ThIS REPORT HAS BEEN DELIMITED

AND CLEARED FOR PUBL.C RELEASE

UNDER DOD DIRECTIVE

5200.20 AND

NO RESTRICTIONS ARE IMPOSED UPON

ITS USE AND DISCLOSURE,

DISTRIBUTION STATEMENT A
APPROVED FOR PUBLIC RELEASFj
DISTRIBUTION UNLIMITED,

.]k 3S[FIE-D'

Armed ervices echnical Information 9gency

Reproduced by

DOCUMENT SERVICE CENTER

'1

KNOTT BUILDING, DAYTON, 2, OHIO

This document is the property of the United States Government. It is furnisihed for tte duration of the contract and shiall be returned when no longer required, or uporn recall by ASTIA
Information Agency,
to the following addresp: Armed Services Technical
Document Service Center, Knott Building, Dayton 2-, Ohio.

NOTICE:

WHEN GOVERNUINT OR OTER DRAWINGS, SPECIFICATIONS OR OTHER DATA

XRT9 FOR ANY PURPOSE OTHER THAN 124 CO)NNECTION WTH i. DEFIN1iELY fELATED
GOVERNMENT PROCURMEJNT OPERATION, THE U. S. GOVERNMENT THEREBY INCURS
NO RESPONSIBILITY, NOR ANY OBLIGATION WHLATSOEVER; AND THE FACT THAT THE

GOVERNMENT MAY HAVE FORlMULATED, FURNISHE,

OR IN ANY WAY

0SUPPLIE-D Z-7AQ-

BE REGARI
IS N)ATO
DATA
OR OTHER
SPECIFICATIONS,
SAID DRAWINGS,
Oi2
ayo
-nt
-atr
Sevc
ouo

BY

-uldn-

PERSON OR CORPORATION, ZOR CONVEYING ANY RIGHTS OR PERMISSION TO MANUFACTURE,

USE OR SELL ANY PATENTP

INVeNTION TRAT MAY IN ANY WAY BE RELATED THERETO.

UNCLASSIFIED.

CHARACTERISTICS OF CRYSTAL-VIDEO RECEIVERS

EMPLOYING R-F PREAMPLIFICATION

by
W. E. AYER

'C

Technical Report No. 150-3

September 20, 1956

Prepared under
Air Force Contract AF 33 (600)47784

Stanford Electroydes L.:':,;ratories

STANFORD UNI'ERSII Y
STANFn)O, CAUORNMA

P---

T-~

T---~

~h

IY

F-pw

w'-WW-

Stanford Electronics Laboratories

Stanford

University

SLanford, California

1*

CHARACTERISTICS OF CRYSTAL-VIDEO RECEIVERS

EMPLOYING R-F PREAMPLIFICATION

by
W. E. Ayer

Technical Report No. 150-3

Septerrnbev 20, 1956

Pre~pared under

Air For-ce Contract AF 33(600)-27784

I WWWW

wVww-

I
SUMMARY
The operating characteristics of crystal-video receiving
systems with r-f preamplification are discussed, and equations
given for the maximum sensitivity (Smax). and for the gain
required in the preamplifier to realize Smax . Use of the
equations is demonstrated by means of specific examples.
An analysis of the behavior of this type of system is
presented in the Appendix, the derivation of the Smax equa~ion
being given there.

-iii-

TABLE OF CONTENTS
Page

1.0

Introduction

2.0

The simple crystal-video receiver-.,

3.0

Effects of preamplification. .

3.1

.
.

. . . ,

.
.

.. 4. .

a.

0.

a.

Microwave pulse receiver. 0 a

C-w detection system. . .
. .

Conclusions . . .

Appendix:

..

..

. *

.
.

.
.

.
.

.
.

2
.

3.2 Preamplifier gain required to realize Smax. #

Practical examples . . ae v
. . . . .a. . . &
.
.
4.1
4.2

5,0

0 0 a 6 a 0
*

Determination of maximum system sensitivity,

Smax

4.0

. *

'7
7

9
10

.*

..

Derivation of maximum sensitivity of crystalvideo receiver with r-f preaiplification . . .

12

LIST OF ILLUSTRATIONS
Figure

1.

