Document 12961196

advertisement
! ! $& !#!'#
# %!" " #"#&
"
Methods for enhancing the dynamic range of the intensity noise technique
for high-frequency photoreceiver calibration are proposed and experimentally demonstrated. These methods combine recently developed EDFA*
technology with spectral filtering techniques. The intensity noise
calibration technique is portable, easy to use, and field deployable.
& $" & & !
""30 2$ "' 0 "2$0(8 2(-, -% 2'$ %0$/3$,"7 0$1.-,1$ -% ,
-.2(" * 0$"$(4$0 (1 (+.-02 ,2 2- $,130$ 2' 2 2'$ 0$"$(4$0 (1
"-+. 2(!*$ 5(2' 2'$ # 2 20 ,1+(11(-, 0 2$ 300$,2*7 ,3+9
!$0 -% 2$"',(/3$1 $6(12 2- "' 0 "2$0(8$ 2'$ %0$/3$,"7 0$9
1.-,1$ -% -.2(" * 0$"$(4$01 ,$ +$2'-# (1 2- "-+.32$ 2'$
-30($0 20 ,1%-0+ %0-+ 2'$ 2(+$9#-+ (, (+.3*1$ 0$1.-,1$
0$/3$,"7 #-+ (, 2$"',(/3$1 13"' 1 2' 2 31$# (, 2'$ *(&'25 4$ "-+.-,$,2 , *78$0 $+.*-7 %0$/3$,"79
15$.2 1(,31-(# * +-#3* 2(-, -% 2'$ -.2(" * (,2$,1(27 31(,&
'(&'9%0$/3$,"7 (! +-#3* 2-0 '(1 **-51 "-++$0"( **7
4 (* !*$ 0$1.-,1$ +$ 130$+$,21 %0-+ 8 2- 8
'$ -.2(" * '$2$0-#7,$ 2$"',(/3$ 31(,& 25- #9 * 1$01
' 1 #$+-,120 2$# " . !(*(27 %0-+ 8 2- 8 2
5 4$*$,&2' -% + '$1$ 2$"',(/3$1 ** (,4-*4$ 1.$9
"(%(" 20 #$9-%%1 +-,& %0$/3$,"7 "-4$0 &$ $6.$0(+$,2 *
"-+.*$6(27 ,# 1$,1(2(4(27 ,# ,-,$ 0$ "-+.*$2$*7 1 2(1% "9
2-07 2 (1 #$1(0 !*$ 2- ' 4$ 2'$ " . !(*(27 2- +$ 130$ 2'$
!0- #! ,# %0$/3$,"7 0$1.-,1$ 5(2' 1(+.*$ 03&&$# -.2(" *
(,1203+$,2
'$ (,2$,1(27 ,-(1$ 2$"',(/3$ -%%$01 2'$ .-11(!(*(27 -% +$ 9
130(,& %0$/3$,"7 0$1.-,1$ "' 0 "2$0(12("1 -% .'-2-0$"$(4$01
"0-11 2'$ $,2(0$ %0$/3$,"7 1. , -% +-#$0, $*$"20(" * 1.$"9
203+ , *78$01 %-0 $6 +.*$ )8 2- 8 %-0 2'$ '(1 (,2$,1(27 ,-(1$ +$2'-# 5 1 %(012 #$+-,120 2$#
31(,& 1$+("-,#3"2-0 -.2(" * +.*(%($0 1 1-30"$ '(1
2$"',(/3$ (1 -% . 02("3* 0 (,2$0$12 !$" 31$ 2'$ ,-(1$ $6(121 2
** %0$/3$,"($1 1(+3*2 ,$-31*7 .$0+(22(,& 4$07 0 .(# -.2(" *
0$"$(4$0 "' 0 "2$0(8 2(-, ##(2(-, **7 , 3,.-* 0(8$# 1'-029
"-'$0$,"$9*$,&2' -.2(" * 1-30"$ (1 31$# '(1 (1 #4 ,2 &$-31
$!03 07 $5*$229 ") 0# -30, *
!$" 31$ (2 + )$1 2'$ +$ 130$+$,21 (++3,$ 2- .-* 0(8 2(-,
#0(%21 ,# 2(+$94 07(,& (,2$0%$0$,"$ $%%$"21 %0-+ +3*2(.*$
-.2(" * 0$%*$"2(-,1 2'$0$!7 **-5(,& 12 !*$ 0$.$ 2 !*$
+$ 130$+$,21
#"#& " $"
,2$,1(27 ,-(1$ 2$"',(/3$1 2 )$ #4 ,2 &$ -% 2'$ !$ 2(,&
!$25$$, 4 0(-31 -.2(" * 1.$"20 * "-+.-,$,21 -% !0- #9
! ,#5(#2' 1.-,2 ,$-31 $+(11(-, 1-30"$ ,7 25- 1.$"20 *
*(,$1 5(** !$ 2 -0 +(6 2- "0$ 2$ , (,2$,1(27 %*3"23 2(-, 5(2'
%0$/3$,"7 $/3 * 2- 2'$ %0$/3$,"7 #(%%$0$,"$ !$25$$, 2'$
25- *(,$1 '(1 "-,"$.2 (1 (**3120 2$# (, (& (,"$ 2'$ -.2(9
" * ! ,#5(#2' -% 1.-,2 ,$-31 $+(11(-, 1-30"$1 " , $ 1(*7
$6"$$# 2'-31 ,#1 -% &(& '$028 2'$ (,2$,1(27 !$ 2 ,-(1$ 5(**
' 4$ 1(+(* 0 %0$/3$,"7 "-,2$,2 '$ %*3"23 2(-,1 (, -.2(" *
(,2$,1(27 0$ 0$%$00$# 2- 1 1.-,2 ,$-3191.-,2 ,$-31 -0
1.91. !$ 2 ,-(1$
'$0$ 0$ + ,7 1-30"$1 -% !0- #9! ,#5(#2' 1.-,2 ,$-31
$+(11(-, -2 130% "$1 13"' 1 23,&12$, *(&'2 !3*!1 " ,
.0-4(#$ -.2(" * 0 #( 2(-, 0 ,&(,& %0-+ 2'$ 4(1(!*$ 2- 2'$ % 0
(,%0 0$# $+("-,#3"2-0 1-30"$1 13"' 1 $#&$ $+(22(,& *(&'29
$+(22(,& #(-#$1 1 .0-4(#$ (,"0$ 1$# .-5$0 #$,1(2($1
-4$0 5 4$*$,&2' 0 ,&$ -% !-32 ,+ 2(** '(&'$0 .-5$0
#$,1(2($1 " , !$ -!2 (,$# %0-+ 1-*(#912 2$ 1-30"$1
* Erbium-doped fiber amplifier.
