Space electronics
Settling on the moon
Four-beam radar system that guided Surveyor 3 to the moon will also
furnish the velocity and range data to land the Apollo lunar module
By Charles J. Badewitz
Rya n Aeronautical Co., San Diego, Calif.
A la ndin g rada r that proved its capability by guid-
ing Surveyor 3 to its landing on th e moon last
month also will provide the data needed to land
th e Rrst mann ed Apollo lun ar modul e (Ll\l) on the
surface of the moon.
As th e two-man L)'[ descends toward the moon's
surface at abou t 60 mil es p er hour, its landing radar
will be switched on at an altitude of 50,000 feet.
About three minutes later and 10,000 feet nearer,
tll C radar will b eg in to make continuous range
m eas urcm ents , and at 25,000 fe et it will b egin co m-
puting th e velocity of th e module. The range and
velocity dat::! will be pro cesse d along with auxiliary
data from other sensors and th e res ults fed to th e
landin g control sys te m to in sure a so ft and safe
landing. On ce th e spacccraft land s, th e radar's
work is cOll1ple ted and it rema in s on the surface
of th e moon as part of the lun ar module's descent
s tage.
Th e land in g radar u ses a three-beam doppler
velocity senso r and a sin gle-b ea m alitmetcr to
measure th e spacecraft's velocity and range rela-
tive to th e lun ar surface.
The rad ar return signal in each of th e veloci ty
sensor's rece iver beam s is mi xed with a portion of
th e transmitted signal. Th e output of th e mixers
are audio doppler frequencies that are proporti onal
to th e spacecraft's velO City components along th e
respective beams. Frequency data converters break
The author
Charles J . Badewitz is director of
engineering at the Ryan
Aeronautical Co.'s Space Systems
Division. He was chief project
engineer on the lunar module
landing radar program and
desig ned a number of specia l-
purpose airborne doppler
navigation compu t ers.
110
down th e doppl er frequcncies into velocity co m-
ponents in vertical , latel:al , and longitudinal an-
tenna directions and relate them to the spacecraft's
coordinate system .
The· radar altim e ter's return signal is also mixcd
with a portion of th e altimeter's tran smit ted pulse.
This produces a frequ ency proportional to the
spacecraft's height above th e moon plus a fr e-
quency corresponding to the vchicle vc loc:tv al ong
th e altimeter bcam. Th e latter is rcmoved in th e
range data convertcrs by utili zing the doppler sig-
nals in velocity beams 1 and 2.
After process ing by th e radar, th e vclocity and
range data is sent to th e lunar modul e's guidance
computer in a serial binary form , and as puls e
train s and doc analog vol tagcs to th e astron au ts'
visual displays.
The landing radar's data will be combin ed with
altitude and velocity signals from th e space vehi-
cle's inertial mcasur in g unit. Th ese d ~lta inputs are
weighted by the spacecraft's h eight above the
moon 's surfa ce, and program ed to assure the guid-
ance computer receives the mos t accurate input.
Errors in th e in ertial system tend to in crease with
tim e. Landin g radar errors, on th e other hand , are
de termin ed primarily by veloci ty and altitud e,
th erefore th ey decrease as th e velocity and altitude
decreases.
Th e initial effect of th e radar data on fh e space-
craft's guidance system is small , and increases as
th e descent progresses. However, as th e data from
th e landin g radar is factored into th e gu idan cc COI11-
puter, it reduces th e effects of accumulated errors
in th e in crtial unit; thus the landing radar updates
th e inertial measuring unit.
Integrated Antenna
The four radar b eams are generated by two
planar arrays on th e single integrated antenna as-
scmbly pictured on page 113. Th e groups of alter-
nate slotted waveguides in th e large center section
Electronics I May 15, 1967 IiO.'!!!
 DOPPLER VELOCITY SENSORS

INITIAL
BRAKING
MANEUVER
5,500 FEET
PER SECOND

Powered descent begins when lunar modu le 's altitude is about 50,000 feet. Throughout the tilt maneuver, the four
radar beams are kept pointing at the lunar surface automatically by the guidance computer. The repositioning
is accomplished by rotating the antenna of the la nding radar around the Y axis into either of two positions.