Page

3
4

Block diagram of receiver

Threshold sensitivity (Sx) of crystal - video receivers as a function of band ;idth (By) . ...
. .
Variation of sensitivity with r-f gain. * .....
Variation of system sensitivity with preamplifier

gain*. oo o
. 9:
e......m
:*

5
A.1
A.2

C-w detection receiver. .

Model of noise spectrum . .
.
Idealized noise output spectrum

.
.

iv

o *

c
..

.
.

......

a
.

1
4
5

8
9
12
15

CHARACTERISTICS OF CRYSTAL-VILZO RECEIVERS

EMPLOYING R-F PREAMPLIFICATION
1.0

INTRODUCTION

The development, in recent years, of high-gain broadband

traveling-wave tubes has greatly increased the potential usefulness of receiving systems of the crystal-video type.

A .ery

simple receiver can be realized by preceding a simple crystalvideo .ystem with a traveling-wave tube operating as an amplifier in the radio-frequency range of interest. The general
arrangement of such receiving systems is illustrated in
Fig. 1.

R-f amplifier
Bandwidth = Br

Power gain = Gdetector

Square-

lawbnddh

Video amplifier

andwidth

Noise figure = F

FIG. l.--Block diagram of receiver.

Addition of the traveling-wave-tube preamplifier can result
in an appreciable increase in threshold sensitivity, making
such an arrangement sitable for many applications where
superheterodyne receivers would otherwise be necessary.
A brief discussion of the operation of crystal-video
systems without preamplification is presented in Sec. 2.

Data

on the threshold sensitivities nbtainable with practical microwave detectors is given in graphical form.
Seution 3 is devoted to nonsideration of the effect-,, of
various amounts of preamplification upon overall. system sensitivity.

The derivation of the necessary relations is indicated,

and the working equations presented.

-1 -I

5 w-

In Sec. 4 the design procedure is demonstrated by means

of a practical example. Also, a circuit arrangement for indicating the presence of a c-w signal is described, and performance capabilities are cleculated.
In the last section the characteristics of this general
class of receivers is reviewed briefly, performance being
compared with that of superheterodynes.
2.0

THE SIMPLE CRYSTAL-VIDEO RECEIVER

As the name implies, a crystal-video receivar consists of

a crystal detector followed by a suitable video amplifier.
The choi-ce of video bandwidth is made in the usual way, on the
basis of the type of signal to be rpcelved. Thus, if the
applied signals are rectangular pulses, the bandwidth is
typically made equa2 to 0.6/6 where 6 is the pulse length.
R-f selectivity can be provided by meons of tuned circuits
ahead of the detector. It should be noted that such selectivity will not improve the sensitivity of this type of receiver,
because the noise level is set by the crystal detector.
The sensitivity of crystal-video receivers is inherently
low because of the low rectification efficiency of the detecting element at small signal levels. The effects,threshold is determined by the magnitude of the video resistance of the detector, and the bandwidth of the detector-Video amplifier combination. The video resistance behaves
like an ordinary resistor in producing thermal noise.
A variety of silicon crystal types.--iN23B, lN26, IN32,
etc.--have been used at this laboratory in various detector
mounts. The video resistance for t-?se rectifiers generally
falls in the range of 5,000 to 20,000 ohms, which is sufficiently great to override noise produned in the video amplircer
if reasonable care is taken in amplifier design. A particular
practical precaution worth mentioning is the use of wirewound resistors in un-bypassed current-carrying circuits near

-2-

the front end of the amplifier.

If carbon or composition

resistors are used, considerable amounts of "excess" noise

are produced, and overall performance will be degraded.
*receivers

In Fig. '? data on the sensitivity of crystal-video

is presented. These curves are based upon the performance of wideband detector mounts developed at this labo"atory.
These mounts have been designed to work from a
50-ohm source.

The threshold sensitivity figure (S ) gives

the amount of pulse power which must be applied in order to

produce an output Just discernible in the noise on a synchronized CRT display. This measure of sensitivity corresponds

roughly to pulse power equal to average noise power. The

data are based upon input pulse lengths approximately equal to
0.6/B --conditions for maximum sensitivity. The "good" and
v
!poor" lines indicate the limits of performance for crystals
selected at random.
The sensitivity, as plotted in Fig. 2, varies inversely
with
. This characteristic results from the square-law
nature of the detector. The noise output power will vary
directly with the video bandwidth; the square-law action causes
the output signal amplitude to increase as the square of the
input signal.

