM.

Gerald

Eo

Mahoney

is

a newDIC

Staff

Member

working

for

Charlie

Adams

for

the

summer.

Until

recently

he

was

a Teaching Fellow.

Ph.D

candidate

at

Boston

University.

Staff

Termination

William

Klein

New

Non-Staff Personnel (R.A.Osborne)

Jean

Bowse

has

joined

Group

62

as a

secretar,y.

Judith

Brask

is

a

new

secretary

in

the

Publications

Ofi"ice. She

is

a

transfer

from

Group

3.

Lillian

Connors

is

a

new

secretar,y

for

Frank R,yder.

Frank Veechia

is

a

new

technician

in

the

Construction

Shop.

Be

transferred

from

the

Flight

Control

Laboratory.

James Johnson

is

also

a

technician

from

the

Flight

Control

Laboratory

who

will

be working

in

the

Inspection

Department.

Jean Randall

is

a

new

member

of

the

Drafting

Room.

M-2870

page 21

LIBRARY

ACCESSIONS

LIST

The

following

material

bas

been

received

in

the

Library,

Barta

109.

Library

Files

No.

Author 'or Source

A9

"0.

He~mer

B136

Brainerd;

Koehler;

Reich

and Woodruff

B227

Jacques

Oattell(ed.

)

0141

IBM

0142 Univ.

of

Illinois

0143

John,L.

Jones

Cl44

Harwood

G.

Itolsq

0145

Joint

IRE

AIEE

ADM

0146

Joint

IRE

AIEE

ACM

C147

Vol.I-

Univ.of

Penn.

Vol.IV

0148 g -

oj

Vol. I Univ.

of

Penn.

Vol.II

0149

Eo

G.

~ogbetliantz

Herbert

B.

Hilton

Charles

R.

Williams

Title

.Randomness

Ultra-High

Frequency Techniques

American

Men

of

Science

IBM

Assembly Program 701

to

701

Illiac

Programming

A Survey

of

Automatic Coding Techniques

for

Digital

Computers

Numerical

Solution

of

Three

Simultaneous

Second-Order

Differential

Equations

Arising

in

the

Low

Energy

Meson

Theory

of

the

Deuteron

Proceedings

of

the

Western Computer Conference

1953

Trends

.in

Computers: Automatic

Control

and

Data

Processing

-

West.Comp.Conf.

1954

Theory and Techniques

for

Design

of

Electronic

Digital

Computers

Report

on

the

ENIAC

On

the

Solution

of

Linear

Equations

by

Diagonalizing

the

Coefficients

Matrix

Problem Programming

for

the

EDVAC

A Review

of

ORDVAC

Opera.ting

Experience

Circular

Functions

C150

0151

C153

C154

0155

Xing,

:Brown,

and Ridenour

Photographic

Techniques

for

Informa.tion

Storage

0156

0157

D41

D42

D37

D38

Russell

J • Brooke

M.

E.

Davis

OE

Fritz

John

NBS

Alan L.

Leiner

NBS

John

Todd

NBS

John Todd

NBS

Applications

of

Computing Machines

to

the

Problems

of

General

Staffs

The

Use

of

Electronic

Data

Processing

Systems

in

the

Life

Insurance

Business

The

Elliott-N.R.DoC

..

Computer 401

Mark

I .

On

Integration

of

Parabolic

Equations

by

Difference

Methods

System

Specification

for

the

DYSEAO

The

Condition

of

Certain

Matrices

II

The

Condition

of

the

Finite

Segments

of

the

Hilbert

Matrix

H-2870

Library

Aooessions

oont.

No.

Author

or

Souroe

D39

John

!odd

NBS

F173-

101

Herbert

J.

Brun

F173

....

111 Harold

A.

Sohapiro

F104

R.

.O;rpser

page

22

!it1e

Experiment on

the

Inversion

of

a 16 X 16

Matrix

Evaluation

of

an'In~egra1

Equation

on WiI

A Cheoker

Playing

Program

for

the

WI

Computer

Computer Searoh

for

Eoonomioa1

Oper~tion

of

a.

Hy'dro

!henna.l

Eleotrio

System

Distribution

List

for

S

&:

EO,

or

Group 6345

Internal

Distribution

External

Distribution-Biweekll

OnlZ

Oe

Adams

R.

Gil~ea

B.

D.

Gavri

1,

31-264

D.

Arden

F.

Helwig

F.

J.

Corbato,

6-003

S.

Best

E.

Hoy

M.

H.

Hellman, Wl-316

Instrumentation

He

Briscoe

D"

Helwig

2 copi····

H.

Carr

D.

Combelio

J&

Cox

M.

Demurjian

H.

Denman

Boo

Bello

Bu

Humphrey

Bo

Fellows

C.

Fleming

J

..

Frankovich

J.

Rapozynsld E.

H.

Jao obsen, 6-410

E.

Ropley

J.

D.

C.

Little,

6-003

M.

Mackey

R.

Miller,

24-510

M.

Marean

(5

copiesProt.

Morse ( 7

copies),

6-109

R.

Parker

of

bi-

A.

RaIson, 2-383

J.

Porter

weekly)

E.

A.

Robinson,

2QE-222

H.

Pareohanian

D.

Ross, Bldg. 32

F.

Shaw

A.

Siegel

J.

Thompson

A.

Vanderburgh

D.

N.

s.

M.

D.

Ste

rnli

gh

t,

East

P.

Hobbs, 41-219

Gravitz,

41-211

M.

Chen, 41-211

Friedman, 24-510

Campus

J

Box

-----~.~---------------------

P.

Bagley -W

F.

Heart

Dr. Loeb -W

J.

Forgie

- B

R. Nelson - B

Dr

Q

Neumann

-

W·

W.

He

Thomas

- W

W.

Wolf'

J.

W.

Forrester

Ro

R~

Everett

So

H'i!I

Dodd

A.

Mo

Falcione

L. Schmit, 41-219

s. Sydney, l-385A

J.

F.

01

Donnell,

12-127

B.

Marrows, Graduate House

B.

Campbell, 20B-l33

H.

Paul,

20B-l33

H.

Lin, 33-316

M.

C.

Newstein, 6-003

.,::'

1.1

.

i:

,

~"

..

)

C\

:'~.

,"

~

cr'.

~

(

..

..1

4

",I

(I

,

l..

• ,

~

\ •

t.~,

~

J

I,

~

!

Dr. Brown, 4-202

F.

M.

Verzuh, 7-306

Bldg. 32 -

Library

:.<

·t~

:.~

~

19

~;

C)

,'L,

14-Notes Only

Ae

Meckler

6-320

KII

Vozof'f 24-032

GD Dinneen

Bldg.

B-333

Linooln

Air

Force Cambridge Research

Center

Attention:

Document

Room--CRQ-SL-1

230 Albany

Street

Mathematics Branch

Division

6 - Librar,y (15

to

20

copies)

Office

of

Naval Research

M's by

Be

Paine--1

dop,r

to

V.

Savio Washington 25, D.C.

AV

s--1

to

Ro

Nelson

One

copy

to

Barta

Reception Desk

(Envelope)

A,P~Kromer

(10

copies

of

M-Notes

for

I.B.M.)

M 2870_Biweekly_Report_June_13_1954 2870 Biweekly Report June 13 1954

M-2870_Biweekly_Report_June_13_1954 M-2870_Biweekly_Report_June_13_1954

Navigation menu

Versions of this User Manual:

Views

Navigation