are the transmitting antennas; the receiver anten-
nas are th e outrigger sections around it. Th e
E-p lane slottcd guides in th e center section tran s-
mit th e altim eter frequcncy-modu lated continu ous
wave; and th e alternate H-plan e slots transmit a
c-w velocity sensing signal.
The thrce narrow b eams from th e velocity scnsor
an tenna emanate from th e planar arlay in th e same
mann er as legs are ex tendcd to form a tripod .
Th e attitudes of th e b ca ms during the initial
braking and hover man cuvers are illustrated by th e
drawing shown above. The multiple b eams are
generated at fix ed angles relative to a line drawn
perpendicu lar to th c face of th e planar array. Th e
outrigger rccc ivin g arrays, which have th c same
slot or ientation as tll(' " clocity tran smitter array ,
are th e three broadside arrays . Thcir gco mctri cs
make each receivin g b eam axis colin ear with th e
corresponding transmitted b eam axis.
Th c altim et er's reccivin g array is boresighted so
that co in cidcnce of its main lobe with th e altim eter
tran smitter's main lobe is achi eved, yieldin g in turn
th e desired two-way pattern for the altimeter.
Perpendicular beams
Th e antenna is mounted on a rotating pedes tal
on th e underside of th e spacecraft's descent stagc.
Th e antenna can be rotated about the spacecraft's
Y axis , into ei th er of two positions, automatically
by th e guidance computcr, or manually by th e
pilot as th e eraft's attitude changes. This keeps th e
rad ar beams as nearly p erpcndicu lar to the lu nar
surface as possible.
Th e diagram on page 114 details the basie op-

iii.'! Electronics I May 15, 1967 111

 I
VELOC ITY
II
I
, - - RECEIVER -..
ARRAYS

BEAM I I

Ii !
I
____._______ ~
BEAM 4 "

L- ___________ ~L
Lunar module's landing radar combines transmitter velocity and altimeter signals with the signals in
return beams to produce doppler frequencies proportional to the space.craft's velocity and range
from the moon . Data converters break down the doppler signals into velocity components in the X, Y,
and Z direction. These are input signals for th e module' s guidance computer.

era tion of th e radar's velocity sensors. Th e b eam quencies proportional to th e component of th e

angles and th e energy pa ttern shown are for b eam spacecraft's velocity along th e b eam of concern .
1, but the pa ttern s and angles for the other two Th ese frequ en cies, D D ~ , and Da, are comb ined to
velocity beam s are similar. produce th e orthogonal " velocity components su ch
Two orthogonal sets of velocity components are as Vxa , Vy ., and Vza .
measured in th e coordin a te sys tem of the antenna. F or eX'a mpl e, th e component of velocity relative
The Vxn , V)'a, and Vza components are shown to the antenna axis, V x ., is computed as follows:
on page 114, where V xa coin cides with the center
line of three velocity sensor b eam s and V)'a lies in Vxa =
A CDI + Da) (1 )
th e plane of velocity sensor b eam s 1 and 2. The 2 cos A cos ~ 2
coordin ate sys tem for Vxa ' , V.r . ' and Vza' velocity
co mponents is rota ted about the V)'a axis through W here:
an angle, t, to cause Vxn ' to lie in th e plane of A = electromagnetic radi ation wa velength
beams 1 and 2. The primed velocity components A = beam 1 an tenna angle
are used only for the pilot's visual d isplays . ~ = beam 3 antenna angle
T o determine veloci ty, the radar return signal is DJ = doppler shift along beam 1
mixed with a portion of the tran smitted signal. D3 = doppler shift along beam 3
T hi s is a process of d irect radio-frequency to audio- The fo llowing equ ation computes the range to
frequ ency conversion which produ ces doppler fre- the lunar surface. The K(Dl +
D 2 ) term is the cor-

fd
( QUADRATUR E
PAIR I
+ (QUADRATU RE PA IR I ERROR FREQUEN CY

VELO CITY
SENSOR rl SINGLE -
SIDEBAND
MODULATOR
f-+
I· F
AM PLIF IER --.. MIX ER 1 LOW- PASS
FILTER
--.. INTEGRATO R --+
VOLT AGE -
CONTROLL ED
OSC ILLATOR
....
f
Ie
R-F
(QUADRATUR E
OSC ILLATOR
PAIR )

LM 's frequency tracker is a closed loop servosystem that processes sensor and reference signals.

