Laser Systems

Dispersion gas flow

Dispersion and gas flow effects in the ring
ring laser
laser gyro

F. Aronowitz and W.
Aronowitz and W. L. Lim
Lim
Systems and
Systems and Research
Research Center,
Center, Honeywell, Inc.
Inc.
2600 Ridgway
2600 Ridgway Parkway, Minneapolis,
Minneapolis, Minnesota
Minnesota 55413
55413

Abstract Dispersion Theory

effects of
The effects of dispersion
dispersionand and gas
gasflow
flowononthe
the operation
operation ofof the
the ring A set
A set of
of equations has been derived
derived which
which describe
describe thethe intensities
intensities
laser gyro
laser gyro are
are summarized.
summarized. The theory is shown shown toto predict a dim- and frequencies of the oppositely directed
directed beams
beams inin the
the laser
laser gyro.
gyro.
ple in the anomalous
ple anomalous dispersion properties of
dispersion properties of the atomic
atomic transi-
transi- The equations have been developed for the weak signal case, using
case, using
tion. Experiments
tion. Experiments werewere carried
carried out
out using
using aa 43.4
43.4 cm
cm and 21.7 cm terms to
terms to third order in
in the
the polarizations
polarization. 5To
Toavoid
avoidthe
thecoupling8,7
coupling6 ' 7
visible laser gyro.
visible laser gyro. Scale factor and nullshift
Scale factor nullshift results
results showed
showed con-con- between the two beams leading to frequency synchronization,
synchronization, the
sistency and
sistency and confirmed
confirmed the existence of of the
the dimple. The
The peak -peak
peak-peak equations are considered only only for
for rotation rates much higher than
relative magnitude of the scale
relative scale factor dimple was found to be ap- the lock -in threshold.
lock-in threshold. The
The equations for the the intensities
intensities II,Ii, I2
I 2 and
and
proximately 25
proximately 25 xx 10"6
10 -6. Gasflow
. Gas flowwas
wasvaried
variedbyby unbalancing
unbalancing the the frequencies GUI,
frequencies w1, cu2
w2ofofthe
the two
two beams
beams can be written
writtenas as88
discharge currents in the two two legs
legs of the laser gyro. A A nullshift sen-
sitivity of
sitivity of 0.0066
0.0066 deg/hr/^amp
deg /hr /µampofofunbalance
unbalance waswas found. Analysis
Analysis
of the data showed
of showed the
the nullshift
nullshift sensitivity
sensitivity to velocity
velocity flow to be be Il = (c/L) - - 81212] (1)