Thus a doubling of the video bandwidth will

cause a 3 db increase In output noise, while a similar increase

in output signal will Le realized by a 1.5 db increase in
applied sigral.

D. J. Grace, "Analysis of two broadband untuned crystal

detector mounts," TR No. 14, (Nbonr 25132), rlectronics
Research Laboratory, 5t*.nford University, 10 June 1952.
2

D. J. Grace, 'Arialysis of a broadband untuned ur~stl

mount for X-band operation," TR No. 7, (Nonr 22510), Electronics Recearch Laboratory, Stanfo'-rd University, 10 March

1953.

-r-7

-77

77.

>

ICYi_
_

_,

c.

____i

0 H

o%

w.qp

Iqsatl.

_A0

Ulr

E'1

3.0

EFFECTS OF PREAMPLIFICATIoN

A qualitative picture of the effect of r-f amplification

on system sensitivity can be obtained by considering progressively greater amounts of gain ahead of the crystal deLector. For unity gain in the preamplifier the sensitivity
will be the same as for a simple crystal-video system, since
the noise at the output of the system will C:e due only to the
crystal detector. As che r-f gain is increased, there will
generally be a considerable region in which the system sensi.tivity increases linearly with r-f gain. For sufficiently
large gains, the output noise due to the preamplifier will
override that due to the detector. Further increase in gain
will no longer increase sensitivity, once the threshold is
cet by the r-f amplifier. The behavior described is sketched
in Fig. 3.
3.1

DETERMINATION OF MAXIMUM SYSTEM SENSITIVITY, S

An analysis of system performance has been made for the

maximum-sensitivity condition, corresponding to predominance

Maximum sensitivity

,
0

-Crystal-video

aenettivity

R-f amplifier gain, db---

FIG. 3.--Variation of sensitivity with r-f gain.

of r-f amplifier noise at the output. This analysis is presented in the Appendix.
The maximum sensitivity is given by:

Smax = -114 (Ibm + F + 10 iOgl o

[Be'

()

where -114 dbm is the available thermal noise power per

megacycle at the input of the preamplifier, F is the noise
figure of the preamplifier in db, and Be is the effective
bandwidth of the system in megacycles. The effective bandwidth is given by:

Be = [2BEv
r -B
vv

(2)

where Br and B v are the preamplifier and video bandwidths,

respectively, in magacyales. When the r-f bandwidth is at

least 10 times the video bandwidth, Be is given quite accurately by:

Be = [2B
3.2

]v3l

(3)

PREAMPLIFIER GAIN REQUIRED TO REALIZE Smax

The gain requirement for realization of

.uz- t. tem
sensitivity can be deduced from a plot of the type shown in
Fig. 3. With the sensitivity of the crystal-video portion of
the system known, the slop..ng line cor'esponding to small
amr- s of pream.plifier gain is established quantitatively.
The horizontal line, corresponding to maximum possible system
sensitivity, is found fron Eqs. (D% and (2) or (I) and (3).
The intersection of these two stra.ght lines gives the preamplifier gain required for the output naoise due to the preamplifier to equal that due to the Ctector. This conditlon
corresponds to an output signal/noise ratio which is 3 db
worse than tnat which will be obtained wrien there is sufficient
r-f gain to completely override crystal noise. Since the

- 6-

detector action is square-law, a 1.5 db increase in input

signal will bring the output signal/noise ratio up by 3 db,
so that, for this particular amount of preamplifier gain, the
system sensitivity will be 1.5 db below Smax*
A plot of the variation of sensitivity with r-f gain in
the region of G - Isax - Sxj is given in Fig. 4. When the
gain is 3 db greate;, than IS
- Sxl, the sensitivity is
only 0.5 db lesn than S
Accordingly, a reasonable practical rue for determinIng galn is:

Gt

Sax -

+
*

(4)

where all quantities are in terms of decibels.