112 Electronics I May 15, 1967 i."!

 rection factor \\"hi ch compensates for doppler shift
in th e rangc beam.

(2)

where
R = range to the lunar surface
c = speed of light
S = a ltimeter 's t ransmitter frequency deviation
r ate
fa = a lt imeter 's received frequency, including
range and doppler components
K = scaling constant
D2 = doppler shift along beam 2
The radar antenna assembly divides each dop-
pler signal in quadrature pairs. The b eam's quad-
rature outputs are amplified and u"sed as input sig-
nals to th e frequency trackers in the electronic as-
sembly. Th en recombination of the quadrature
pairs and sin gle sideb and comparison techniques in Landing radar antenna is covered with special thermal
finish to protect it against extreme temperatures in space .
the frequ ency trackers validate th e signals and The center section contains the altimeter and velocity
measure magnitude and sense of velocity. sensor transmitting antennas. The outrigger antennas
Th e radar's receiver u ses narrow ~band filter cir- at either side of the center section are for the receivers.
cu its (l ab eled frequency trackers in the b lock dia-
gram) to detect doppler signals, and to separate
th e true doppler signal from noise b ackground. Radar system characteristics
Th e opera tion of the frequency t rackers is traced Type of system
b y the block diagram b elow, left. Th e tracker is velocity sensor c-w, 3 -beam
b asically a closed loop servosys tem whose inputs rada r altimeter f-m / c-w
are the doppler qu adrature pair, fd , from the radar
Altitude capability
velocity sen sor, and another pair called the carrier velocity 5 to 25,000 feet
qu adra ture p air, fe, from a fixed r-f oscill ator. altimeter 10 to 40,000 feet
Th e input circui t of the frequ ency tracker is a
sin gle sideband modula tor which accepts the dop - Velocity capability Vxn - 2,000 to + 500 / fps.
pler quadrature p air input and another feedback Vyn - 500 to + 500 / fps.
qu adra ture pa ir, approxim ately equal to fe + f,I, Vzn - 500 to + 3 ,000 / fps .
from a voltage-controlled oscillator. The single
Altimeter antenna
sideb and modula tor output is amplified b y an in-
type planar array
termedia te-frequency amplifier and mixed with the gain 50.4 db
fe quadrature pair in a second mixer. beamwidth 3.9 0 E plane
The output of that mixer is the error frequency of 7 .5 0 H plane
the trackin g loop. This signal is filtered, integrated,
Velocity sensor antenna
and u sed to drive the voltage-controlled oscillator.
type planar array
The oscilla tor' s output is" applied to the frequency gain 49 .2 db
converters and also fed back and u sed as a carrier beamwidth 3.r E plane
frequency for the single-sideband modulator. Un - 7 .3 0 H plane
der tracking conditions, the output of the single
Transmitters
sideb and modulator is very nearly the same as th e
type solid state
carrier frequency fc • frequency
Once the true doppler signal is acquired, it is velocity sensor 10.51 Ghz
tracked with a high order of accuracy, about 0.1 %, radar altimeter 9 .58 Ghz
and the tracker output frequency is almost the same
as the center frequency in the received doppler Output power
veloc ity sensor 200 mw min . (50 mw per beam)
signal b and.
altimeter 175 mw min.
E ach b eam's frequency tracker sends a doppler
signal to the velocity and ran ge data converters Altimeter
along wi th a d-c step voltage th at indicates acqui- modulation sawtooth f -m
si tion and authentication of a true lun ar return frequency 130 hz
signal. The conver ters then translate the doppler deviations 208 Mhz and 42 Mhz
sign als in to orthogon al velocity sets in the coordi- Response time
na te sys tem of the an tenna. The guidance compu ter 25 ,000 feet 0 . 1 sec
uses this data to steer the lunar mo dule. E quation

