Receiver Signal to Noise Ratios for IPDA Lidars Using Sinewave and Pulsed Laser Modulation and Direct Detections Xiaoli Sun and James B. Abshire NASA Goddard Space Flight Center Solar System Division, Greenbelt, Maryland, USA The 16 th Coherent Laser Radar Conference Long Beach, California, USA June 20-24, 2011 Work Supported by: NASA ESTO IIP, Goddard IRAD & NASA ASCENDS Programs Contact: xiaoli.sun-1@nasa.gov 1
Context & Application: Lidar for NASA s planned ASCENDS Mission Active Sensing of CO2 Emission over Nights, Days, and Seasons Measuring the ratio of the reflected laser power/energy to the transmitted ones at on-line and off-line wavelength to determine total column absorption. ~450 km polar orbit Simultaneous measurements: CO2, O2, and col. height Target: ~ 1ppmv in ~100 km (10 sec) along track Objective of this work: - Achieving the required SNR with minimum average laser power (electrical power from the spacecraft) => Comparing signaling efficiency of coherent detection and two direct detection approaches: Sine-wave laser intensity modulation and lock-in detection Pulsed laser and box-car averaging detection 2
Laser Transmitters for Pulsed and Sine-wave Modulation trigger Laser seeder Off-line Laser seeder On-line Wavelength control Intensity modulator Pulses or Sinewaves Function generator Intensity modulator + Laser Amplifier Energy/power monitor Similar architecture for both pulsed and sine-wave approaches Pulsed modulation requires the laser amplifier to have a high peak power limit Similar architecture but no intensity modulation for coherent detection 3
Laser Transmitter Modulations 2-λ case (On-line & Off-line only) CW Lasers for Coherent Detection (no modulation) Two lasers at on- line and off- line wavelength simultaneously Sine-wave Laser Intensity Modulation for lock-in detection: On- line, 1+sin(ω on t) Off- line, 1+sin(ω off t) + Total, 2{1+cos[(ω on - ω off )t] cos[(ω on +ω off )t]} Laser peak power = 2 x average power Pulsed Laser Intensity Modulation: Laser peak power = 1/(pulse_dutycycle) x average power Range resolved measurement, capable of deteclng clouds 4
Coherent DetecLon Receiver: OpLcal signal Local oscillator laser Detector Receiver Block Diagram - for Coherent Detection Band pass filter on-line 90 0 Sin(w on t) X X Lowpass filter Lowpass filter Sqrt(X 2 +Y 2 ) Freq locking Laser seeder light from transmitter D/A off-line channel (same circuit but different intermediated frequency) Microprocessor The local oscillator laser must be coherent to the received laser signal in both longitudinal (frequency) and spatial modes at the detector. 5
Receiver Block Diagram - for Sine-wave Intensity Modulation and lock-in detection Lock- in DetecLon Receiver: OpLcal signal Detector Band pass filter on-line X X Lowpass filter Lowpass filter Sqrt(X 2 +Y 2 ) 90 0 Sin(w on t) D/A Microprocessor off-line channel (same circuit but different intermediated frequency) Direct detection of the optical signal followed by a lock-in type detection 6
Receiver Block Diagram - for Pulsed Modulation and Direct Detection Pulse DetecLon Receiver: OpLcal signal Detector Histogrammer Microprocessor 7
Signal and Noise Calculations Coherent Detection: (following derivation in Optical Communications by Gagliardi, Wiley 1976): At high local oscillator laser power, mean = stdev η c P LO 2 P LO + η c Background light int ensity W : Hz Noise bandwidth (Hz) : BW c = 1 2T int 2 P LO I bg BW c I bg = P bg Δλ Average signal power # wavelength samples (2) = hc /λ laser T int P LO Detector quantum efficiency Photon energy Received integra6on 6me Local oscillator laser power 8 η c Pbg Δλ T int Coherent heterodyne efficiency Background light power Receiver op6cal bandwidth Received integra6on 6me
Signal and Noise Calculations Sine-wave Modulation and Lock-In Detection Calculate mean signal amplitude in time domain Calculate noise in frequency domain, following derivation by Gagliardi 1976 Single photon detection, equal signal amplitude at on-line & off-line wavelength) mean = 1 2 sin 50% signal power in DC component due to intensity modulation stdev 2 ( + I sin bg Δλ) BW c 2 BW c = 1 2T int Noise stdev approximated to increase by sqrt(2) due to root mean square of the quadrature multiplication for the sine wave magnitude estimation. 9
Signal and Noise Calculations Pulse Detection: Single photon detection, equal signal amplitude at on-line and off-line wavelength mean = f pulses T int pulsed 1 f pulse stdev = f pulses T int pulsed 1 f pulse + I Δλτ bg pw pulses τ pw f Pulse rate, Pulse width 10
Signal to Noise Ratios Coherent Detection: SNR c mean stdev = η c 1+ T int η c 2 I bg η c n sig if I bg = 0 (The quantum limit) n sig = average number of detected signal photons over the integration time Sine-wave Intensive Modulation and Lock-in Detection: SNR at each wavelength sample point SNR sin = 1 2 2 sin T int ( + I bg Δλ) sin 1 2 2 n sig 11
Signal to Noise Ratios Pulsed Laser Modulation and Detection: SNR at each wavelength sample point SNR pulsed = pulsed T int pulsed +α dty I bg Δλ 1 1 T int n sig (The quantum limit) with α dty = τ pw f pulse << 1 the pulse duty cycle 12
Signal to Noise Ratio Comparison For Equal Powers: SNR at each wavelength sample point: pulsed = sin SNR pulsed SNR sin = 2 2 + η +α det dty I bg Δλ I bg Δλ 4 for = 2 (online &offline only) For Equal SNRs, Ratio of the required average signal powers: let I bg = 0, and SNR pulsed = SNR sin pulsed = sin 1 2 2 2 = 1 8 = 1 16 for ( = 2) 13
Laboratory Measurements Directly modulated laser diode as the laser transmitter Rectangular pulses at 10 khz, online and off line alternatively, equal amplitude Sum of 2 sincewaves, 50 KHz and 51 khz, equal amplitude Single photon counting PMT detector Sinewave signal through a 50±5kHz BPF to an ADC at 1Msamples/sec All subsequent signal processing in a PC 0.2 sec (5Hz) receiver integration time,100 trials for each measurement Background count rate: 0.23Mcts/s dark; 3.0 Mcts/s simulated daytime 14
Signal and Power Spectra of Sinewave Modulation On- line sine wave signal Power Spectra of the band- pass filter output Sync signal From the bottom up 1.System noise floor 2.Detector noise only (signal blocked) 3.Signal on Sum of on- line and off- line signal Same as above, expanded horizontal scale Sync signal 15
Results from Model and Measurements ( = 2) Lab measurement results agree with theory At a given detected signal photon rate (average power), SNR pulsed /SNR sin 4 At a given SNR, P pulsed / P sin ~ 16 Larger difference at low signal and high background noise levels 16
Summary Comparison of Coherent, Sine-wave and Pulsed Modulation Coherent Sine-wave modulation Pulsed modulation Peak to average power ratio P pk = P average P pk =2 P average P pk =(1/dutycycle) P average Wavelength sampling Simultaneous Simultaneous Sequential Primary advantage Ratio of SNR to quantum limit Background light sensitivity Clouds and aerosol detection Column height measurements Near quantum limited performance Low laser peak power mature technology High Laser signalling efficiency 1 1/(2sqrt(2N ) 1 Low High Medium No No Yes No No Yes 17