4.0

PFRAOTICAL EXAMPLES

Calculation o! the performance chrp:teristics of

pveapli.fier--trytital-'ideo combinations is a relatively
straght-forwa-d prced, re. .n order to illustrate the calrcautions,

culations and, as %:ell, to point out certain pr'actical pretwc. examples will b,' considered.
4.1

MICROWAVE PULSE RECEIVER

Consider the following system characteristics:
S

= -52 dbm

= 0.6 Mc

B3r

-.

?0

M.

F = 20 db

These numbers are consistent with piuise receiving applications

in the kMc range, where a T-W tube followed by a tuned cavity
is utilized as the preamplifier. (Note that the cavity must
follow the 'T1 if'high sensitivity is to be realized--otherwise

7.7

A-6

> 4

C)
I
S-8

-6

-4

-2

+2

Gain of preamplifier relative to

+4
s..

+6

+8
S,

+0
db

FIG. 4.--Varlatlon of, system eenpitivity

wiLh preamplifier gain.

-8-

all of the TWT noise will be applied to the crystal.

In some

applications an additional cavity ahead of the tube might be

needed to prevent overloading).

(2 x 20 x .6

Be

Smax

LThen

From Eq. (3),

-114 dbm + 20 db

-114 + 20 + 7 ---87 db

-t

4.9 Mc

10 loglo(4.9)

The 'rnimum preamplifier gain is then, from Eq. (4):

01a
4.2

1-87 +521

+3

38 db

C-W DETECTION SYSTEM

An arrangement for determining the presence of a c-w

signal is illustrated In Fig. 5. The purpose of the modulation
of the r-f input is to produce a video component in the
envelope of the signal which, when detected, will pass through

F.

Lossy
Modulator

Tunedne
Preamplifier

Detector

(Ikfv:

__

FIG

5.--C-w detection receive.

the tuned video amplifier.

The actual modulator could be a

ferrite devicc, or perhaps a transmA..Ion-line structure with

seml-conductor diodes whose charactoristics are changed by
the modulating voltage at the frequency fv
For this sort of application, where it is desired to close
a relay or perform some other simple function when a continuous

signal is yr.

ved, a very narrow video bandwidth may be

employed.
Assuming, for sake of simplicity, that the modulatcr
goes from a "full-on" (zero loss) to a "full-off" (infinite
loss) condition in a square-wave fashion at a fundamental
frequency fv' the maximum system sensitivity may be calculated
in much the same manner as was used in the preceding example.
On the basis of a 1 Ice video bandwidth, the crystal-video
sensitivity for a "typical detector" is found from Fig. 2 to
be -63 dbm.

Assuming an r-f bandwidth Br of 20 Mc and a pre-

amplifier noise figure of 20 db, Eqs. (3) and (1), respectively,

yield an equivalent bandwidth of 0.45 Mo, and an Smax of
-97 dbm.

From Eq. (4), the required preamplifier gain is

found to be 37 db.
There are two points regarding the performince of such a
device which should be discussed.
a modulator is needed at all:

First, one might wonder why

;hat is, why not use a d-c

amplifier to provide "direct detection",*f t*Ie c-w signal?

Tho answer lies in the fact that there fs.a.arge component of
1"noise" at d-c at the output of the detector which would make
the system sensitivity extremely low.
Appendix.
A second question is:

This is shown in the

why not introduce the modulation

by appropriate variation of the voltage on a control element

in the preamplifier?

The trouble with this technique is that

the entire noise output of the preamplifier becomes in effect

a "carrier" for the modulation, and the sensitivity is again
drastically reduced.
5.0

CONCLUSIONS
If it can be assumed that sufficient r-f selectivity is

obtainable to limit Br to the actual value required for the

intended application, then the ser sitivity of the type of
receiver considered here can be made essentially equal to that

10

realizable by superheterodyno techniques.

For, such cases the

simplicity of the preamplifier crystal-video combination may

well indicate its use in preference to a superheterodyne.
Again, in many applications the simpler system wiil not
provide the ultimate in sensitivity, but its performance may
be entirely adequate for the specific problem at hand.