A.'!!I Electronics I May 15, 1967 113

, . - - - - - - - - - - - - -- - - - - - -- ---- - -- - - - - _.
Why did Surveyor 3 bounce?
When Surveyor 3 hopped and
skipped on the moon last month,
many people were concerned that
the same thing could happen to
the Apollo lun ar module when it
reaches the moon's surface. If it
does happen, it probably won't
be the fault of the landing radar.
Apollo will use an updated version
of the radar system aboard Sur-
veyor 3. The landing rad ar will only
indicate the range to the lunar sur-
face during the last seconds of
Right ; the actual landing will be
made by the astronauts.
What really happened to Sur-
veyor 3? The scientists at the Jet
Propulsion Laboratory at Pasa-
dena, who control the Surveyor
program haven't made known the
results of their investigation into
the cause-but one can speculate.
During a nOlmal Surveyor de-
scen t, the spacecraft slows to about
five feet per second at 100 miles
altitude. At this point, the craft
__._ - -- -- - - - - - - - - - ,
becomes practically weightless be-
cause the moon's gravitational pull
is 1/6 the force of the earth 's, and
the Surveyor's retrorockets are
thrusting in the opposite direction
to the pull. All the while, the space-
ship's guidance computer is receiv-
ing velocity and range signals from
the landing radar. As the craft
passes the 14 foot mark, the radar
generates a range mark which the
Surveyor's control system uses to
cut off the engine. Surveyor then
falls free the remaining distance
to the moon.
A landing with the engines still
burning is actually a softer landing
than normal - because of the
weightlessness. But the spacecraft
would come down , touch, boun ce,
and touch again, and continue to
do so until the ship's momentum
is dissipated in the bounces.
Ryan Aeronautical Co. engineers
tend to disclaim early theories
that the radar caused the engine
r
cutoff failure by sending confusing
signals to the guidance computer.
If one of the three velocity sensing
beams were pointed down to a
hole, it may have generated some
contradictory data concerning the
Surveyor's velocity across the lunar
surface. However, the engine cutoff
is made by a signal generated by
a fourth beam-the altimeter-
and if this beam was pointing
down a hole the spaceship would
have thought it had farther to go.
On ce before at the test station
at Holloman Air Force Base, New
Mexico, the Surveyor's engines
didn't cut off and it bounced. TO
damage was done to the space-
craft or any of the instruments on
board . That time the scientists
said it was the fault of the space-
ship's instrumentation , not the
radar. In the case of Surveyor 3,
Ryan engineers think the same
thing may have happened.
W.J .E.

1 shows how the sum of the doppler signals from
b eam 1 and b eam 3 are multiplied by a constant
coefficien t (which is a function of the beam angles)
to yield Vxn . V.rn and Vza and the range are obtained
in a similar manner.
The measurement of range is accomplished in an
analogous manner, th e only difference b eing that
some com p en sa tion for th e doppler shift along the
range b eam is required. This is provided in the
velocity and ran ge data converters by summing
th e doppler frequ encies in b eams 1 and 2, scaling
th e summation , and subtracting th e result from
V,.
_BEAM 3
BEAM I
/
----.",,;'
the doppler frequency in the altimeter return signal.
The measured range and the velocities are h eld
ready in the radar's signal data converters for read -
out when requ ested by th e guidance computer.
The computer samples the velocity and range
data sequentially in a cycle that repeats every 80
milliseconds.
Spurious signal rejection
The basic task of the frequ ency trackers is to de-
tect relative motion b e tween the vehicle and th e
lun ar surface. However certain physical objects on
th e lunar module, such as hold-down b elts that
come loose wh en th e LM and th e command and
service module separate, may b e in the radar b eam
and moving-thus causing spurious signals. To
avoid acquisition and tracking of th ese signals,
unique circuitry h as b een incorporated into fre-
quency trackers to examin e all signal s during th e
search mode and reject all except the true doppler
return signal.
Th e logic used in re jecting spurious signals is
based upon recognizin g that returns from th e vi-
brating physical elements contain double side-
b ands. Tru e doppler signals h ave ju s t one s ideband.
Th e technique that is used to discriminate against
spurious signals is simply one of examining the
signal level in th e positions of both sidebands.
The tracking filter examin es th e level in one s ide-
band; and a noise-sampling filter examin es th e level

---- --./ /
.;'
/
in the other. A comparison is made and if the signal-
to-nois c ratio exceeds a prescribed value, th e acqui-
Radar velocity sensor's beam 1 geometry is typical of sition circuitry allows the frequency tracker to
the landing radar' s four beam patterns. track the signal.

