Photonic A/D conversions (Part II)
Optical folding-flash ADC
September 15, 1995 / Vol. 20, No. 18 / OPTICS LETTERS 1901
Optical folding-flash analog-to-digital
converter with analog encoding
B. Jalali and Y. M. Xie
1
photodiode transimpedance amplifier
\ r--,M,,,-i/
parallel
digital
word
/
comparator
Fig. 1. Direct conversion of an analog signal into gray
code through an array of MZ interferometers with binary-
scaled active lengths.
2
In a
MZ interferometer the output intensity I 0 varies witl1
r.::
0
analog voltage V a as
(1)
where I i is the input intensity and V 1T is the half-wave
voltage. To generate the least significant bit (LSB) i11
an N-bit quantizer, the V 1r of the modulator must scale
as
(2)
where Vrs is the full-scale input analog signal.
3
The
required geometrical scaling of V with resolution
1T
renders this technique impractical for ADC's of even
modest resolution. For example, a 4-bit ADC with
Vrs = 10 V will require V 1r = 630 mV, which is beyond
what can be achieved with state-of-the-art MZ interfer-
ometers based on either LiNbO 3 (Ref. 6) or electro-optic
polymers. 7
In this scheme
the most significant bit (MSB) is generated by a MZ
interferometer with V1r = Vrs/2 and biased at point B
on the transfer characteristics.
The second bit is ob-
tained by a MZ interferometer with the same active
length (same V 1r) but biased at point A. This ensures
two zero crossings within the full-scale range, as shown
by the output waveforms in Fig. 2(b).
4
(a) C
-v...
I ..
pulsed
MSB
laser
(b) ~ MSB-1
A
~~=r-~--~
MSB-2
B :;LSB
',! I I
I
i0 0 ,0 j
MSB-1:_;:,
I l
illlllliii.-...t==!====:::lt:::=:;:::::::ic:=--illlli
(c)
:o
MSB-2>;---.1'---t- --"l~--,f--~F--~ l - - - 'l -'---i
0 Va v"-
Fig. 2. (a) Transfer characteristics of the MZ interferome-
ter, (b) block diagram of the proposed ADC, (c) output show-
ing the digital gray-code representation of the analog input.
5
The phase of the folded waveform can be controlled
through the de bias point of the interferometer. For
example, when the interferometer is biased at point C',
its waveform is described by
l e, = Ii[1 - (1rVa)]
cos V 1r (3)
2
and is 180° out of phase with that at bias point
A. When the waveforms corresponding to bias points
A [Eq. (1)] and C' are multiplied in the voltage domain,
the resulting optical intensity has twice the folding
frequency:
(4)
6
Multiplication in the optical domain can be performed
by series connection of two interferometers, as shown
in Fig. 2(b).
-
The LSB has twice the folding frequency of the
MSB - 2 bit and is generated by multiplication of four
phase-shift waveforms corresponding to bias points A,
C', B, and B':
Ii [ 1 - cos
/LsB = IAic,IBIB, = 128 (41rVa)]
v 1T • (5)
This has the correct folding frequency and phase to
represent the LSB.
7
,.-.. MSB-2
::s••
,
•
.
Threshold , • .., , ~ / ,,,,- ..._ ,
'✓ MSB •
•I
... 1 •. .....
•
,., ·.
I
• •' ~
/ • •
/ \ '
/ I
MSB-1 ' •
,,
'
-5.3 0 5.3
Analog Input Voltage (V)
.
Fig. 3. Experimental data derived from two series-
connected LiNbO 3 MZ modulators.
8
11 68
Performance of a Time- and Wavelength-Interleaved
Photonic Sampler for Analog- Digital Conversion
T. R. lark, J. U. Kang and R. D. Esman nior M mb r, IEEE
I. t
t
WDM ...
RF In
memory
+
processing
t
... ... ...
t
Fig. 1. Photoni c ADC system.
9
~
1 •
,• .... ••
•
•
•• ''
••
'' -
G>
!l 0.8 '" I
I.,,,,•• h• • •
, '
I
0
ll.
-g
0.6 - ••• I
'••••• \-
''
..E
.!:! 0 .4 - ,•••
I ~
z
0
0.2 -,
, -'
0 '
1520 1540 1560 1580
Wavelength (nm)
(a)
0.06
0.05
-
> 0.04
-..