APPEINDIX:

DERIVATION OF MAXIMUM SENSITIVITY 0? CRYSTAL-VIDEO

RECEIVER WITH R-F PREAMPLIRICATION

In order to simplify the calculation of system performance, an idealized noise spectrum will be assumed at the input
to the detector.

The noise, '-hich in reality is continuously

distributed in frequency, will be considered as a discrete

line spectrum.

The components will be arbitrarily assumed

to be spaced uniformly at 1 cycle intervals across the r-f

bandwidth Br, each amplitude corresponding to the thermal
noise in 1 cycle of bandwidth. The phase of each such "equivalent noise signals" will be considered to be random with
respect to all the other components.

This equivalent noise

spectrum will appear as shown in Fig. A.l.

43

wI

___

llHLIII

..

FIG. A.l.--Mode). of noise spectrum.

The detector will be considered to operate in its square-

law region, as given by!

eo

(A-1)

-3i.4 -

HP

where ei is the instantaneousH voltage at~Lthe input to the

detector, and eo is the corresponding output voltage. K.

accounts for the overall effcien - of detector mount and
,

crystal.

1'I

Calculation of the noise at the output of the detector is

accomplished by substituting the expression for the model of the
Input noise spectrum into the relation for the detector
characteristic:
e. =Kx[Vn sin(wlt + 01 ) + V n sin(w2t +
L.

+.....

(A-2)

+n ain cunt + On]

where V
cycle.

is the peak value of the applied noise voltage per

If the r-f bandwidth B is expressed In cycles/sec,

there will be a total number of components in the noise spectrum

model equal to IBl

For convenience in manipulation, the expressions for
Ssin(coit + 01), sin(a 2 t + 2), etc., will be represented by
a,, a2 , etc. Taking the common amplitude term outside of the
brackets and substituting rms for peak voltage (vn for Vn) ,
the expression is:
eo = 2KxVn 2 [aI
[

-a2 + ... SBr] 2

(A-3)

Expanding, the output voltage is:

2KxVn 2 {l2 + a22 + a 3

+ 2a1 3

+.

--

+ 2a 2 a33

I.....+ 2ala 4 + 2a2 a5

'24
2
'A-4)

Remembering that the vario-us noefficients a1 , a,, et.,

represent sine waves, the squarcd Germs reduce to d-c and
second harmonic components, while the product terms reduce
Lo sum- and difference-frequency components.

Restricting the

analysis to cases where the video bandwidth B v is eqiual to or

Sltss than the r-f bandwdth Br, only the d-c and differencefrequency components need be considered.
-13 -

There will be IBr

d-c components, each of amplitude 1/2.

The difference frequency terms at 1 cycle/sec result from the

products of adjacent components 2ala,, 2a2 a3 , etc. There will
- 1 such components, each of unit amplitude.
Similarly, at m cycles per second there will be IBr - Iml

be a total of IBr

components, each of unit amplitude.

Because of the random phase of the components oC the original r-f spectrum model, the difference-frequency components
occurring at each cycle in the video band must be added powerwise.

The expression for the video output is thus:

[0a) +

e O = 2Kxv
xn
x

+ CIBr -211

[IBrV

-111

+Ir

cos(Pt +

-2
) +

cos(11vt+b

o~at%
CIBr - 3]1 cos(6't+

......
(A-5)

The output noise spectrum, in terms of the squares of the

rms values of each component, is sketchcd in Fig. A.2.

To

find the total noise voltage output within a video bandwidth

Bv, the spectrum is integrated over B v . This miist be done
power-wise, and then the square root taken. Assuming that rhe
video amplifier response approaches, but does not include the
d-c term, the expression is:

By
V

2B

Kxvn2[}

dJ

(A-6)

0+r
Integrating, the output noise voltage is found to be:

Vno = KxVn 2 [2BrFv - Bv

(A-7)

The threshold sensitivity of the system corresponds to

the condition oi the signal voltage at the output of the

- 14 -

1A

KC2vn 4Br 1J2

2Kv

04

Video frequency

-k-

Br
FIG. A.2.--Idealized noise output spectrum.
detector (Vso) being equal to the above determined noise
voltage. (It is assumed here that the noise due to the crystal
itself is negligible: This is necessary for the realization
of maximum system sensitivity).
The signal output voltage of the detector will be:
es= Kx[V s sin wst] 2

(A-8)

where V. is the peak value of the signal voltage applied to

the detector. The peak output voltage for, a c-w input, or
f . a pulse which is sufficientlv wide to pass through thc
video ampifier, will be given by tne d-c term obtained when
the squaring is performed. Tha:
KxV s

=
Vs

2-

K-xV-s2

where

VS is the rms value of the input Vignal.

and noise output voltages,

-15

Equating sigral

(A-9)

V
where v

(A-l0)

is the thermal noise voltage per cynIe of r-f band-

width, as previously defined.