SP-SP Beat Noise
Power Spectral Density SE ()
.2(" * 2$"',-*-&7 5(** .* 7 , (+.-02 ,2 0-*$ (, !3(*#(,& 2'$
"-+(,& (,%-0+ 2(-, 13.$0'(&'5 7 2'0-3&' (21 " . "(27 2.0-4(#$ '(&' (,%-0+ 2(-, 2'0-3&'.32 ,# *-59*-11 20 ,1+(19
1(-, 1(+3*2 ,$-31*7 .2(" * 1-30"$1 13"' 1 1$+("-,#3"2-0
* 1$01 .0-4(#$ , -.2(" * " 00($0 5'-1$ (,2$,1(27 (1 +-#3* 2$#
5(2' (,%-0+ 2(-, 2- !$ 1$,2 -4$0 %(!$09-.2(" " !*$ -0 '(&'9
# 2 90 2$ "-++3,(" 2(-,1 * 1$01 27.(" **7 -.$0 2$ ,$ 0 2'$
9+ -0 9+ *-59*-11 5 4$*$,&2'1 (, -.2(" * %(!$0 .2(9
" * 0$"$(4$01 "-,4$02 2'$ (,%-0+ 2(-, +-#3* 2$# -,2- 2'$
-.2(" * " 00($0 2- ! 1$! ,# $*$"20(" * 1(&, *1 1 #$+ ,#1 0$
+ #$ %-0 +-0$ (,%-0+ 2(-, 2'0-3&'.32 2'$ ! ,#5(#2'1 -%
-.2(" * 0$"$(4$01 0$ (,"0$ 1$# "-++$,130 2$*7
Amplified
Spontaneous
Emission (ASE)
4000 GHz
o
Frequency .-,2 ,$-3191.-,2 ,$-31 1.91. !$ 2 ,-(1$ 0(1(,& %0-+
+(6(,& -% 2'$ 4 0(-31 1.$"20 * "-+.-,$,21 %0-+ 2'$0+ *9*()$ -.2(" *
,-(1$ 1-30"$
 Hewlett-Packard Company 1995
SE (n)
ASE
SI (f)
Si (f) [ȧR(f)ȧ2SI (f)
Bo
'+!# -&,1,"',"# !2//#+1 '0 -/,-,/1',+) 1, ,-1'!) '+1#+0'16
+" +,1 #)#!1/'! $'#)" *,/# /#)#3+1 .2+1'16 1, !,+0'"#/ '0
1&# -,4#/ 0-#!1/) "#+0'16 ,$ 1&# ,-1'!) '+1#+0'16 $
4&'!& '0 0!)#" 1, &3# 2+'10 ,$ 4110
7 &'0 .2+1'16 '0
/#)1#" 1, 1&# ,-1'!) $'#)" 0-#!1/2* 6 1&# /#)1'3#)6 0'*-)#
#5-/#00',+
SE(n) (W/Hz)
Power Spectral Density
(a)
[
Po
Bo
Bo
no
Frequency n
(b)
S I (f) (W 2/Hz)
Power Spectral Density
2
Po d (f)
P o2
[
Bo
SE(n)SE(n)
0
Bo
Frequency f
(c)
&# '+1#+0'16 +,'0# 1#!&+'.2# $,/ ,-1'!) /#!#'3#/ !)' /8
1',+ &# ,-1'!) $'#)" 0-#!1/2* 0 *#02/#" ,+ + ,-1'!) 0-#!8
1/2* +)67#/ ! &# ,-1'!) '+1#+0'16 0-#!1/2* 4&'!& '0 -/,-,/8
1',+) 1, 1&# -&,1,!2//#+1 0-#!1/2* 0 *#02/#" ,+ + #)#!1/'!)
0-#!1/2* +)67#/
02!& 0 $' #/8,-1'! *-)'$'#/0 0## -%# &# ')'16 1,
!,2-)# 1&#0# 0,2/!#0 ,$ /," +" )'%&1 #$$'!'#+1)6 '+1,
0'+%)#8*,"# ,-1'!) $' #/ '0 )0, '*-,/1+1 ,2-)#" -,4#/
"#+0'1'#0 !+ /+%# $/,* ,21 -+* $,/ )'%&1 2)
1, %/#1#/ 1&+ *+* $,/ *-)'$'#" 0-,+1+#,20 #*'08
0',+ $/,* $' #/8,-1'! *-)'$'#/ &# &'%& -,4#/ "#+8
0'1'#0 $/,* $' #/8,-1'! *-)'$'#/0 /# -/1'!2)/)6 4#))802'1#"
$,/ '+1#+0'16 +,'0# %#+#/1',+
'% '))201/1#0 &,4 1&# 01/#+%1& +" )'+# 0&-# ,$ 1&#
,-1'!) '+1#+0'16 #1 +,'0# /# %#+#/1#" $/,* 1&#/*)8)'(#
0,2/!# 02!& 0 *-)'$'#" 0-,+1+#,20 #*'00',+ '% 0&,40 1&# ,-1'!) -,4#/ $/,* + 0,2/!# ,$ 0-#!1/)
+"4'"1& , '+!'"#+1 ,+ 1&# &'%&8$/#.2#+!6 ,-1'!) /#8
!#'3#/ 1, # 1#01#" &# -,4#/ 0-#!1/) "#+0'16 ,$ 1&# ,-1'!)