4.4 deg /hr /cm /secofofvelocity

deg/hr/cm/sec velocityunbalance.
unbalance. ¡ß111
I2 = (c/L)I2[a2 -p2l2-02iIi]
I2 (2)
Introduction
coi QI ++0101+ +y111
w1 = Q1 T12 I2 .
£Ji++T1212. (3)
There are three major
There major sources
sources of of error
error inin the
thelaser
lasergyro:
gyro:lock -in,
lock-in,
nullshift and
nullshift and scale
scale factor.
factor.'1 IfIf the
theeffects
effects of
of lock -in are neglected,
lock-in CO2
GJ 2 ==&2
Q2 ++ 02
02 ++t;2I2
&I2++T21 I 1.
ToJi. (4)
source of
the source of the
the nullshift
nullshift and
and scale
scale factor
factor error is dispersion
dispersion of of the
the
active medium,
active medium, commonly
commonly called called mode
mode pulling
pulling andand pushing.
pushing.'2
This paper discusses
This these dispersion
discusses these dispersion effects
effects on
on the
the operation of In the equations:a aisisthe
amplitudeequations:
the amplitude andO0 are
gain,ß/3and
excessgain,
theexcess self
theself
are the
the laser
the gyro. In particular, the
laser gyro. the modification
modification to to the
the dispersion
dispersion and mutual
mutual gain
gain saturation,
saturation, respectively.
respectively. InInthe
thefrequency
frequency equa-
equa-
caused byby collisional
collisional radiation
radiation trapping3
trappings andand motion
motion of of the
the gas
gas are
are tions:
tions: Q is the cavity frequency, ao is
Q is is the dispersion
dispersion correction due
considered.'
considered. 4 to the gain medium; ¿£ and
and TT are dispersion corrections due to gain
are dispersion gain
Section IIII gives
Section givesaa description
descriptionof ofthe
theerrors
errors introduced
introduced when when the
the saturation. Subtracting
Subtracting Eq.
Eq. (3)
(3) from
fromEq.Eq.(4),
(4),an
anexpression
expression can
can be
dispersive properties of
dispersive properties of the
the active medium are are considered.
considered. Scale
Scale obtained for the beat frequency
frequency as
factors experiments
factors experimentsverifying
verifyingthe the dispersion
dispersion properties
properties are are de-
de- F
scribed
scribed inin Section
Section III. Section
Section IV describes
describes the generalization of
CO = Q + (02 - QI) +
CA.) = Q + (02 - 00 + (T2, - S1)11 -- (T12
(T21 - - S2)
( Tl2 - £2) 12
I2 (5)
(5)
dispersion properties
the dispersion properties toto take
take into account motion of of the active
active In Eq. (5)
(5) Q
Q is
is the
the empty
empty cavity
cavity frequency
frequency difference.
difference.
gain atoms.
gain atoms. Experimental results are
Experimental results are presented,
presented, confirming
confirming the The effect
effect of
of the
the homogeneous
homogeneous saturation
saturationcan canbe
beexpressed
expressed ex-
ex-
theory. plicitly by letting8
letting8
- of ßi R/G , (6)
This work
This work was
was partially
partially supported by
by the
the Naval
Naval Air
Air Systems
Systems Command under
under Con-
Con-
N00019-75-C-0306.
tract N00019 -75 -C -0306. Ti Tij-OißiR/G,i#j
*j == 1,2 (7)
1528 received
1528 receivedAugust
August14,
14, 1978.
1978.This
Thispaper
paper was
was presented
presented at
at the SPIE
SPIE seminar on
on
Laser Inertial
Laser Inertial Rotation
Rotation Sensors,
Sensors, August 30 -31, 1978,
August 30-31, 1978, San
San Diego,
Diego, and appears
appears in
in In Eqs. (6)
(6) and (7), R is
is the
the saturation
saturation parameter
parameterand
andGGisis the
the laser
laser
SPIE Proceedings Volume 157.157. gain.

376 // OPTICAL
376 OPTICAL ENGINEERING
ENGINEERING / Vol.
Vol. 18
18 No.
No. 44 / /July- August 1979
July-August 1979

Downloaded From: https://www.spiedigitallibrary.org/journals/Optical-Engineering on 06 Jun 2019

Terms of Use: https://www.spiedigitallibrary.org/terms-of-use
 DISPERSION AND
DISPERSION AND GAS EFFECTS IN
FLOW EFFECTS
GAS FLOW THE RING
IN THE RING LASER GYRO
LASER GYRO