This may be rewritten in terms

of available powers at the input to the detector:

Ps

Fn [2BrBv - By2

(A-1l)

Putting noise puwer and bandwidths on a megacycle basis, and

converting to decibel measure:
PSdbm = -114 dbm + 10 log 10 [2BrB v - Bv2]1

(A-12)

The above relation applies at the detector input. This

is readily translated to the input of the preamplifier by
adding in the effect of the excess noisiness of the preamplifier. The resulting relation for the maximum sensitivity of
the system is:
Smaxdbi

= -114 + F + 10 log 1 0 [Be]

(A-13)

Where F is the pre;^kmplifier noise figure in db ana B_ is the

effective system bandwidth:
Be = [2BrB v - By 2

(A-14)

When Br is at least 10 times Bv, Be is given with good

accuracy by:

Be

[2rBv]l/

16

'

(A-15)

and Afral00

DA.J4.lO..lll

t
te.
,MNt..,
Stg..I Corp. Cal... It.. lake
"..., b" .l l.,al.
0241. h. 41,
Set_.,I
.. harry
20A--3
t
.,.1w1 1111,,

~l

Gil.

A Z5...
NLt
... Vr,
... 11,..

D..lltj..IU

l.0 4l.P.
Oi...
GAld

.... Dk.. D0..In"t~, Di.,

C1.9,1.4lA
At
Off
I M1..
chief.8
t82410
A,-.
Ch~ekf.t1.D.

All.Atpt.
2

All.

f C
it

54*

tt
I.~.5. A5llt

hW
l
a,.

Iil
Wee

De ta1

Ill

Utn.l.

,..Ccl~sI
Of-,A.'.,

1-1

8 .

Vi.,lto..

C,..lah.t,
A= .dl.

,.-.

All I~.i.I
art.... 71..

Al.00

Talkh

.
-Ithn

erb
H.

A".

It~lI

01

Itil,

to

U.Iut. 1.t..Lta
A2t%-~t
G
L . l..-.

t.4,.
At00

tllyI
PatIdaalo

Ph14f2a.
0

7t-lt14
1~~
0
L,tt..
WrgtAtDe-a Cotter

florins 111i.l
At
An..

11
0tlWa1

A0~

3-4.

A41-1.1

b11.
..
.

La
2

C.,.

Al.
_-

Off~
.1

Co.,.
AU.6-ag
Uol.

I
O1322, .INI,.
Goland,
4..,.w. Aintroat
1004 t.l..
AUt.
0,

III
A

U.

00..0Cattle..
Aln. acwsdt M1

lots Va...

...

OffCt'. 0 Natal R......

it

--At'.pt0
00,10.01080 D"..
All.
IALOl

All.
AR
j1ll..

0.H.

40..I

Al 51001I

C~d.UIO

Stoy4 hrvt1.5.C.ttOI
All.

....

InitiatioS.1.
.1,1.1P.D.
9-7 tA..r..brtMre.
Am

W.72

1loo, mos

7.h and.

I.C

0402allte

h...
.C.

1Ui

0,,,.,..,

.1

,1

1 ,t5.

.tto.

Lateral.t

N. 2, C
84
U.I
id'ibi~s
.%

al:ZI

L4011t11,

CId
b..~

l,4
York IIt. M.t...~~.
TookA.0..,.
00t10
battles1U1
, AJrI.
54...mfr-At

CAV.7.tl~l.0
C.U ..
I"~l All
0..2n.. UCLOI.
5.1110...t,
I1tot
ll
IN All.'slDNowo
S.Jot A ,l0l
.t
C..
Cere
0
l1
. A'01-0n.t.
arl0t.it
M . OhCAA.
Alitl. LI. nor
Vll0I
INCL00.7
2ta Aim. Cap. 0. A. st.l.
I

U.S CtIvl~roS.
tlhdllisA.