$'#)" n 0!)#" 0, 1&1 '10 2+'10 /# 41107 '0 !,*8
*,+)6 20#" 1, !&/!1#/'7# 1&# ,21-21 ,$ ,-1'!) 0,2/!#0
&'0 .2+1'16 '0 16-'!))6 *#02/#" 20'+% + ,-1'!) 0-#!1/2*
+)67#/ '% 0&,40 16-'!) 0-#!1/2* $,/ n 4'1& $2))
4'"1& 1 &)$ *5'*2* +"4'"1& , +" !#+1#/ $/#.2#+!6
n, &# 3#/%# ,-1'!) -,4#/ , '0 $,2+" 6 '+1#%/1'+%
,3#/ 1&# ,-1'!) 0-#!1/2* 1&1 '0
, ŕ
4&'!& %'3#0 -#( 3)2# $,/ 1&# 0-#!1/) "#+0'16 ,$ --/,5'8
*1#)6 ,, 1 0&,2)" # +,1#" 1&1 )) -,4#/ 0-#!1/) "#+8
0'1'#0 "'0!200#" '+ 1&'0 --#/ /# 0'+%)#80'"#" 4'1& #+#/%6
,+)6 1 -,0'1'3# $/#.2#+!'#0
R
$ , δ$ n n
4&#/# δ$ /#-/#0#+10 "#)1 $2+!1',+ +" "#+,1#0 1&#
0'+%)#80'"#" 21,!,//#)1',+ '+1#%/1#" ,3#/ -,0'1'3# $/#8
.2#+!'#0 &'0 /#02)1 002*#0 2+-,)/'7#" +" 0'+%)#8
0'"#" -,4#/ 0-#!1/) "#+0'1'#0 &'0 #5-/#00',+ '0 3)'" $,/
,-1'!) /"'1',+ 4'1& 1&#/*)8)'(# +,'0# 011'01'!0 02!& 0
*-)'$'#" 0-,+1+#,20 #*'00',+ &# #$$#!10 ,$ 0&,1 +,'0#
/# +,1 '+!)2"#" '+ #.21',+ #!20# '1 '0 #0'#/ 1, !!,2+1
$,/ 1&'0 +,'0# $1#/ 1&# ,-1'!) '+1#+0'16 '0 !,+3#/1#" 1, -&,8
1,!2//#+1 &# /#02)1 ,$ #.21',+ '0 %/-&'!))6 '))201/1#" '+
'% ! +1#+0'16 #1 +,'0# #5'010 2- 1, $/#.2#+!'#0 "#1#/8
*'+#" 6 1&# 0-#!1/) 4'"1& ,$ 1&# 0,2/!# &#0# $/#8
.2#+!'#0 /# 16-'!))6 ,+ 1&# ,/"#/ ,$ 1&,20+"0 ,$ %'%&#/17
&# *%+'12"# ,$ 1&# #1 +,'0# 1 ),4 $/#.2#+!'#0 '0 -8
-/,5'*1#)6 ,ń , 1&# #5!1 3)2# "#-#+"'+% ,+ 1&# )'+#
0&-# ,$ 1&# #)#!1/'! $'#)" 0-#!1/2* ,/# !!2/1# 3)2#0 $,/
16-'!))6 #+!,2+1#/#" )'+# 0&-#0 /# %'3#+ )1#/
&# -,4#/ 0-#!1/) "#+0'16 ,$ 1&# "#1#!1#" -&,1,!2//#+1
'$ '+ 2+'10 ,$ *-#/#0
7 #.2)0 1&1 ,$ 1&# '+!,*'+%
'+1#+0'16 0-#!1/2* #5!#-1 $,/ 1&# $')1#/'+% #$$#!10 ,$ 1&# -&,8
1,"',"# $ 1&# $/#.2#+!6 "#-#+"#+1 /#0-,+0'3'16 '# 1/+08
$#/ $2+!1',+ $,/ 1&# -&,1,"',"# '0 %'3#+ 6 $ 4&'!& &0
2+'10 ,$ *-#/#0411 1&#+ 1&# -&,1,!2//#+1 0-#!1/2* !+
# #5-/#00#" 0
'$ ] $
$
&# /#!#'3#/ 1&#/*) +" 0&,1 +,'0#0 /# +,1 '+!)2"#" '+
#.21',+ &# 3)2# $,/ 1&#0# 14, +,'0# 0,2/!#0 4')) "#1#/8
*'+# 1&# 0'%+)81,8+,'0# /1', $,/ 1&# *#02/#*#+1
1#!&+'.2# 002*'+% 1&1 1&# ,-1'!) #1 +,'0# 0'%+) '0
)/%#/ 1&+ 1&# 0&,1 ,/ 1&#/*) +,'0# 1&'0 #5-/#00',+ !+ #
20#" 1, "#1#/*'+# 1&# *%+'12"# $ ,$ 1&# $/#.2#+!6 /#8
0-,+0# ,$ 1&# -&,1,"',"# &'0 *#02/#*#+1 '0 -#/$,/*#"
6 , 0#/3'+% 1&# -&,1,!2//#+1 0-#!1/2* '$ 4'1& + #)#!1/'8
!) 0-#!1/2* +)67#/ $ 1&# 0-#!1/) 4'"1& ,$ 1&# 0,2/!# , '0 *2!& )/%#/ 1&+ 1&# $/#.2#+!6 /#0-,+0# ,$
1&# -&,1,"',"# 1&#+ $ !+ # 002*#" !,+01+1 +" 1&#
/#0-,+0'3'16 0.2/#" $
,$ 1&# -&,1,"',"# '0 "'0-)6#"
"'/#!1)6 ,+ 1&# #)#!1/'!) 0-#!1/2* +)67#/
+# -/#3',20 "'$$'!2)16 '+ 1&# -/!1'!) 20# ,$ 1&'0 "#1#!1,/
!)' /1',+ 1#!&+'.2# '0 1&# 0*)) 3)2# $,/ 1&# '+1#+0'16 #1
+,'0# -/,3'"#" 6 16-'!) 0,2/!#0 &'0 &0 -/#0#+1#" -/, )#* $,/ &'%&8$/#.2#+!6 !)' /1',+ #!20# 1&# 1&#/*)
+,'0# )#3#) 00,!'1#" 4'1& 4'"# +" /#!#'3#/0 '0 16-'!))6
.2'1# )/%# + 1&'0 --#/ 4# 0&,4 1&1 6 !,* '+'+% 1&#
/#!#+1 "#3#),-*#+1 ,$ #/ '2*8",-#" $' #/ *-)'$'#/0 4'1&
0-#!1/) $')1#/'+% 1#!&+'.2#0 4# !+ ,3#/!,*# 1&# -/#3',20
)'*'11',+ ,$ 0*)) 0'%+) 01/#+%1&
n"n
 Hewlett-Packard Company 1995
# /2/6 #4)#118!(/" ,2/+)
To understand how to optimize the ASE for detector calibraĆ
tion, the concept of relative intensity noise or RIN, which
has units of Hz-1, will be introduced. This parameter can be
thought of as the fractional intensity noise associated with an
optical source. The definition for RIN in terms of the optical
intensity spectrum is:
RIN(f) = SI(f)/ P 2o.
–100
1 nm
RIN (dB/Hz)
(3)
Thus RIN is the spectral density of the optical intensity at a
given frequency divided by its integrated value at zero freĆ