E
C
C>
0.03
0.02
i:ii
0.01
0
-5 0 5 10 15 20
Time (ns)
(b)
0.025 • • • •
~
>
0.02 - .
-..
E
0 .01 5 - .
C
C>
i:ii
0.01 - -
0 .005 - , ~ .'v
-
0
0 200
"' .V'
400
w
600
II'
800 "" 1000
Time (ps)
(c)
Fig. 2. (a) Spectrum of photonic sampler input (dashed line) and
eight-channel sampler output. Pulse train of sampler w ith (b) 320 MSPS
and (c) IO GSPS sampling rates.
10
IEEE PI IOTONICS TECHNOLOGY LETTERS, VOL. 12, NO. 9. SEPTEMBER 2000 1237
Phase-Encoded Optical Sampling for
Analog-to-Digital Converters
J. C. Twichell and R. He lkey
A high-speed optical sampling system for electrical
signals has been developed using a gain-switched diode laser and
a dual-output Mach-Zehnder interferometer.
The optical phase
shift between the branches of the interferometer is highly linear
in the applied electrical signal. The phase shift is encoded in the
two outputs of the interferometer and is recovered through digital
signal processing.
Analog-to-digital (A/D) conversion with 78-dB
spur-free dynamic range is demonstrated. the phase-encoded
sampling technique allows high-resolution (12-bit) conversion
with high linearity at practical laser power levels.
11
RF Input
Short-
Pulse Phase Shifter Digital Vour
Optical Signal
Processor 0010011 • • •
Source
Mach•Zehnder
Interferometer
Photodetectors
(a)
Optical Pulses
RF Input
-- -
--------- ,_ ----
(b)
Fig. 1. (a) Schematic illustration of analog-to-digital converter system that
uses optical sa1npling. The 1node-locked laser generates a train of short,
precisely timed optical pulses that enter the interferometer and sa1nple the RF
input. (b) Representative system wavefonns.
12
The outputs of the two legs of the interferometer
are both a sinusoidal function of the phase shift, but the
differential response of one output is inverted relative to the
other output. This dependence allows the phase shift 8¢ in an
ideal 1nodulator to be calculated fro1n the intensities of the two
output legs as follows:
(1)
The 111easure of the 8<j, (and therefore, Vin) is independent
of intensity noise in the laser pulses. To test this feature, 1nea-
surements were made in which the laser intensity was inten-
tionally modulated. The data demonstrated that the intensity
noise was suppressed by at least 50 dB .
13
IEEE PHOTONICS TECIINOLOGY LETTERS, VO L 12. NO. 12, DECEMBER 2000 1689
Photonic Analog-to-Digital Conversion by
Distributed Phase Modulation
M arc Currie, 1\llembel'; IEEE, Thomas R . C la rk, 1\llembe1; IEEE, and Paul J. Matthews. 1\llembe1; IEEE
EO •oo1il3 tor
Fig. 1. Conceptual drawing of a 3-bit pipeline photonic ADC syste1n based on
WDM and distributed optical phase 1nodulation.
14
-Vrc 0 +Vrc
0 1 2 3 4 5 6 7 Decimal
0 0 0 0 1 1 1 1
0 0 1 1 0 0 1 1 Binary
0 1 0 1 0 1 0 1
F ig. 2. Photonic transfer function maps the input signal voltage to digital
optical outputs in a 3-bit photonic ADC system. Binary values are achieved by
thresholding the data at mid-scale.
15
Polarization Photodetector ,
45° splice --
'·
Control - -- ·;),.._ -~
''
PM lso
Multi-A \ Phase Mod.
Source Phase Mod. PM lso
(a)
---- -- ---
In ut Sawtooth
---
... ... MSB
...
...
..
' ...
' ...
- - , ...
... LSB
' ...
... "
"'
'' .. ~--
-Vrr, 0 +Vrc
(b)
F ig. 3. (a) Experimental 2-bit fiber-based ADC syste m configuration. (b)
Output data w ith extrapolated values (dotted lines) exte nding to a driv ing
voltage of ± Vrr.