At.. Me
~ ~.t ~1LI.

PU

11005SationLibrar

r..I

000.C0SOIA
.aot-IN.

Al.

Stodar

R~g..adl.0n
Mai F0awa,..01

Va 4fld
DOW
Er.

T.1,

WhLwa

1tSS~ O

51

744

h.

IoS
7=Alln~

At.111
D1.1
Ol31

10.2.ClIt

AU.,
..
llisr00 00,1.1ls,, A Olat

.tll
O

Vt.1

SW rtatinal.Ail
1
c....04"In.sfl
Alt 0n.4lo.t-lCandor

o11o

Al. Dl.~:.
AO..I.
T.Clannial
4*11104a

0660..1

061,h10,11111.
Allsrv" A

.,

S0

Al'..

I~~~~~~~
.A

Mel
15.1.1

I'0.

D.C.4

.1An.l

w7
I LIO.,tl

CO..al.,.,

1111.,1lni .1Il.laM ro .I

of100.
.4
D$ Patro
0 tot
.......h IA*7

LIM.. "r.-C
4111. it
Tenn*.
01

s.

Dallt.
ttll

1e0C,

AN.&.

000.1,21010080.
0a-.6

Aill.-sok
C

101
:
21rn.
DC.

IM"
C

Ovtal.mt

coloadne

A-535 .

b~shiuI.,..l
461A-Mln
.1C
r"" hand
014,
:
,1.
WAan arl tt
6

...

Usve..t.lnn
t .1501

D.I..04
A..e

n C.

eidIe4
It0IS. lt
am.S
ll LIA1
Aw
e
Q 1,2t

e.

al..CO

iaClO

ad'.W

.341

mC.

We"
.,,
Il

slim..43
0.p.h4g'p
A.......

Is,..,

NO.1

,-All

Iioooa4.511.10.

.4114l. Me.4.l.
A tll1 n l..104d..ilw"1
Li.brarian

6.t.'.. "M

944l

0.t...,I.
.,IO
1
50Million11
CWe..

Af00.3I42

Lirary

.On

.P...Z

S.C.

t. )1'$. 00t1
ricll14.L

U.

A111.1SI

laoi1121111.
a .. 1..
DeI.s.
..3 0Il LoI..ilnf

54a.111.1I1,.

1 4 ,013a
. ab ..

~.UI

Spb5

C..1Dala

Vlaarlolo

a.

ib..w.I

Wl ..
3; ..t
RV0 Patt,44 111010
s. CI,.
).t., J.

Alt.

t..ia,4

M.

Aft-

.WA.M10lM4
at

W !a-idr

Mltt
OTattlk

71;Mnneo

b.,.ti.1oir.
t. . . na...IIS

0
r ... Lat

7.82ae. t.

CL
oial

"Lloiaw

nlif..it r0t519"nIta"~oid o.
4

Alt. 4.01.V;.IO~

51taI.ranL1..to
Gish land

Gee,

wt

510044lo
Kto,~~lq
D..
11.3 .00 .v R.1.

ICAlorp.~l.
Dtllltnro

- C-10 Age.,,

D aW
pl11.

.0
Aim5A .3.
L..Se
Llbo.

Chri,.1
A.~il
.. t1..
.
Ca0 ll j 9.6h."l,,tI
It:: Tak
.111
18211.4rl..he~h
York.,Ie
Mo. Yorkl
111.11
I
All.-

cente

~~1

CtI.: 'M
3.,,.

N...
03.117.

0111.-h
J.: ... t,

h4.. &-.I2.
t2001
..... -11.4

Ig

...
.C. n

Caoios

Alt.

seat. Courselll712*l
, 1 OttanllO..
AtheAIt,
00.,

W5.-

Meta..~,tt~.0L..
1. 000,t
.11.
..
P-1 - ,,
A
l~~r

I.J

113.teho.
r
Co.-"..~also rn
A".b..lpo

411.