It is more convenient
is more convenient to
to express
expressEq.
Eq. (5)
(5) in
in terms
terms of
of the
the sum
sum and transitions of
of the
the neon
neon isotopes).
isotopes). In
In the
theprevious
previous experiments,8
experiments, 8 the
difference of
difference of the two intensities,
intensities, experimental accuracy was
experimental accuracy was limited
limited so
so that
that this dimple
dimple was not
observed.
I= (I2 + II), i = 2 (I2 Il) (8)
(8) It should
should be noted
noted that
that since
since the
the dominant
dominant contribution
contribution So
S o is
is
negative, A
negative, A isis defined
defined as
as aa positive
positive quantity.
Then using Eqs. (6)
using Eqs. -(8), Eq. (5)
(6)-(8), (5) becomes
becomes
w == QQ ++(o2
o> (o2- -ol)01)+ 2+ g2- (£
T)-1 T)- [(2
i -R
[(2/G) (02(32
R/G) (02/52- -o1ß1)
Scale Factor
Scale Factor Experiment
A scale factor
A scale factor experiment was performed to checkcheck for the dimple
dimple
+ (T12 - T21)++ (Sl
+ (Ti2-T21 ) - 2)]
(£l-£2 I
)]I. (9)
(9) at the power peak.
peak. A
A HeNe
HeNe ring laser,
laser, operating
operating at
at the
the0.633
0.633 µm
/urn
transition, was
was used.
used. The
The laser
laser was
was filled
filled with
with an
an equal
equal mixture
mixture of
The term
The term in i gives rise to
gives rise to aa bias,
bias, and will
will be
be neglected.
neglected.
the Ne21 and Ne22
Ne20 and Ne22isotopes.
isotopes.The
Thecavity
cavity length
length was 43.4 cm, giving
giving
Since each of
Since each of the coefficients
coefficients in in Eq. (9)
(9) signifies
signifies its
its being
being
a longitudinal
longitudinal mode
mode spacing
spacing of
of 691 MHz. For improved accuracy,
accuracy,
evaluated at
evaluated at the particular frequency,
frequency, aa Taylor
Taylor expansion
expansion can be
the laser was constructed
constructedfrom
fromaasolid
solidblock°
block9 of
of Cer -Vit with inter-
Cer-Vit
made about the average
average ofof the
the two
two cavity
cavity frequencies.
frequencies. nal mirrors.
nal
Then Eq. (9)
(9) can
can be
beexpressed
expressed as as
Readout was
Readout was implemented
implemented for for the laser
laser by
by combing
combing the two two
w
co == QS(1 -A) ,
(1-A) (10) beams to
beams to form
form aa fringe
fringe pattern.
pattern. The fringe pattern
pattern was
was monitored
where by a detector
detector with
with logic
logic designed
designed to give
give aa digital
digital pulse
pulse for each
fringe maximum. Each Each pulse
pulse then
then corresponded
corresponded to to the
the gyro
gyro turn-
turn-
ing through
through 1.5 seconds of ofarc.
arc. Data
Data was
was obtained
obtained by by turning thethe
-A--2-
A= (L) [So +SR +ST
21(p
+s] (11)
(11)
ing
gyro at
gyro
1.5 seconds
at a constant rate andand measuring
measuring thethe total
total number
numberof of counts
counts
as the
as the gyro
gyro turned
turned through a complete revolution.
revolution. The accuracy of of
In Eq. (11),
(11), Kp
K/u is the Doppler
is the Doppler width of the transition. the gating
the gating was one count.
was one count. To reduce the one one count quantization
quantization
The S terms for
S terms for the
the various
various contributions,
contributions, have been evaluated'
evaluated8
for an equal
equal mixture
mixtureofofthethetwo
twoneon
neonisotopes
isotopes(Ne20,
(Ne20 ,Ne22).
Ne22). They
They
error, data
error, data were
were typically
typically obtained
obtained by by sectoring
sectoring through
through ten
revolutions.
are a function
function of the frequency tuning of of the
the laser.
laser.
Figure 2 shows typical scale factor data obtained
obtained withwith the
the 43
43
shows typical scale factor
Figure 11 shows
shows aa plot
plot of
of the
the SS coefficients.
coefficients.STSTand
andS^S are usually
small compared to
small compared to SR.
SR. The effect of SR is to
SR is to reduce
reduce the effect of
effect of
cm laser
cm laser as
as a function of frequency tuning.
frequency tuning.
SSo. It can
0 . It can also
also be
be seen
seen that
that aa very
very small
small dimple
dimple was
was predicted
predicted at
at
the peak
peak gain (midway
(midway between
between the centers
centers of the
the two
two atomic
atomic
8, 23, 570
8,23,570
o o 0 0 0o 00
0 0
0 0
0 0°
O 0
a;
UJ
0
o
.2 SR 00
1—
o
0

o
823, 560
823, 560 o

`
0
C£
i0
—
o
0 o

0
Z

scc
0

AV
S

p `í/ i
CO

823, 550
823,550 '
° 2 __

ó -200
-200 -100
-100 00 100
100 200
200

LL S TO TAL
FREQUENCY
FREQUENCY MHz
MHz

2. Scale
Figure 2.
Figure Scale factor vs
vs frequency
frequency tuning
tuning for
forAA ==0.633
0.633 µm
/^m and
and aa
-.4 -
-.4 - cavity length of
of 43
43 cm.
,.,-,

oz
_ The rotation
rotation rate
ratewas
was200200deg /sec. Note
deg/sec. Note that
that unlike
unlike previous
previous
on
data,'8 the
the scale
scale factor
factor data
dataare areconcave,
concave, opening
opening downwards.
,
o, data,
This isis interpreted
This interpreted as due to the fact
as due fact that
that for
for single
downwards.
single mode
mode opera-
á
U tion, the frequency tuning range was limited. Hence, for the data
tion,
shown in
shown in Figure
Figure 2, the tuning
tuning range
range waswas over
over the
the central
central region
region
only, which consisted
only, entirely of the dimple.
consisted entirely
For visible
visible laser
laser gyros
gyros ofof this
this size,
size, only
only scale
scale factor
factor versus
versus fre-
fre-
quency turning data
quency turning data with
with aa negative
negative curvature
curvature have have been
been ob-ob-
8 Sc.'
served
served in in our laboratory.
laboratory. To To allow
allow single
single mode
mode turning over over a
large
large range,
range, aa smaller
smaller sized
sized gyro
gyro having
having aa larger
largerlongitudinal
longitudinal
mode spacing is required.
, i
The experiment was was repeated with aa similarly
similarly constructed gyro gyro
-400
1