16
.lOUR AL OF U HT WAVE T CH OLOGY OL. 21, n 12 D EMBER 2 3 3 85
Photonic Tin1e-Stretched Analog-to-Digital
Converter: Fundan1ental Concepts and
Practical Considerations
Yan Han and Bahram Jalali Fi. !lo w, IEEE
Time
trctch DigiHier
M
hr >
- 2J,tF
:1· Ti me
(a)
T = llf~
Time Stretch
- Tim,
fl~
Segmec,.,{
I
=~1 "A-
"VV
4 t'\Jvv
4 ~ -
Di gitizer
-
x4 ,,-,.~
- /\ V
Digitiz er
Dlgltlw
Delay n~ I I\ I\ - I Di-gi-tiz- er
V \/"\.f
Clock f/4 - Vvtj
T
(b)
x4
4T
d . D is Fig. 2. Co nce pl1ial bl ck dia ram For th tim e-str etch ct DC. (a)
iti - r, d <t 1 f in gle-channel syst 111 For apturing a tim -limited si~nal. b) M ulli ll ann el
syst m For ca pturing a c ntinu ous-t itne signal. [n onlrnst to th e
lS 11 hn I sarnpl -in t rl a ed s, stem (Fig. I ), each hann 1 Ii.ere is sampled at or
above th e N quist rnt .
17
i \./'Jnput RF Signal Stretched RF Signal
SC
Dis persion ~ ]/1, 0
Time ~ -~
' " "' aw Dispersion
~
~J
~-~""---------.i Modulator ~ ~-"'""-- - PD
Source Fiber L1 Fiber L 2
b ~
\ Chirped Optical Pulse
l=1300nm ).. =149Dnm ). =1561nm
e
=! -15
a>
'g .20
-10
~xoooxix ""i i
{ -~5
-30
1500 1520 1540 1560 1580 1600 1620
Wavelenoth (nm)
Fig. 3. Functional block diagra m for the photonic time-str etch preprocessor. A linearly chirped optica l pulse is o btained by dispersing an ultra-shon
ultra-wide-ba nd pulse (the SC). Time-to-wave length mapp ing is achieved when this pul se is in tensity-modu lated by the electrical signal. The inset is the diagram
of a three-stage SC generation. SC: Supercontinuum. PD: Photodetector.
Supercontinuum generation: Supercontinuum generation is a process where light with very broad
spectral bandwidth (i.e., low temporal coherence) and usually at the same time high spatial coherence
is generated. This is usually obtained by propagating optical pulses through a strongly nonlinear
device as e.g. an optical fiber. Of special interest are photonic crystal fibers, mainly due to their
unusual dispersion characteristics, which can allow a strong nonlinear interaction over a significant
length of fiber. In some cases, tapered fibers can also be used.
18
100
80
§' 60
6
:;;
:,:
0
'2.
40
20
0
-60 -50 .40 -30 -20 -10 0 10 20 30 40 50 60
time (fs)
-;::.
I
t::: 0
E
CD
~
~-20
·.;
C:
Q)
-0
~-40
ti
Q)
'2.
"'Q) -60
:,:
0
'2.
-80
300 400 500 600 700 800 900 1000 1100 1200 1300 1400
wavelength (nm)
Supercontinuum, simulated with the program ProPulse by R. Paschotta, assuming that 20-fs laser pulses
propagate through a 2-mm long photonic crystal fiber. The time domain (upper graph) exhibits a complicated
multi-peak structure, while the spectrum (lower graph, with logarithmic scale) has a significant power spectral
density over more than one optical octave. The simulation took into account dispersion, the Kerr nonlinearity
(leading to self-phase modulation and four-wave mixing) with self-steepening, and Raman scattering.
19
Time Stretch
Swjtc n
Digitiz,er
Passive
In Fig. 19(a), the signal is split Sr:ilitwr
into channels and then gated ______f\41....___ Digitiz,er
prior to the time-stretch block
I.
in each path. Segmentation is '
achieved by triggering the :.: ), : Digiti~er
switches with a multiphase Time
clock. In Fig. 19(b), the Dig itizer
segmentation is performed in a ~ >
T
completely passive method (a)
using time-to-wavelength
Ti me Stretch
mapping inherent in a chirped
optical carrier. The RF signal An a log Input
Di,g itizer
is first modulated onto the
chirped optical carrier. A Dlgftlzer
passive optical filter then
performs wavelength
Digitiizer
segmentation which, owing to Ti me I Wavelength
the time-to-wavelength
mapping, corresponds to the Digitizer
desired time segmentation.
T Tim,e
(b)
Fig. 19. onceptnal block diagrams for hvo possible impleme11truim1s of the couti.tmous-time time-stretch AD . egmentahou is parfonned (a) u__si:ng a passi\•e
splitter and time gatiug and (b usi11g a passive optical filter.
20