quency. An equivalent definition equates RIN to the variance
of the optical intensity hDI2(f)i in a oneĆhertz bandwidth
divided by the average intensity squared. Since the optical
bandwidth Bo is usually much larger than the electrical
detection bandwidth, the RIN from an ASE source is usually
considered to be constant, equal to its lowĆfrequency value.
From Fig. 2c, it can be seen that this value is approximately
given by:
RIN ] 1/Bo.
Bo = 0.5 nm
–110
2 nm
5 nm
–120
–130
1
10
100
1000
Frequency (GHz)
(4)
Relative intensity noise (RIN) for a GaussianĆshaped optical
field power spectrum centered at 1.55 mm.
This result is valid for unpolarized light with thermalĆlike
noise statistics. A more accurate value for equation 4 reĆ
quires knowledge of the line shape of the ASE spectrum.
Normally, for a given source, one would increase the average
optical power incident onto a photoreceiver to increase the
photocurrent beat noise and hence the measurement sensiĆ
tivity. With the development of EDFAs (erbiumĆdoped fiber
amplifiers), the average optical powers attainable will easily
saturate highĆfrequency photoreceivers. This means increasĆ
ing optical power is no longer an issue, and for a given optiĆ
cal receiver, a power just below its saturation level should be
used for calibration purposes. Since incident optical power
can now be considered a constant, depending on the detecĆ
tor saturation level, the concept of RIN becomes useful beĆ
cause increasing its value increases the SNR of the intensity
noise technique. According to equation 4, this means that
decreasing the spectral width Bo of the ASE source will imĆ
prove the SNR for detector calibration. The only limitation is
that eventually there will be a rollĆoff in the highĆfrequency
content of the beat noise.
different optical bandwidths. The plotted curves correspond
to the case of an unpolarized ASE source with a Gaussian
optical spectral shape. The highest RIN is achieved at the
lowest frequencies with the narrowest optical bandwidth. At
frequencies comparable to the optical spectral bandwidth,
rollĆoff in the RIN becomes apparent. The RIN for the 5Ćnm
spectral width is relatively constant to several hundred gigaĆ
hertz. Expressions for RIN as a function of frequency are
shown in Table I for the case of Lorentzian, Gaussian, and
rectangular optical field spectrums. The Lorentzian shape
corresponds closely to the spectrum of commonly used FabryĆ
Perot optical filters. The Gaussian spectrum is sometimes
used to describe the spectra of EELEDs or lasers operating
below threshold. The rectangle function approximates some
interference filters, gratingĆbased monochromators, and
chirped fiber gratings. The normalized (to unity) optical field
spectrum is also included in Table I for reference. The line
width Bo in each case is the FWHM (full width at half maxiĆ
mum) of the optical field spectrum.
As the optical bandwidth is reduced, the maximum freĆ
quency separation of beating components also decreases. In
Fig. 3, the RIN is shown as a function of frequency for four
Comparing the lowĆfrequency RIN values for each of the
different spectral shapes, the rectangle spectrum delivers the
most RIN (1/Bo), followed by the Gaussian (0.66/Bo),
Table I
Relationship between Unpolarized Optical Field Spectrum and RIN
Optical Field Spectrum Shape
Normalized SE()
ǒn *B n Ǔ
Rectangle*
Gaussian
o
o
NJ
Lorentzian
o
o
1
1)4
= rectangle function.