114 Electronics I May 15, 1967 &i.."!

 Space electronics
Digital tester okays Apollo's altimeter
Specially designed test set counts frequ ency and averages it
over the pulse period to measure slope and linearity accurately
By V.M. Andreone and J.H. Poi rie r
Electron ics and Space Systems Division, Ryan Aeronau t ica l Co., San Diego, Calif.
Before it leaves the earth, the radar that will guide
the Apollo lunar module to a safe touchdown on the
moon's surface will b e ch ecked out with equipment
that has no precedent or peer.
Like most of th e electronic hardware aboard the
lunar module, th e radar altimeter has some ex tra-
ordinary tes t requirements. In almost every sub-
assemb ly, th e Ryan Aeronautical Co. engineers en-
Glossary of terms
Downsweep average slope, m:
The f req uency deviation over t he downsweep
pe ri od di vided by th e downsweep pe riod, T.
m = .1 (T df (t) dt = ( fJ! _ fL)
TJo dt
where fJ[ and fL are the high and low limits of the
mod ulated transmitted frequency.
linearity, I:
The percentage deviation from the downsweep
average slope, m, of the measured average slope
for any 10% portion of the total downsweep period.
I = [ (~i)* - m ] 1~0
*a verage value over 10% segment of T
Downsweep/Upsweep ratio, a:
The ratio of the downsweep time period to the
upsweep t ime period.
Sweep cycle, ts:
The t ime required to complete one downsweep and
one upsweep cycle .
i.."! Electronics I May 15,1967
countered test problems that cou ld no t be solved
with existing equipment.
For example, the frequ ency modu lator, the heart
of th e radar altimeter that makes continuous m eas-
urements of th e spacecraft's distan ce from th e lunar
surface, required gear with a measurement accur-
acy 10 tim es b e tter than the tighte st modulator
tol erance. And th e equipm ent mu s t determin e in-
accuracy in an altimeter that will operate at 25
times the h eight (50,000 fee t) and twice th e fre-
qu ency (X band, 5.2 to 10.9 Ghz) of conventional
airborne puls e radar altimeters.
Since no commercial tes t equ ipm en t could meet
specifications, Ryan engin eers developed th eir own
digital tes t set. Made with off-the-shelf integra ted
circuits , it samples the modulator's radio-frequency
1
T output and counts th e chan ges in the output fre-
qu ency as a function of tim e. By averaging the fre-
qu ency changes, the r-f wave's sweep rate or slope
is determin ed as well as th e r-f pulse's lin earity
and upsweep and downsweep times.
Th e di gital averaging measurement technique
can also b e u sed to plot any period ic frequency-
modulated signal as a function of time, provided
that th e puls e time p eriod is greater than the tes t
equipment's 700-microsecond sampling time.
Altimeter operation
Th e output frequency of the radar altimeter's
modulator varies as a sawtooth function, shown
on page 116, from a center fr equency of 99.79
megahertz. Frequency multipliers raise th e output
to an X-band center frequ ency, which is transmit-
ted to th e lunar lan din g surface. The radar return
signal is received, amplified , and the doppler fre-
quency component ex tracted.
A portion of the downsweep section of the
transmitted and return signal is sampled by the
115
 qu ency b etween f'J' and fit an d inversely propor- r
tional to m, th e average sweep rate or slope. As
equation 1 reveals, small varia tions in sweep rate
have a large e ffect on range measurement. Speci fi -
fe r---~~---7~----~~--~------~~~ call y, a 1% error in average slope is reBec ted as a
f L -- - - - - , 2 % crror in range.
I
Th e altimeter's freque ncy modulator operates in
r- --+ T t
',
t---
three mod es. In mode 1, th ere is a h igh deviation
I--- ts --1 / BLANKING SIG NAL
frequ ency. It is u sed wh en th e sp acecraft descends
from an altitud e of 2,500 fee t to th e lun ar surface.
/ t --- In thi s mode, th e out put fr equ ency vari es from

:: U~U
99.998 Mh z to 99.582 Mh z, a d eviation o f ± 208 kh z.