,606 400 -200

-200 0
0' 200
200 400
400 600
600 of
of smaller perimeter. The
smaller perimeter. The path length length was
was 21.721.7 cm,
cm, giving
giving aa
FREQUENCY TUNING -MHz
FREQUENCYTUNING - MHz longitudinal
longitudinal mode mode spacing
spacing of of approximately
approximately 1400 1400 MHz.
MHz. ThisThis
allowed aa larger tuning range
allowed range with
with single
single mode
mode operation.
operation.
Figure
Figure 1. Components of
1. Components of the
the scale factor correction
scale factor correction as as obtained
obtained Figure 3 shows scale factor
shows scale factor versus
versus frequency
frequency tuning
tuning data for a
from the solution
solution of
of the
the theoretical
theoreticalequations
equations given
given ininRef.
Ref. 4.
4. rotation rate
rate of
of 250
250 deg /sec. The predicted
deg/sec. predicted dimple
dimpleisis clearly
clearly seen.
seen.

July-August
July- 1979 // Vol.
August 1979 Vol. 18 ENGINEERING //
No. 44 / OPTICAL ENGINEERING
18 No. 377
Downloaded From: https://www.spiedigitallibrary.org/journals/Optical-Engineering on 06 Jun 2019
Terms of Use: https://www.spiedigitallibrary.org/terms-of-use
 F. ARONOWITZ AND W.
F. ARONOWITZ W. L.
L. LIM

411,840 - BEAM (1) BEAM (2)

1400 MHz
(ONE WAVELENGTH)

411,830
411, 830

V >0 V< 0

411,820
411, 820 -

LASER FREQUENCY
LASER FREQUENCY TUNING
TUNING

Figure 3.
Figure 3. Scale
Scale factor vs frequency
frequency tuning
tuning for
forAA = 0.633
0.633 pm and
and aa
cavity length
length of
of21.7
21.7 cm.
cm.
CATHODE

The solid line isis aa smooth

solid line smooth curve
curve drawn
drawn through the datadata points,
points, Figure 4.
Figure 4. Gas flow convention.
Gas flow convention.
showing
showing the
the high
high resolution
resolution of
ofthe
the data.
data. Although
Although nono attempt
attempt was
made to
to parameter fit the data, aa theoretical
theoretical plot
plot of
of Eq.
Eq. (10)
(10) was
made. The
The ten -count difference
ten-count difference between the extremes
extremes was found As has been
As has been shown,
shown, the flow
flow of the active
active atoms
atoms isis due to
to the
the
and the two minima
minima were were spaced
spaced approximately
approximately asas shown. Langmuir
Langmuir effect." Alongthe
effect. 4 Along the tube
tube center,
center, the flow
flow is is from
from the
the
anode
anode to the
the cathode."
cathode. 4 Hence,
Hence, the
the velocity
velocity flow
flow isis positive
positive and
Velocity Flow Nullshift negative on the left and rightright sides
sides of the tubes, respectively.
respectively. Note
that for
for earth's
earth's rate
rateto
to be
be considered
considered positive,
positive, thethe frequency
frequency of the
It has been shown4
shown4 thatthat aa nullshift
nullshift (output beat frequency
frequency in the CW beam (Is) (I2) is higher than
than the
the frequency
frequency of of the
the CCW
CCWbeambeam(I1).
(1^.
absence of
absence of an
an input
input rotation rate) can can occur
occur ifif the gas atoms have a As
As aa consistency
consistency check,
check, the
the discharge
discharge currents
currents were
were unbalanced
unbalanced
non-zero
non -zeroaverage
averagevelocity.
velocity.This
This isis due
due essentially
essentiallyto to the
the fact
fact that and
and the
the nullshift
nullshift measured.
measured. Figure
Figure 5 shows
shows the results of the
results of
when light travels through aa moving moving medium,
medium, the theindex
index of of refrac-
refrac- measurement.
measurement. The slope slope of the best -fit straight
best-fit straight line
line through
through the
is different
tion is different from
from that
that of
of aa stationary medium. Hence, for op- op-
positely directed
positely directed beams
beams traveling
traveling through the the moving
moving medium,
medium,
the material is is anisotropie
anisotropic with respect
respect to
to the
the index
index of of refraction.
refraction.
This isis equivalent
This equivalent to to an
an optical
optical path difference, which results results in a
nullshift beat frequency.
nullshift
It has
has been
been shown4
shown4 that the dominant
dominant effect
effect in in the
the HeNe
HeNe gasgas 1.0
mixture, which has has aa non -zero velocity
non-zero velocity flow,
flow, is is due to the shift in
the dispersion
the dispersion characteristics
characteristics of of the
the active
active medium.
medium. This This is
equivalent to
equivalent to Doppler shifting
shifting thethe oscillation
oscillation frequencies
frequencies of the
two beams
beams and can be represented
represented by writing"
writing4
|2 5 i ->|2
42,1 ? i ±± KV
42,1 KV/K^
/Kµ (12)
In Eq. (12)(12) 4| is the oscillation difference from the
is the oscillation difference from the peak peak power fre-
fre-
quency and is
quency is normalized
normalized with respect to the Doppler Doppler width
width Kµ.
K/L(. -0.2 -0.1 0.1 0.2
V
V isis the
the non-zero
non -zerovelocity
velocityflowflowofofthe
thegas,
gas, defined
defined as
as positive
positive when
when DIFFERENTIAL
DIFFERENTIAL CURRENT
CURRENT -- ma
ma
the active atom flow is in the direction of positive
is in positive rotation rate. AA (12-1!)
(12 - 111
rotation is is positive when w2 co2 > wt.
coi.
By
By using
using Eq.Eq. (12)
(12) in
in Eq. (9) and carrying out the Taylor
Eq. (9) Taylor expan-
expan-
sion, aa more
sion, more general
general formform ofof Eq.
Eq. (10)
(10) is obtained:
is obtained:
w
co == QQ (1 -A) --22 KVA
(1-A) KVA (13) 1.0