ǒn *B n Ǔ Nj
2
exp *(4ln2)
*
RIN(f)
(fu0)
ǒn*n
Ǔ
B
o
2
1
Bo
NJ
ǒBf Ǔ
o
ǒ ǓNj
1 Ǹ2ln2 exp *(2ln2) f
B o Ǹp
Bo
1 1
Bo p
2
1
2
ǒ
1 ) fńB oǓ
o
= triangle function.
February 1995 HewlettĆPackard Journal
 Hewlett-Packard Company 1995
(m !(# !
The fiber-optic amplifier is a key element in modern high-data-rate communications.
Its large optical bandwidth, [4000 gigahertz, makes it effectively transparent to
data rate and format changes, allowing system upgrades without modifications to
the amplifier itself. The most prevalent fiber-optic amplifier is called the erbiumdoped fiber amplifier (EDFA). It provides amplification in the third telecommunications window centered at a wavelength of 1.55 m. The EDFA has four essential
components, as shown in Fig. 1. These are the laser diode pump, the wavelength
division multiplexer (WDM), the erbium-doped optical fiber, and the optical isolators.
To achieve optical amplification, it is necessary to excite the erbium ions situated
in the fiber core from their ground state to a higher-energy metastable state. A
diagram of the relevant erbium ion energy states is shown in Fig. 2. The erbium
ions are excited by coupling pump light ([20 mW or greater) through the WDM
into the erbium-doped fiber. Commonly used pump wavelengths are 980 nm and
1480 nm. The ions absorb the pump light and are excited to their metastable state.
Once the ions are in this state, they return to the ground state either by stimulated
emission or, after about 10 ms, through spontaneous emission. Light to be amplified
passes through the input isolator and WDM and arrives at the excited erbium ions
distributed along the optical fiber core. Stimulated emission occurs, resulting in
additional photons that are indistinguishable from the input photons. Thus, amplification is achieved. Optical isolators shield the amplifier from reflections that may
cause lasing or the generation of excess amplified spontaneous emissions (ASE).
Good EDFA design typically requires reduction of optical losses at the amplifier
input and minimizing optical reflections within the amplifier. EDFAs with greater
than 30-dB optical gain, more than 10-mW output power, and less than 5-dB noise
figure are readily achieved in practice.
4S
3/2
670 nm
4F
820 nm
4I
9/2
980 nm
4I
11/2
1480 nm
9/2
4I
13/2
10 ms
Metastable State
1550 nm
Ground
State
4I
15/2
Fig. 2. Relevant erbium ion energy levels.
ASE is generated in optical amplifiers when excited ions spontaneously decay to
the ground state. The spontaneously emitted photons, if guided by the optical fiber,
will subsequently be amplified (by the excited ions) as they propagate along the
fiber. This can result in substantial ASE powers (u10 mW) at the amplifier output.
A typical spectrum of signal and ASE at the amplifier output is shown in Fig. 3.
The ASE can extend over a broad spectral range, in this case in excess of 40 nm.
Output Spectrum
(5 dB/div)
ErbiumDoped
Fiber
WDM
Input
Signal
Amplified
Output
Fusion
Splice
Isolator
Fig. 1. Schematic of an erbium-doped fiber amplifier showing essential optical components.
The WDM is a wavelength division multiplexer.
".% .&84 #9 4)& /2&.4:*". / ' 4)& 3/52$& *3
0/,"2*:&% 4)& 7*,, #& 47*$& "3 ,"2(& '/2 ",, 4)2&& $"3&3
' 4)& /04*$", 2&$&*6&2 2&30/.3& -&"352&-&.43 "2& 0&2'/2-&%
*. 4)& ',"4 2&(*-& 3&& *( ./ 30&$*'*$ +./7,&%(& /'
4)& 3/52$& ,*.& 3)"0& *3 2&15*2&% /7&6&2 *' 4)& 30&$42",
,*.& 3)"0& '/2 4)& 3/52$& *3 "$$52"4&,9 $)"2"$4&2*:&%
"%%*4*/.", -&"352&-&.4 3&.3*4*6*49 $". #& /#4"*.&% #9 53*.(
."22/7&2 /04*$", '*,4&23 ".% $/22&$4*.( '/2 4)& 2/,,;/'' *. 4)&
2&30/.3& %"4"
&!# #"#' " $
" # %# $#
/ %&-/.342"4& 4)& '*,4&2&% *.4&.3*49 ./*3& 4&$).*15& -&";
352&-&.43 7&2& 0&2'/2-&% "4 " 7"6&,&.(4) /' - /. "
2&$&*6&2 7*4) ;: &,&$42*$", #".%7*%4) )& -&";
352&-&.4 3&450 *3 3)/7. *. *( )& -*22/2 "4 4)& &.% /'
4)& 2&35,43 *. 47/;0"33 (&.&2"4*/. )*3 %/5#,9
"-0,*'*&% *3 4)&. '*,4&2&% #9 ". /04*$", #".%0"33 '*,4&2
7*4) " ;.- 30&$42", 7*%4) ,, /04*$", $/..&$4*/.3 7&2&
'53*/.;30,*$&% 4/ &,*-*."4& /04*$", 2&',&$4*/.3 )& 02&3&.$&
 Hewlett-Packard Company 1995
Green
540 nm
Amplified
Signal
Laser Diode
Pump
Isolator
2H
11/2
ASE
1520.00
1572.00
Wavelength (nm)
Fig. 3. Spectrum of amplified spontaneous emissions (ASE) and amplified signal at the
amplifier output.