I
1 A large frequen cy d eviat ion is rcgu ired for a more
Frequency of Apollo radar altimet er' s modulator varies
as a sawt ooth fun ction with a center frequency of 99.79
Mhz . A b:anking pulse synchronized to the r-f wave is also
produced. Frequencies fll and f r. are the upper and lower
band limits of the sawtooth; and T and t' are the
downsweep and upsweep pulse periods .
altim eter's range process in g circuitry, an d the in-
s tantaneous frequ ency difference, Llf, between th e
two is de tcrmin ed . Thi s frequ ency differ ence is
related to th e range b e tween th e spacecraft and the
moon 's s urface by th e equation
accurate range de termina tion . Th e second mod e is
switched in when th c spacecraft is at an altitud e
above 2,500 fee t. In this mo ck , th e fr equ cncy onl y
deviates 4l.6 kh z. In both mod es 1 and 2 th e output
wavefo rm s are th e same with equal swcep times.
Th e third mod e is call ed th e d cviation -inh ih it
mode. As th e nam e impli es, th e output frequ ency
is a cons tant, unm od ulated 99.79 Mh z. ~r o d c 3
is actually a built-i n tes t mod e for centcr fr e-
qu ency checks and is not u til izcd durin g High t.
In additi on to th e r-f output, th e modulator also
generates a b lankin g Sig nal. Th is s ignal has th e
same tim e peri od as th e do \\'nsweep porti on of th e
r-f pul se ; and is synchroni zed wilh th e sw eep fun c-

2~{ )
tion. Th e blankin g signal is present on ly during
M = (fr - f R ) = m ( ± fd (1) operation in modes 1 and 2.

\Vh ere Tight test tolerances

f.r = (ranf'mittC'd frequency Th e do\\'n sw('ep average slope proved to b e the
fR = ]'('cf'ivf'd frf'quency mo s t d ifllc ult modu lator parameter to te s t. It mu s t
H range be measured w ith in an accuracy or 0.0.')(/; , 10 tim es
fel clopplf'r fl'equf'ncy
111 aver[lgf' s\ycep m(e = rate of change of the be tter than th c slope tolerance itself. In the low
tl'[ln smitied frequency as a function of tim e d eviation mod e, th e d igital tes t set had to detect a
c ve locity of lig ht. change of 500 parts-per-mi lli on in a slope of ap-
T he range is proportional to th e difference in fre- proximately 12 megahertz p er second. Thi s is
equi valent tu a chan ge in s lOl;C of 6,000 hertz p er
second. Th erefore, over one swc(' p cycle, th e ma\ i-
mum allowab le frequ ency d eviation is 42 h z.
Sin ce lin earity meas uremcnts arc qualiti vely th e
sa me as th e sweep slope measurement, th e .S( ·~
fH lin earity speci fi ca ti on is eas il y im plem cn ted h y
.... .... wh atever techniqu e is chosen to d etec t changes in
.... ....
,, slopc. ~1 (' a s urin g center fr eg ueney with in 0.03%

, accura cy is not a major prob lem si nce the ab solu te

valu e of th e center fr equ ency is very large. I-
Nt :'\ The tes t se t meas ures th e sw('ep slope b y direc tl y
fe ·· - - - - - - - - - - sampling the fr equ ency modu lators r-f output. Thc
di gital averaging tcchniqu e is unaO:ccted b y cen ter
fr eqll ency Bu ctuation s and more than mee ts th e
tes t's accuracy requircmcnts. Lincarity and cen ter
frequ ency error mcas uremen ts rcs ult naturall y fr om
)
th e slope measurement.
fL ~
. ~--~------------___+~--~~- Th e quantity Ll~ = NJ - NJfJ is independ ent of

~' 6'~- center freguency. Th e larger th e sa mplin g intervals,

Frequency modulator'S average slope is rep resented by
the solid line between f ll and f... Nonlinearities are
depicted by the dashed inverte.d S-curve. The tinted
areas are equivalent to the change in transmitter fre-
quency during pulse sample intervals N , and N,o.
th e larger th e value of t. N for a given va lue of m;
thu s, th e resolution improves as the samplin g in -
terval in creases.
H oweve r, th cre is a trad eo ff to b e consid ered.
As th e samplin g interva l in cre a ses, th e slop e aver-
age is taken over a p rog ress ively s maller po rtion