The "A" coefficient is is as

as defined
defined inin Eq. (11).
(11). Thus,
Thus, the
thedispersion
dispersion
effects due to
effects due to aa splitting
splitting ofof the
the gain
gain curve
curve with
with respect
respect to to the two
beams
beams isis equivalent
equivalent to to the dispersion effects due
dispersion effects due to
to the
the rotational Figure 5.
Figure Differential current
5. Differential currentsensitivity
sensitivityfor A A==0.633
for 0.633 pm
^m and
and a cavity
splitting of the frequencies
frequencies with a degenerate
degenerate gain
gain curve.
curve. length of 43
43 cm. A sensitivity
sensitivityofof0.0066
0.0066 deglhrlpamp
deg/hr/^amp was found.
found.
should be
The dimple should be observable
observable with
with aa nullshift
nullshift measurement,
measurement,
as well
as well as aa scale
scale factor
factor measurement.
measurement. Nullshift
Nullshift measurements
measurements data was
data was 0.0066
0.0066 degdeg/hr/ptamp
/hr /µamp ofofcurrent
current unbalance. The fact that that
were carried outout using
using the
the previously
previously described
described 43 cm laser gyro. an offset
offset of -0.05 deg /hr was found indicates that
deg/hr that balanced
balanced cur-cur-
Since lock
Since -in effects
lock-in effects mustmust be be compensated,
compensated, the the gyro
gyro waswas rent does
does not
not guarantee
guarantee balanced
balancedvelocity
velocity flow.
flow. Construction
Construction
mechanically dithered.9
mechanically dithered. The readout
9 The readout was
was designed
designed so so that
that the
the asymmetries
asymmetries could could bebe the cause.
cause. As
As will
will be
be discussed,
discussed, thisthis offset
offset
counts due
due to
to the
the sinusoidal
sinusoidal mechanical
mechanical dither
dither were
were automatically could be caused by by aa velocity
velocity flow
flow unbalance
unbalanceofof0.01 0.01cm /sec.
cm/sec.
compensated and
compensated and not observed
observed in in the
the output. The sense
sense ofof the
the slope
slope isisof
ofmore
moreinterest.
interest. From
From Figure 5, 5, an in-
in-
Figure 44 shows
shows aa schematic
schematic of of the laser gyro with the the sign
sign con-
con- crease
crease in the discharge current
current ininleg
leg22over
over leg
leg 1,1, gives
gives aa positive
positive
vection. The gyro was designed
vection. designed with parallel excitation
excitation legs.
legs. The nullshift. As
As can be seenseen from
from Figure 4, a relative increase
increase in cur-
cur-
cathode was
cathode was common.
common. The The component
component of of earth's
earth's rate
rate normal
normal to rent in
in leg
leg 22 gives
gives rise to aa negative
negative net
net velocity
velocity flow.
flow. Using
Using Eq.
the plane of the laser is shown and has has magnitude
magnitudeof of10.64
10.64 deg
deg/hr.
/hr. (13),
(13), this results
results in a positive
positive nullshift
nullshift in
in agreement
agreement with withthe theex-
ex-
This corresponds
This corresponds to 677 counts for aa 100 -second sample
100-second sample interval. periment.