/' -5,4*0,& /04*$", 2&',&$4*/.3 $/5,% *-0"24 5.%&3*2"#,& *.;
4&.3*49 2*00,& /.4/ ". /4)&27*3& $/.34".4 30&$425-
7)*$) 7/5,% #& %*''*$5,4 4/ 3&0"2"4& '2/- 4)& 2&$&*6&2 '2&;
15&.$9 2&30/.3& )& #".%0"33 '*,4&2*.( 2&35,4&% *. "002/8*;
-"4&,9 " ;% *-02/6&-&.4 *. . "6&2"(& /04*$",
0/7&2 /' 7"3 *.$*%&.4 /.4/ 4)& 2&$&*6&2 4)*3 7"3
#&,/7 *43 3"452"4*/. ,&6&, /' 3*.( " -/2& $/-0,*;
$"4&% %5", %;! /04*$", )&4&2/%9.& 3934&- -&"352&;
-&.43 "4 " 7"6&,&.(4) /' - 7&2& 0&2'/2-&% '/2 $/-;
0"2*3/. 7*4) 4)& *.4&.3*49 ./*3& 4&$).*15& )& *.4&.3*49
./*3& 4&$).*15& 3)/73 &8$&,,&.4 "(2&&-&.4 7)&. $/-0"2&%
7*4) 4)& 34".%"2% )&4&2/%9.& -&4)/% "3 *.%*$"4&% *. *( !' #! #"#' " $
3 %*3$533&% &"2,*&2 '/2 " '*8&% "6&2"(& /04*$", 0/7&2 4)&
30/.4".&/53;30/.4".&/53 #&"4 ./*3& *. 4)& 0)/4/$522&.4
30&$425- *.$2&"3&3 "3 4)& /04*$", #".%7*%4) / *3 2&%5$&%
)*3 *.$2&"3& *. 3*(.", 342&.(4) *3 %&3*2"#,& #54 *' 4)& /04*$",
#".%7*%4) *3 2&%5$&% 4// -5$) 4)& )*();'2&15&.$9 $/.4&.4
/' 4)& #&"4 ./*3& 7*,, 34"24 4/ 2/,, /'' -"+*.( *4 5.35*4"#,&
&#25"29 &7,&44;"$+"2% /52.",
Laser Diode
Pump
Erbium-doped
Fiber
SE(n)
SI (f)
Wavelength
Division
Multiplexer
Mirror
Isolator
Bandpass
Filter
Bo = 1 nm
Optical
Receiver
Si (f)
HP 8565E
Electrical
Spectrum
Analyzer
9 kHz to 50 GHz
EDFA
Experimental arrangement for frequency response measurement of a SONET photoreceiver using the bandpassĆfiltered intensity
noise technique.
for highĆfrequency detector calibration. In practice, this
tradeĆoff between signal strength and frequency content beĆ
comes a problem when trying to characterize highĆfrequency,
lowĆgain optical receivers. Because of the large input noise
figures (typically 30 dB) associated with highĆfrequency elecĆ
trical spectrum analyzers, large values of photocurrent beat
noise are required.
The periodically filtered intensity noise technique solves this
tradeĆoff problem by allowing the magnitude of the beat
R(f) (linear)
Heterodyne 1.32 mm
0
0
1
2
3
Frequency (GHz)
(a)
R(f) (linear)
Intensity Noise 1.55 mm
0
1
2
3
Frequency (GHz)
Measurement of a SONET receiver frequency response using
(a) a 1.32Ćmm NdĆYAG optical heterodyne system and (b) the 1.55Ćmm
filtered intensity noise technique.
Details of the periodically filtered technique are shown in
Fig. 6. Fig. 6a shows the ASE from a source such as an
erbiumĆdoped fiber amplifier passing through a FabryĆPerot
filter and on to the test optical receiver. The transmission
characteristics of the FabryĆPerot filter are determined by its
free spectral range, FSR, and its finesse, F. FSR is the freĆ
quency separation of the transmission maxima and F is the
ratio of the FSR to the width of the transmission maxima.
From a signal processing point of view, the FabryĆPerot filter
has a frequencyĆdomain transfer function for the optical field
given by H(n). The squared magnitude of this transfer funcĆ
tion is illustrated in Fig. 6b. For the ideal case, the transmisĆ
sion through the filter would be unity at frequencies sepaĆ
rated by the FSR. After passing through the filter, the power
spectral density of the input optical field SE(n) will be transĆ
formed to H(n)2SE(n). Fig. 6c shows the filtered spectrum,
which is incident on the optical receiver. Strong beat signals
occur in the intensity separated by frequency intervals equal
to the FSR of the FabryĆPerot filter. As described earlier in
equation 1, the power spectral density for the optical intensity
SI(f) can be obtained from an autocorrelation of the input
electric field spectrum:
SI(f) ] P 2oδ(f) + H(n)2SE(n) H(n)2SE(n).
0
(b)
noise to be increased without loss of its highĆfrequency conĆ
tent.4 This is accomplished by passing the amplified spontaĆ
neous emission through a FabryĆPerot filter and then on to
the highĆfrequency optical receiver. This reduces the average
optical power while maintaining the magnitude of the sponĆ
taneousĆspontaneous beat noise at periodic frequencies
spaced at intervals equal to the free spectral range of the
filter. The result is an increase in RIN at periodically spaced
beat frequencies. If needed, optical amplification can then
be used to boost the average optical power to a value just
under the saturation level for the receiver.