11 6 Electron ics I M ay 1 5, 1967

 RESET
RE GI STER
Nl N2 N3 N4 N5 N6 N7 Na Ng NlO

R-F INPUT 5 6 7
4 a
3 DECK 1 9

BLANKING
SIGNAL

--L SWE EP RATE

~(o-------IL~~ ____j------------------t---r--t--i-------------------r~R~E~S~ETT~
-=- INITIATE
TEST UP COUNT
PU SHBUTTON 5 - DIGIT REVERSIB LE
ACCUMULATOR AND
DIGITAL READOUT
DOWN COUNT

After m ixing and f ilt er ing, modu lato r 's r·f ou t pu t pas ses throu g h AND g<Jtes in'"o t he dig i' <1 1 t c s ~ set s
cou nt accumulator. Sweep ra te (ave rage slope) measurem ents are made by puttir.g S, in the swee p rat e
po sition and averag ing t he transmitter frequen cy over one cycle. This det er m ines t he most
signifi ca nt digits. Averag ing t he freq uency ove r 1,0 0 0 cycie s det ermin es th e I c c: s ~ significa nt digit s.

LRES ET 1RESET
i
I
I
10 - Mh z
CRYSTAL OSC f----
lO - Mh z DIVIDE
BY 7,000
-
AND
GATE
..... 10-STAGE SHIFT REGISTER
I REF LI MITER
I -- ~

NlO
I ENABL7 r
R- F I 1100-Mhz I
INPUT I ISOLATION AMP LI
I
a:
o
I-
I
I LO W-PASS
<l:
-'
=> I
90-Mhz
MIXER 1 FILTER AND l -
Cl CRYSTAL OSC I
o
:::!'
I LIMITER
>-
I
u
Z
W CONTACT
M O MENTAR Y~
=>
o INITIATE TEST ~'--
w
a: PU SHBUTTON CLEAR F 5- DIGI T
"- --lI
...L
r DELAY AND
If ESET
r-
-
I
BINARY
SET F
GATE
ACC UMULATOR AND
DIG ITA L DISPLAY

RE SET
RE SET AN D
RECYCLE
LOGIC

BL ANKI NG
51 GNAL I
J-, BLAN KING -
IN VERTER
SI G AM PL

Center frequency measurements are made by the test set' s configuration , shown in tint. One microsecond
after the test button is pressed, the binary circuit enables the AND gate and the accumulator cou nt s
r·f pulses until the end of the shift register 's t ime interval, N IO• This count is proportional to the
center frequency. To measure upsweep and downsweep time, the test set is switched to the configuration
represented by th e blocks outlined in solid color. The black -bordered blocks are used during bot h tests.