378
378 // OPTICAL
OPTICAL ENGINEERING
ENGINEERING //Vol.
Vol.18
18No.
No.44/ July-
/ July-August 1979
August 1979
Downloaded From: https://www.spiedigitallibrary.org/journals/Optical-Engineering on 06 Jun 2019
Terms of Use: https://www.spiedigitallibrary.org/terms-of-use
 DISPERSION
DISPERSION AND
AND GAS
GAS FLOW
FLOW EFFECTS
EFFECTS IN
IN THE LASER GYRO
THE RING LASER GYRO

Further experiments
Further were performed,
experiments were performed, in
in which the
the nullshift
nullshift
was measured
measured asas a function of frequency tuning. To
To obtain
obtain max-
max-
imum sensitivity, the laser
laser was operated
operated with
withone
onedischarge
dischargeleg
legex-
ex- LEG ==00
Ii LEG
tinguished.
Figure
Figure 66 shows the results
shows the results obtained
obtained with
with a discharge
discharge in leg
leg 11
only. The
only. The contribution
contribution due
due toto earth's
earth's rate has
has been
been subtracted.
subtracted. 18
1.6 ma
2.0 ma
1.8ma
1.4 ma
-16 _ \ 2 LEG = 0
17
17

£ -17
16 ma
1.4 ma
o
LU
o

6 ma
-18
180
180 120
120 60
60 00 60
60 120
120 180
180
ma FREQUENCY -
FREQUENCY MHz
-- MHz

7. Differential current
Figure 7.
Figure current bias
bias vs
vs frequency
frequency tuning
tuningfor
forAA == 0.633
0.633
pm and aa cavity length
p<m and length of
of 43
43 cm.
cm.
-19

180 120 60 0 60 120 180 18 -

FREQUENCY - MHz

Figure 6.
Figure 6. Differential current bias
Differential vs frequency
bias vs frequency tuning
tuning for A
A = 0.633
and aa cavity length of
^m and
Nm of 43
43 cm.
cm.

Because ofthe
Because of thelarge
largecavity
cavitylength,
length, the
the tuning range was restricted
so
so that data were
were obtained
obtained only
only over
over the central
central dimple
dimple region.
region. 17
Since the
Since the discharge was only in leg 1,
1, as can be seen
seen from Figure 4,
the velocity flow
flow was positive.
positive. Using Eq. (13),
(13), aa positive
positive velocity
velocity o
LU
flow gives
flow gives aa negative
negative nullshift,
nullshift, which
which isis in in agreement
agreement with with the o
data shown in Figure 6. 6. Note
Note that
that higher
higherdischarge
discharge currents
currents give
give
rise to a larger velocity
rise to flow and a more
velocity flow more negative
negative nullshift,
nullshift, as
as ex-
ex-
pected.
pected. Since
Since "A""A" isis positive
positiveand and data
data are
are in the central
central dimple
dimple <
OQ
region
region only,
only, thethe nullshift
nullshift versus
versus frequency
frequency tuningtuning isis concave,
concave,
opening downwards. This This shape
shape isis similar to thethe shape
shape ofof the scale 16
factor data shown
shown in Figure
Figure 2. 2.
Similar data were obtained with the
Similar data the discharge
discharge onlyonly inin leg
leg 2,
2,
and areare shown
shown in in Figure
Figure 7. 7. As
As expected,
expected, the thenullshift
nullshiftisis positive
positive
since
since the velocity flow isis now
velocity flow now negative. Also, Also, increasing currents
result in an even more positive nullshift. Because of the change in
the sign
the sign of of the velocity flow, the sense
velocity flow, sense of the frequency
frequency tuning
tuning
characteristics alsoalso change.
change.
-180
-180 -120
-120 -60
-60 00t 60
60 120
120 180
180 200
200
Equation (13)
Equation (13) waswas numerically evaluated. A velocity
numerically evaluated. flow of
velocity flow of
-3.5 cm
-3.5 /sec was
cm/sec was chosen.
chosen. Figure
Figure 88 shows
showsthe the resultant
resultant nullshift ver- FREQUENCY -
FREQUENCY MHz
-- MHz
sus frequency
sus frequency tuning.
tuning. As As expected
expectedthe the shape
shape isis similar
similar toto that
that
found experimentally for the case of a negative velocity, velocity, asas shown
Figure 8.8.
Figure Velocity flow
flow bias
bias theoretical
theoretical curve
curve for
for I,I, leg
leg == 00 and
and v =
in Figure 7.
in 7. This
This confirms
confirms thethe assertion
assertion that
that over
over the tuning range -3.5 cm.
considered, only
considered, only the simple region region isis observed.
observed.
The shape
shape of of the
the theoretical
theoretical curve
curve isis slightly
slightly broader
broader than
than the
the
experiment. This may be due to uncertainities
experiment. uncertainities in in the
the collisional
collisional ef- As shown
As shown by
by the
the data presented
presented in Figure
Figure 3,
3, the
thenullshift
nullshift sen-
sen-
fects which
fects which cause
cause thethe dimple
dimple or or questionable
questionable use use ofof the
the low
low gain sitivity to current
sitivity to current unbalance
unbalance can
can be
be obtained.
obtained. The
The theoretical
theoretical
third -order model.
third-order model. Further investigation
investigation is is needed
needed to to resolve
resolve this analysis allows an estimate to
analysis allows to be
be made
made ofofthe
thenullshift
nullshiftsensitivity
sensitivity
question. to velocity flow.
flow. Using the nullshift result found in Figure 8, at the
Using the