February 1995 HewlettĆPackard Journal
(5)
This expression is valid for unpolarized amplified spontaĆ
neous emission (see equation 1). The result of equation 5 is
illustrated in Fig. 6d. Intensity noise peaks are evident at
frequency locations separated by the FSR of the FabryĆPerot
filter. Relative to the dc signal, these peaks are a factor of
F/p larger than the unfiltered case shown in Fig. 2c. For a
typical finesse of F = 100, this corresponds to an increase in
signalĆtoĆnoise ratio of about 15 dB for photodiode characterĆ
ization. The penalty for the increased SNR is that beat signals
only occur at specific frequencies. This constraint is not very
severe because this frequency spacing can be set to any
 Hewlett-Packard Company 1995
ȧH(n)ȧ2SE (n)
SE (n)
ASE
Bo
Fabry-Perot
Filter
SI (f)
Si (f) [ȧR(f)ȧ2SI (f)
H (n)
H(n ) 2
Filter Transfer Function
(a)
)& ] :&: & + :&:
FSR
1
H(n ) 2 S E(n)
Filtered Spectrum
[
P o2
Bo
Bo
no
Frequency n
Power Spectral Density S I (f)
(c)
2
Po d (f)
[
P o2 F
Bo p
H(n) 2SE(n) H(n) 2SE(n)
0
Bo
Frequency f
(d)
.
! +.#* $)!'1!, )++4231!3)-' 3(% /%1).$)#!++8 &)+3%1%$ )-3%-;
2)38 -.)2% #!+)"1!3).- 3%#(-)04% " 1!-2&%1 &4-#3).- .& 3(% !"18;
%1.3 &)+3%1 # /3)#!+ &)%+$ 2/%#314, )-#)$%-3 !3 3(% /(.3.1%#%)5%1
$ (.3.#411%-3 2/%#314, !2 ,%!241%$ .- !- %+%#31)#!+ 2/%#314,
!-!+89%1
WDM
0.98-mm
Pump
EDFA
p
ȍ
*
ǒ
Ǔ
) &**
Dn
.1 #.,/!1)2.- /41/.2%2 3(1%% $)&&%1%-3 3%#(-)04%2 6%1%
42%$ 3. ,%!241% 3(% &1%04%-#8 1%2/.-2% .& 3(% /(.3.1%#%)5%1
(% 1%24+32 .& 3(%2% 3(1%% ,%!241%,%-32 !1% 2(.6- )- )' (% #415% +!"%+%$ 6!2 ."3!)-%$ 42)-' 4-&)+3%1%$ .1
Optical
Receiver
Er3+
Er3+
Isolator
@
! " . $%,.-231!3% 3(% !".5% 1%24+3 3(% !11!-'%,%-3 )++4231!3%$
)- )' 6!2 42%$ (% ."3!)-%$ &1., 3(% 36.;/!22
24/%1&+4.1%2#%-3 2.41#% '%-%1!3%$ !- ./3)#!+ 2)'-!+
6)3( ! 2/%#31!+ 6)$3( .& !".43 -, #%-3%1%$ !3 m,
()2 3(%- /!22%$ 3(1.4'( !- ./3)#!+ )2.+!3.1 6()#( /1%;
5%-3%$ 3(% 36.;/!22 24/%1&+4.1%2#%-3 2.41#% &1., "%#.,)-'
! +!2%1 (% .43/43 .& 3(% )2.+!3.1 6!2 3(%- 2%-3 3(1.4'( !
2)-'+%;,.$% &)"%1 !"18;%1.3 &)+3%1 6)3( &)-%22% !-$
&1%% 2/%#31!+ 1!-'% 9 . "..23 3(% !5%1!'%
/.6%1 3(% &)+3%1 .43/43 6!2 ./3)#!++8 !,/+)&)%$ "8 ! &)-!+
/.23!,/+)&)%1 5%1!'% .43/43 /.6%12 .& 2%5%1!+ ,)++);
6!332 #!- "% ."3!)-%$ 42)-' 3()2 !11!-'%,%-3 42).- 2/+)#%2
"%36%%- 3(% 24/%1&+4.1%2#%-3 2.41#% 3(% &)+3%1 !-$ 3(% /.23;
!,/+)&)%1 6%1% 42%$ 3. ,)-),)9% 1%&+%#3).-2 ()2 )2 ),/.13!-3
"%#!42% ,4+3)/+% 1%&+%#3).-2 6)++ !$$ !,/+)34$% 1)//+% 3. 3(%
)-3%-2)38 /.6%1 2/%#314, (% /(.3.1%#%)5%1 #.-2)23%$ .& !
()'(;2/%%$ ;m,;$)!,%3%1 -!2 /;);- /(.3.$).$% &.+;
+.6%$ "8 ! 31!5%+)-' 6!5% !2 ,)#1.6!5% !,/+)&)%1 ;9 %+%#31)#!+ 2/%#314, !-!+89%1 $)2/+!8%$
3(% /(.3.#411%-3 /.6%1 2/%#314,
Periodically Filtered Intensity Noise Source
Mirror
.
6(%1% Dn )2 3(% 2/%#31!+ 6)$3( .& 3(% !"18;%1.3 &)+3%1 !-$ )2
')5%- "8 Dn ()2 1%24+3 )++4231!3%2 3(!3 3(% ,!'-)34$%
.& 3(% /(.3.#411%-3 "%!3 2)'-!+ #!- "% )-#1%!2%$ 1%+!3)5% 3. !
