Payload measurements with digital signals Markus Lörner, Product Management Signal Generation Dr. Susanne Hirschmann, Signal Processing Development
Agenda ı Why test with modulated signals? ı Test environment ı Discussion of measurement parameters to assess the signal quality EVM BER ı Overview on DVB-S2(X) signals ı Generate and measure DVB-S2(X) signals ı Live demo 16.03.2017 A&D Symposium 2
Physical layer test needs Function blocks which require RF testing: ı Digital satellite transponders ı Satellite components like PA, frequency converters, switching modules ı Ground modems ı Receivers ı LNBs, low noise block downconverters 16.03.2017 A&D Symposium 3
New approach in satellite design ı Complex digital signals with high bandwidths ı Pent pipe or complete digital system? ı Move from CW to more realistic signals ı High bandwidth simulation with MCCW as basic alternative ı More details with true digital signals to do EVM and BERT on true use case 16.03.2017 A&D Symposium 4
Comparison MCCW vs true modulated signal Original signal Multi-Carrier CW 8000 tones CF 8.9 db True modulated signal 16.03.2017 A&D Symposium 5
Comparison MCCW vs true modulated signal Gain compression through a 20 db PA Multi-Carrier CW 8000 tones CF 8.9 db True modulated signal 16.03.2017 A&D Symposium 6
Comparison MCCW vs true modulated signal AM/AM distortion through a 20 db PA Multi-Carrier CW 8000 tones CF 8.9 db True modulated signal 16.03.2017 A&D Symposium 7
Comparison MCCW vs true modulated signal AM/PM distortion through a 20 db PA Multi-Carrier CW 8000 tones CF 8.9 db True modulated signal 16.03.2017 A&D Symposium 8
Conclusion ı Different effect on MC CW than on true modulated signal ı More spot focused on MC CW Effects over level variation can not be seen ı Effect on EVM can not really be judged by MCCW tests Variation (=EVM) at points at different positions in constellation diagram differs New approach: ı VSA measures with the actual signal that is to be transmitted via the satellite transponder ı Analog effects (e.g. amplifiers distortions) on this concrete signal are evaluated ı Because it is the link quality that counts in the end 16.03.2017 A&D Symposium 9
Next generation solution 16.03.2017 A&D Symposium 10
Physical layer test approach Component or module test ı Same setup for PA or RF front end tests, just may differ in frequencies on both sides Signal Generator R&S SMW200A Signal Analyzer R&S FSW PA as DUT example Power supply P IN PA P OUT V CC 16.03.2017 A&D Symposium 11
Conversion to Baseband General concept of vector signal analysis ı Goal: assess quality of a measurement signal ı How: Compare against ideal signal ( reference signal ) Can be extracted by the measurement software or directly loaded into the software DUT Extract Reference Signal Measurement Signal Reference Signal Sync Compare and Extract Measurement Results VSA 16.03.2017 A&D Symposium 12
Synchronizing the measurement signal ı Estimation model: All residual distortions that are not explicitly compensated for stay in the measurement signal and can be seen in the EVM 16.03.2017 A&D Symposium 13
What is EVM? ı Error Vector Magnitude: Magnitude of error vector between a measured and a reference signal Measurement Signal Reference Signal Error Vector Meas Ref EVM 16.03.2017 A&D Symposium 14
What is EVM? ı Error Vector Magnitude: Magnitude of error vector between a measured and a reference signal Measurement Signal Reference Signal Error Vector Measurement Signal = ideal constellation point + error vector 16.03.2017 A&D Symposium 15
What can we learn from the EVM? ı EVM carries a lot of information on the type of distortion ı Three classical ways to look at the error vector: 1) as Magnitude (EVM) over Time 16.03.2017 A&D Symposium 16
What can we learn from the EVM? ı EVM carries a lot of information on the type of distortion ı Three classical ways to look at the error vector: 1) as Magnitude (EVM) over Time 2) Spectrum 16.03.2017 A&D Symposium 17
What can we learn from the EVM? ı EVM carries a lot of information on the type of distortion ı Three classical ways to look at the error vector: 1) as Magnitude (EVM) over Time 2) Spectrum 3) via the constellation diagram 16.03.2017 A&D Symposium 18
Analog effect: AWGN ı Gaussian distribution of constellation points ı Flat spectrum of the error vector ı EVM can be approximately estimated from the difference to the noise floor 16.03.2017 A&D Symposium 19
Analog effect: Non-linear distortion ı Can be identified from the constellation diagram and the comparison of the spectrums Ideal Points (=Ref signal) + Error vector 16.03.2017 A&D Symposium 20
Analog effect: Phase noise ı Can be identified from the constellation diagram the phase error trace the EVM trace ı Gets more and more important the higher the frequency is 16.03.2017 A&D Symposium 21
Analog effect: Non-flat channel Amplitude flatness ~ 0.3dB 16.03.2017 A&D Symposium 22
Why measure BER? ı EVM describes effect on signal ı BER should make sure data link works correctly ı Throughput counts in the end EVM BER Correlation? 16.03.2017 A&D Symposium 23
Correlation between BER and EVM EVM BER EVM BER Low EVM ensures low BER Zero BER is also possible with a high EVM Caution: EVM must be low enough for given modulation Switching to an APSK may lead to bit errors EVM is the more conservative measure BER check ensures that EVM is low enough for a specific modulation 16.03.2017 A&D Symposium 24
Bit Error Rate (BER) what to measure? Coded vs. uncoded Uncoded Bits Channel Encoder (e.g. LDPC) Coded Bits Modulation =? Coded BER =? Uncoded BER Channel Decoded Bits Channel Decoder (e.g. LDPC) Coded Bits Demodulation & Synchronization 16.03.2017 A&D Symposium 25
Bit Error Rate (BER) what to measure? Coded vs. uncoded ı Coded BER + what the end-user sees - very risky to rely on this measurement waterfall behavior of modern codes ı Uncoded BER can assess the link quality for a certain modulation in a given SNR region 16.03.2017 A&D Symposium 26
DVB-S2/DVB-S2X 16.03.2017 A&D Symposium 27
Why DVB-S2(X)? ı System is optimized for satellite to ground communication ı Variable bandwidth and throughput ı Adaptive coding and modulation from QPSK to 256 APSK ı Steep filtering, especially with DVB-S2X allows for small channel spacing 16.03.2017 A&D Symposium 28
DVB-S2(X) solution in the SMW200A GUI DVB-S2 ı Flow of the signal in the DVB-S2 standard ı Addressing the functions blocks in the SMW GUI 16.03.2017 A&D Symposium 31
Load / channel bonding and interference simulation ı Testing of components with multiple DVB-S2(X) channels ı Interference testing with different signals side by side ı In Multi Segment Waveform tool in SMW200A, different signals can be added together Combination of any signal, any bandwidth possible Frequency and level offsets Sum signal up to 2 GHz wide to create maximal load 16.03.2017 A&D Symposium 33
Measuring DVB-S2(X) signals ı Continuous signal where header and payload may have different modulation schemes /2-BPSK Header data modulation: e.g. QPSK, 8PSK,, 64APSK, 256APSK Payload SOF PLSCode information on e.g. data modulation fixed sequence to identify the start of frame Challenge: General VSA software supports only one modulation type 16.03.2017 A&D Symposium 36
Measurement demonstration. Signal Generator R&S SMW200A Signal Analyzer R&S FSW PA 16.03.2017 A&D Symposium 37