July-August
July- August1979
1979/Vol.
/ Vol.18
18No. 4 /OPTICAL
No.4 ENGINEERING / 379
/OPTICAL ENGINEERING 379
Downloaded From: https://www.spiedigitallibrary.org/journals/Optical-Engineering on 06 Jun 2019
Terms of Use: https://www.spiedigitallibrary.org/terms-of-use
 F. W. L.
F. ARONOWITZ AND W. LIM
L. LIM

peak
peak power,
power, a sensitivity of 4.4 deg/hr/cm/sec
sensitivity of deg/hr/cm/sec was was found.
found. For the beat frequency
frequency is is completely
completely equivalent
equivalent to to aarotation -induced
rotation-induced
high performance laser gyros,
gyros, nullshift stabilities on the order
order oo beat frequency.
beat frequency. Experiments
Experiments were were performed
performed to to measure
measure the
0.01
0.01 deg /hr are
deg/hr are required.
required. This
This means
means that for
for the
the laser
laser used,
used, the
the velocity induced
velocity induced nullshift,
nullshift, by
by unbalancing the parallel dischargedischarge
velocity flow must
velocity flow must be
be maintained to within
within 2323 µm/sec.
/^m/sec. currents which
currents which provide the laser laser gain.
gain. AAnullshift
nullshift sensitivity
sensitivity of of
Infrared lasers
lasers usually show
show a nullshift
nullshift sensitivity
sensitivity to discharge
discharge 0.0066 deg /hr /µamp of
deg/hr//^amp of unbalance
unbalance was found. Nullshift
Nullshift versus fre-
versus fre-
current unbalance
unbalance which
which isis two
two orders
orders of
of magnitude
magnitude higher"
higher4 than
than quency tuning
quency tuning waswas measured
measured for operation with with aa single
single leg
leg dc
dc
that found in visible lasers.
lasers. This
This implies
implies that
that the conditions on the discharge. The results
discharge. results were consistent
consistent with the the scale
scale factor data
velocity flow unbalance
velocity flow unbalance are even more stringent. and confirmed the theoretical radiation
radiation trapping
trappinginduced
induceddimple.
dimple.
A correlation
A correlation of the data
data with
with the
thegyro
gyromodel
modelgave
gave the
the velocity
velocity in-
Conclusions duced nullshift sensitivity
duced as 4.4 deg/hr/cm/sec.
sensitivity as
effects of
The effects of dispersion
dispersionand and gas
gas flow
flowonon the
the operation
operation of of the
the laser
laser The theoretical
The theoretical equations predicted a slightlyslightly broader shape of of
gyro have
gyro have been
been reviewed.
reviewed. The anomalous
anomalous dispersion
dispersion duedue to thethe the dimple
the dimple with
with respect
respect toto frequency
frequency tuning.
tuning. TheThediscrepancy
discrepancy
could possibly
possibly bebe due to uncertainty
uncertainty inin the
theknowledge
knowledgeof of the
the colli-
colli-
gain medium is shown to cause a scale scale factor
factor correction. TheThe colli-
colli-
sional radiation
sional radiation trapping
trapping effect
effect isis shown
shown to to partially
partially cancel
cancel thethe sional parameters
sional parameters or or the need for higher
higher accuracy