&)7%$ !5%1!'% #411%-3 6)3(.43 2!#1)&)#)-' )32 ()'(;&1%04%-#8
#.-3%-3
no
Frequency n
(b)
$%2)1%$ 5!+4% "8 /1./%1 #(.)#% .& 3(% &)+3%1 2 $%;
2#1)"%$ %!1+)%1 3(% ./3)#!+ 1%#%)5%1 &1%04%-#8 1%2/.-2% #!"% ."3!)-%$ 42)-' %04!3).- 6()#( 1%+!3%2 3(% ./3)#!+ )-;
3%-2)38 2/%#314, 3. 3(% /(.3.#411%-3 2/%#314, ,%!241%$
42)-' !- %+%#31)#!+ 2/%#314, !-!+89%1 224,)-' !- ./3)#!+
"!-$6)$3( . ,4#( +!1'%1 3(!- 3(% &1%04%-#8 1%2/.-2% .&
3(% ./3)#!+ 1%#%)5%1 3(% /(.3.#411%-3 /.6%1 2/%#314, )2
')5%- "8
Fiber FabryPerot Filter
WDM
Isolator
Isolator
0.98-mm
Pump
EDFA
WDM=Wavelength Division Multiplexer
HP 8565E
Electrical
Spectrum
Analyzer
9 kHz to 50 GHz
7/%1),%-3!+ 2%34/ 42%$ 3. $%,.-231!3% 3(% /%1).$)#!++8 &)+3%1%$ )-3%-2)38 -.)2% 3%#(-)04%
 Hewlett-Packard Company 1995
%"14!18 %6+%33;!#*!1$ .41-!+
Si (f) (10 dB/div)
B
C
A
(A) Unfiltered ASE
(B) Periodically Filtered ASE
(C) Heterodyne
5
10
15
20
25
30
Frequency (GHz)
curve B, the periodically filtered ASE output was used. The
average optical powers were held constant and equal for
curves A and B. Curve C was generated using an optical
heterodyne technique. Curves A and B experimentally demĆ
onstrate an SNR enhancement of approximately 17 dB beĆ
tween the filtered and unfiltered intensity noise techniques.
Comparison between the heterodyne technique and the two
intensity noise techniques shows very good agreement, illusĆ
trating the flat intensity noise spectrum obtained from the
EDFA noise source.
The use of a broadband intensity noise source for highĆ
frequency detector calibration has several advantages over
other frequencyĆdomain techniques. One important advantage
is that calibration of the frequency response of the optical
source is not necessary since it can be made flat by choosing
the ASE spectral width to be much larger than the frequency
response of the photodetector. This is not the case for other
frequencyĆdomain techniques. For example, sinusoidal intenĆ
sity modulation of an optical source using a LiNbO3 modulaĆ
tor requires careful calibration of the frequency response of
the modulator, which becomes increasingly difficult at high
frequencies. Any errors in this calibration are passed on to
the detector's frequency response.
Compared with heterodyne techniques, the intensity techĆ
nique is rugged and field deployable and does not require
stable polarization alignment. Since the intensity noise is
present at all frequencies, it allows rapid measurements.
Additionally, the long coherence length of laser sources in
the presence of optical reflections makes the heterodyne
measurement more susceptible to environmental effects.
The intensity noise method also has advantages compared to
timeĆdomain impulse measurement methods. For highĆfreĆ
quency measurements, the effects of the oscilloscope must be
deconvolved from the measurement before an accurate caliĆ
bration can be obtained. This can often be arduous since
accurate impulse responses for highĆfrequency oscilloscopes
are difficult to obtain. The advantage of not requiring careful
February 1995 HewlettĆPackard Journal
35
Frequency response meaĆ
surements of a highĆfrequency
optical receiver using three techĆ
niques. The intensity noise curves
were measured at 1.55 m and
the heterodyne measurement was
made at 1.32 m.
source calibration for the intensity noise technique is an imĆ
portant consideration in developing a portable, lowĆcost,
userĆfriendly calibration technique.
In this paper we have proposed and experimentally demonĆ
strated methods for enhancing the dynamic range of the
intensity noise technique for highĆfrequency photoreceiver
calibration. By combining recently developed EDFA technolĆ
ogy with spectral filtering techniques, we have shown that
the magnitude of intensity beat noise can be increased to
values practical for unamplified photodiode calibration.
Using the periodically filtered technique, RIN values larger
than -100 dB/Hz can be achieved over a frequency range in
excess of 100 GHz. The intensity noise calibration technique
has the potential for becoming a portable, easy to use, field
deployable calibration method.
Mike McClendon and Chris Madden performed the optical
heterodyne measurements and Mohammad Shakouri proĆ
vided the 30ĆGHz photoreceiver. Steve Newton provided
important support for this project.
1. D.J. McQuate, K.W. Chang, and C.J. Madden, Calibration of LightĆ
wave Detectors to 50 GHz," Vol. 44, no. 1,
February 1993, pp. 87Ć91.
2. S. Kawanishi, A. Takada, and M. Saruwatari, Wideband FrequencyĆ
Response Measurement of Optical Receivers Using Optical HeteroĆ
dyne Detection," Vol. 7, no. 1,
1989, pp. 92Ć98.
3. E. Eichen, J. Schlafer, W. Rideout, and J. McCabe, WideĆBandĆ
width Receiver/Photodetector Frequency Response Measurements
Using Amplified Spontaneous Emission from a Semiconductor OptiĆ
cal Amplifier," Vol. 8, no. 6, 1990,
pp. 912Ć916.
4. D.M. Baney, W.V. Sorin, and S.A. Newton, HighĆFrequency PhotoĆ
diode Characterization Using a Filtered Intensity Noise Technique,"
Vol. 6, no. 10, October 1994, pp.
1258Ć1260.
 Hewlett-Packard Company 1995
Download