Electronics I May 15, 1967 117

of tlle downsweep cycle. Furthermore, th e longer
the sampling interval, the greater the effect of non-
lin earities on the measurement of average slope.
Linearity measurement
The worst-case nonlinearity in the frequency
modulator output is ± 5% . For a given nonlinearity,
L , the percentage error, p, in the calculated slope
for a given T / M ratio is:
p = L/ [T/ 6 t] - 1
For th ese tes ts, a valu e of 10 w as selected for
T / ~t b ased upon th e maximum allow able slope
error in th e modulator's output, th e tim e allowed for
averaging, and cos t consideration in implementing
th e b alance.
Digital avcraging techniqu es are accura te if a suf-
fi cient number of sweep samples is used.
An error analysis b y a Ryan Aeron autical Co.
ma th ema ti cian, Robert Golden, determined that
averaging th e results of four cons ecutive sets of
4,500 sweep-rate measurements, each in the low
d eviation mod e, and each set having an inh erent
accuracy of 0.5 %, yields a new accuracy of 0.025 %.
By averaging an error over a greater number of
sweeps, accuraci es within 0.01 % can be achieved.
Th e test set also measures th e modulator's center
frequ ency drift to an accura cy of 0.0012 % .
Pulse counters
Th e top diagram on page 117 demonstrates the
mechani zation of th e sweep rate and lin earity meas-
urcmcnt system. Th e clock is a 10 Mhz oscillator
whos e output is divided to form a train of pulses,
each 700 microscconds in length. The 700-,.,.sec
pulse p eriod is the desired sampling interval. The
puls e train is a source of shift pulses for a 10-stage
shift register wired so that only one stage can
contain a log ical 1 at any tim e. A shift pulse ad-
vances that 1 from on e stagc to th e next. As a
result, lin es N 1 throu gh T10 will each carry a pulse
consecutively, for a peri od of 700 ,.,.s ec. vVith th e
down swcep portion of th e bl anking sign al initiating
thi s shifting sequ ence, the ne t result is to divide the
Th e authors
V.M. Andreone designs computers
for data handling. Prior to
joining the Ryan Aeronautical Co.,
he was a systems engineer
developing central data processors
for Apollo tracking ships.
He earned his master's degree at
the University of Pennsylvania.
J.H. Poirier i s a design specia list
at Ryan and ass ist ed in the
developme nt of digita l t est
equipment used to m easure the
accuracy of f-m space alti met ers.
Prior to joini ng Ryan, he was an
inst rume ntat io n engi neer at th e
Naval Electron ics Laboratory.
118
down sweep p eriod into 10 subperiods of 700 f-Lsee.
Th e r-f input from th e modulator is buffered by
passing it through an isolation amplifier, and mixed
with the output of a 90-Mhz crystal oscillator. Th e
difference frequency is filtered and limited and
made availabl e for counting at intervals correspond-
ing to th e selected value of t:,. N.
By means of switch Sl , the frequency over 1 or
1,000 sweep cycles can be averaged_ A test over
on e sweep cycle determines th e most significant
di gits of the measurement. A second nm over 1,000
cycl es defin es th e least signifi cant di gits.
Th e two-deck rotary switch S ~ performs th e
logic that enables th e test set to di stin gui sh b e-
tween nonlin earity and sweep rate tes ts.
To measure sweep rate, switch S2 is placed in
th e SWEEP RATE position. In this position , th e signal
on th e wiper of th e fir st deck is N 1 • Al so N 10 is
conn ected through th e wiper on th e second deck
to input e of gate G ~ . This p ermits th e processcd
r-f signal to pass through to the accumulator during
th ese intervals , for 1- or 1,000-sweep cycles, de-
p ending on th e se tting of St .
During th e Nl tim e interval, th e r-f input drives
th e reversible accumulator in a positive direction ;
during th e NIO interval, the accumulator is driven
in th e opposite direction. The result is t:,. N, which
is accumulated for 1 or 1,000 sweeps, and read out
on th e digital display.
Wh en th e rotary switch is in positions 1 through
9, the gate G ~ at the DOWN input of the accumu la-
tor is inhibited, and a measure of the frequency
within th e selected intcrval is obtaincd. The UP
input is enabled during this interval , and as b eforf"
either 1 or 1,000 consecutive readings may ·b e
stored, and read out at th e end of th e test.
Finding fo
In th e lower diagram on page 117, th e tinted sec-
tions show how th e tcst system is configured to
make mode 3 center fr equ ency measurements. Dur-
ina this tes t, th e blanking pul ses are not u scd.
To make th e test. th e opera tor presses th e initi a te
tes t pu shbutton. This resets th e divide-by -7,000
timing circuit, th e 10-stage shift reg istcr, and th e
5-digit accumulator, and after a 1-f-Ls cc delay; s.ets
th e bin ary logic circuit which enabl es th e tnnlll g
generator and th e AND gate input to th e accumula-
tor. Processed r-f pulses from th e mixer are passed
b y th e AXD gate into th e accumulator. Th e trailing
edge of th e NIO interval pulse, which occurs 7,000
f-Lsec after the start of delay, clears th e binary an,d
terminates th e frequ ency count. Th e accumulator s
digital readout th en shows a count which is pro-
pOl·tional to th e modulator's center frequ ency_
Th e blocks outlined in solid color illustrate the
test system's configuration wh en ch ecking sweep
tim es. For this test, the r-f input is removed and the
blanking signal used. The clock signal is gated
by th e upsweep or downsweep portions of th e
blanking signal into th e accumulator for one cycle.
Th e rcsu lting count of th e clock frequency is pro-
pOl-tional to the up- or downs weep period.
Electronics I May 15, 1967 i.'!