accuracy thanthan provided
provided
differential mode
mode pulling
pulling caused
caused by by the unsaturated
unsaturated dispersion
dispersion by the third -order theory.
third-order
differential
curve. The
curve. The curvature
curvature of of the
the radiation
radiation trapping contribution,
contribution, as as aa
function ofof frequency
frequency tuning,
tuning, is shown to
is shown to dominate
dominate near the peak
gain frequency. This causes
causes a dimpling effect in the mode mode pulling
pulling References
characteristicsnear
characteristics near the
the peak
peak gaingain frequency.
frequency. ThisThis predicted
predicted 1. Aronowitz,
1. Aronowitz, F., Laser Laser Applications.
Applications. Vol. Vol. 1,1, Academic
Academic Press,
Press, New
New
dimpling was confirmed experimentally
experimentally usingusing aa visible
visible HeNc
HeNe ringring York, 1971.
1971.
laser gyro. A
laser A relative
relative scale
scale factor
factorpeak -peak dimple
peak-peak dimpleofof25 25x x101()~6
-6 2. Lamb,
Lamb, Jr.,
Jr.,W.
W.E.,E.,Phys.
Phys.Rev.
Rev.A134,
A134,14291429 (1964).
(1964).
was found. 3. Aronowitz,
Aronowitz, F.,F., Appl.
Appl. Opt.
Opt.11,
11,2146
2146(1972).
(1972).
The tuning range
range over
over which
which the the dimpling
dimpling waswas found
found waswas ap-
ap- 44. Podgorski,
Podgorski, T. T. J.
J. and
andAronowitz,
Aronowitz, F., F., IEEE
IEEE J.J.Quantum
QuantumElectron.
Electron.
4, 11
11 (1968).
(1968).
proximately 700 MHz. Hence aa small small laser
laser gyro
gyro is
is needed to allow
allow 5. Aronowitz,
Aronowitz, F.,F.,Phys.
Phys.Rev.
Rev.A139,
A139.635 635(1965).
(1965).
sufficient single
sufficient single mode
mode tuning
tuning to observe
observe the dimple. 6.
6. Aronowitz,
Aronowitz, F. F. and
andCollins,
Collins, R.
R. S.,
S.,Appl.
Appl. Phys.
Phys. Lett.
Lett.9.,9,55
55(1966).
(1966).
The mode
The pulling and pushing
mode pulling pushing terms were generalized
generalized to in- 7.
7. Macek,
Macek, W. M., M., Davis,
Davis, Jr., D.
D. T. T. M.
M. Olthris,
Olthris, R.R. W.,
W., Schneider,
Schneider, J. R.
R.
clude motion
motion ofof the
the gain atoms.
atoms. The results
results showed
showed a splitting of of and White, G. R.,
and R., Optical
OpticalLasers,
Lasers, P. P. I.I.B.
B.Press,
Press, Brooklyn,
Brooklyn, 1963.
1963.
clude 8. Aronowitz,
8. Aronowitz, F., F., Killpatrick,
Killpatrick, J.J. and
and Callaghan,
Callaghan, S. S. P.,
P., IEEE
IEEE J. Quan-
Quan-
the
the dispersion curvefor
dispersion curve for the
the two
two beams
beams which
which resulted
resulted in in a tum Elec. QE/O,
QE10, 201201 (1974).
(1974).
nullshift.
nullshift. It was
was found
found that
thatthe
thevelocity- induced contribution
velocity-induced contribution to to 9. Killpatrick,
9. Killpatrick, J., IEEE
IEEE Spectrum
Spectrum 4. 4, 44
44 (1967).
(1967). 3

ENGINEERING / Vol.
OPTICAL ENGINEERING
380 I/ OPTICAL
380 Vol. 18 No. 44 / /July-
18 No. July-August 1979
August 1979
Downloaded From: https://www.spiedigitallibrary.org/journals/Optical-Engineering on 06 Jun 2019
Terms of Use: https://www.spiedigitallibrary.org/terms-of-use