It

At~.t M tt.
Johor.ltD-I
l7
lOJo.0100
0.A... f5tLwiat
C-,.
I
.
t a~l.I
AU.:
7 01= .
1 .0

Wl.p.I

fl
W.tl.,~

43.11
Ck-l.

.l

I4.l..2t
sld1It.146.100,1111

allana
a~ti.

05..
MC.. l.~

IGO.t

S~.Lee'l. Cll ....

012,

Liant A-p

aci.

Wuh...#

t.

LbO

5.0.4.
.

14t.Rm
ar

i.y

Dept41

O,.."t CI....41
i
Mr01A.
2.11-101140 ol

ost.

aw.,,

1klI

111W.p..tC
itaE M w

r!U...h R",
011. Me-~T
o'llII*
Al.t M. .. 44 0,114
A.

t
..

14-l..l

1.l 104-.i.l

N,..8tpC.''."
00.10t 00111,..t

II

we#.
All
0 13.

Mriaae..le

9A0olI t 1110.1.
.,A M1..,.

Oh.. .1

ttttt.
V.

~iet

804:
N.

C-

C- L..
i
.llat-.I0

vlw

4 1 Atree

it~U..
..

00O.

0,.1.P*,ft.11'

'.."...
M.

Wa
All.TR

Al

.'

I W10lt-t5.4.p..ltt.I
CO~~~~~~~tl.~~~~40
-1nti

rlKll.*

tl.
PolI

.W
A
rare. Stt.., Florida

A-

0 tnotte

a,.....

Wou,
Monahan
Q11f ..I.
Vt...

1 al Ote.2..t.
C.arnall..4
109S.0111.141. S1~4.1,a
All. CIL 0.3. 010111.. 00721.7f

At..WG
A'-1411

U.1

M.
I.
4.TtIt

At.
lt

I..p.

, 1

General U-...
Cad.,
hir....
L.W.,aseuy
689 Ga..
I.l~
A...Pat00Allt.
Wi.Allf.lt
Ann.: T-lhatal IO..I

nl~l

MyAl
t ....

C-.1

AC~
....
Atvt-A,

Anne~Siut

3;.-% L..l.

It..a
ltt

A1IO
A,
b.ar-

C.10

A-tfllm.&lB..Al.

410~t,44
Olt
.t4 11.11.1
UI
-I L....nt.
L~b.I.I
. - a-,
)l
%2 r8.....a
Alt.11e17,

1114

24-

Allw
..
' teal1,c"

0.

2.17

flatarue

II....
D.C.l..

11.24.
AIM-1
11,
4pltI'40
Oft:. sm. Ia -.
it-01 114."1l0.1

ll00O#,

.
Ont1..

111adi.

15.."darian

WIAAl

1tto~a
ltr.1

I...14I

C2

149F

n~lph

~~..

j?.

I-0-1

.de..

e..A,..4.
AMt. A~l

.- t.,.

D M

Al...
C .rote#4 A.... OI
Qv. A. .G . .&4t.rt.T.WoiAols

* All.

G 1.1.wa

~l

ool.

Io. C111.Mr. .
4

At
.ht....j

.kJ.1.M....
*~ b. . Y.

AS.road
pu.
K...3.
A....y

a-

JA a

A-A

, -ot,

.W.

TonI

..
...

.M.3.

...

,W
l g.

A'r.

sa
37.
w.

wlom
shf.C

C-6166,
O
AltoV.... C.91. LDg.

4.N

I).....

....
,

3.0do-t14.M......

'.15 ..dr...

Co.-

#931l~occkIW

A Am LC's..

L..a., KCM?.

.,

All.. T.M A Tort.

All.

D",

Cr.t....
C l

rr&.A

C
oI

M.

s am

13.

LC4.
IM

Aly. U,. am.

C -.

CC--mod
Aft For...C

C....

.4..

A.

J Also

C-39

.4

41dU.
P 1
.*ro lm.
Ib."t. 3.3.C
-.
sc,.
ta.
Art

I CA

Alt.
to
OR*-A.
hhl.
AAM'
.U~.U.....
p
Al.F...
.Itm
I....

CAL-Mv..Iit- 034

Al-r ,O...I'..,m

?--$et

oho

...

LC.: