Low-IMD Two-Tone Signal Generation for ADC Testing

Transcription

1 18 th International Mixed-Signals, Sensors, and Systems Test Workshop May 15 Taipei, Taiwan Low-IMD Two-Tone Signal Generation for ADC Testing K. Kato, F. Abe, K. Wakabayashi, T. Yamada, H. Kobayashi, O. Kobayashi, K. Niitsu Gunma University Semiconductor Technology Academic Research Center

2 Presented by Fumitaka ABE ( 安部文隆 ) 2

3 Outline Research Background Conventional Method Proposed Method Experimental Results Extension to ΔΣDAC Conclusion 3/50

4 Outline Research Background Conventional Method Proposed Method Experimental Results Extension to ΔΣDAC Conclusion 4/50

5 Research Goal Low distortion two-tone signal generation for communication application ADC testing with low cost AWG by only changing DSP program Conventional Proposed AWG ADC AWG ADC DUT DSP program change DUT IMD3 IMD3 IMD3 IMD3 Low SFDR High SFDR 5/50

6 Two-Tone Generation with AWG AWG Communication application DSP DAC Y ADC DUT Nonlinearity Low quality test Large IMD3 Narrow band IMD3 IMD3 IMD3 IMD3 6/50

7 IMD3 is important for two tone signal! Nonlinear system Two-tone signal Analog Filter IMD3 IMD3 IMD3 Cannot remove with analog filter 7/50

8 Outline Research Background Conventional Method Proposed Method Experimental Results Extension to ΔΣDAC Conclusion 8/50

9 Conventional Method DSP CLK CLK DAC Analog output 9/50

10 DAC Nonlinearity & IMD3 DSP CLK CLK DAC Analog output IMD3 components appear IMD3 IMD3 around 10/50

11 Outline Research Background Conventional Method Proposed Method Experimental Results Extension to ΔΣDAC Conclusion 11/50

12 Four Proposed Techniques Phase Switching Frequency Switching Phase Frequency Switching Pre-Distortion 12/50

13 Four Proposed Techniques Phase Switching Frequency Switching Phase Frequency Switching Pre-Distortion 13/50

14 Phase Switching DSP CLK CLK DAC Analog output 14/50

15 Phase Switching Effects DSP CLK CLK DAC Analog output spurious appear Cancel around 15/50

16 Principle of Phase Switching : phase Fundamental : phase Im Re Inter-Modulation phase IMD3 Im : phase Phase difference by π Re Cancel phase 16/50

17 Four Proposed Techniques Phase Switching Frequency Switching Phase Frequency Switching Pre-Distortion 17/50

18 Frequency Switching DSP CLK CLK DAC Analog output 18/50

19 Frequency Switching Effects DSP CLK CLK DAC Analog output spurious appear Cancel around 19/50

20 Principle of Frequency Switching DAC around Spurious components In principle, IMD components do not appear 20/50

21 Four Proposed Techniques Phase Switching Frequency Switching Phase Frequency Switching Pre-Distortion 21/50

22 Phase Frequency Switching DSP CLK CLK DAC Analog output 22/50

23 Phase Frequency Switching Effects DSP CLK CLK DAC Analog output around around spurious Cancel appear 23/50

24 Principle of Phase Frequency Switching DAC Frequency Switching Interleave 4 signals spurious spurious around around Phase Switching 24/50

25 Four Proposed Techniques Phase Switching Frequency Switching Phase Frequency Switching Pre-Distortion 25/50

26 Pre-Distortion DSP CLK CLK DAC Test Signal Add HD3 components in Din. 26/50

27 Effect of Pre-Distortion DSP CLK CLK DAC Test Signal HD3 territory Spurious around HD5 territory HD7 territory HD9 territory Cancellation 27/50

28 Principle of Pre-Distortion DAC Pre-Distortion Components Fundamental & Spurious 28/50

29 Principle of Pre-Distortion DAC Pre-Distortion Components IMD3 Cancel Fundamental & Spurious 29/50

30 Principle of Pre-Distortion DAC Pre-Distortion Components Fundamental & Spurious IMD3 components disappear 30/50

31 Proposed Low IMD3 Two-tone Generation Change DSP program Nonlinearity DSP DAC CLK CLK No hardware change No need for calibration No need for DAC nonlinearity identification 31/50

32 Simulation Conditions Conventional Test Signal 4 Proposed Test Signal 14bit DAC FFT Sampling Points Two-tone Signal Amplitude (peak-to-peak) 32/50

33 Power [db] Conventional Phase Switching Phase, Frequency, Phase Freq. Switching Around fs/2 Spurious Frequency Switching Phase Frequency Switching Around fs/4 Spurious Normalize Frequency f/fs Around fs/2 Spurious Around fs/2 Spurious 33/50

34 Power [db] Conventional Pre-Distortion Pre- Distortion Normalize Frequency f/fs Conventional Pre- Distortion /50

35 Output Power Spectrum Comparison Disappear Appear Conventional Phase Switching Around Frequency Switching Around Phase & Freq. Switching Around Pre-Distortion Around 35/50

36 Outline Research Background Conventional Method Proposed Method Experimental Results Extension to ΔΣDAC Conclusion 36/50

37 Experimental Conditions AWG Spectrum analyzer Signal Agilent 33220A ADVANTEST R3267 Frequency characteristic Amplitude Resolution Maximum Sampling Rate 1μHz~6MHz 14bits 50MSa/s Frequency band 100Hz~8GHz RBW 10Hz~30MHz RBW : Resolution Band Width Test Signal Two-tone Signal Sampling rate Input Voltage Offset 200kHz, 220kHz 10MSa/s 0.8~2.0Vpp (0.2V steps) 0 37/50

38 Power [dbm] Power [dbm] Fundamental Experimental Results Pre-Distortion -0.60dB Input Voltage [Vpp] Phae, Phae Freq. Switching -1.24dB Conventional Frequency Switching Conventional Phase Switching IMD3 Frequency Switching Phase Frequency Switching Average : -11.9dB Input Voltage [Vpp] Pre-Distortion 38/50

39 SFDR [dbc] SFDR Improvement Conventional Phase Switching Frequency Switching Input Voltage [Vpp] Phase & Frequency Switching Pre-Distortion Phase Switching Frequency Switching Phase Frequency Switching Pre-Distortion db db db db 39/50

40 Phase Switching Power [dbm] Conventional Method Phase Switching Around Fundamental HD3_3f1,3f2 Sampling Frequency_fs Frequency [khz] 40/50

41 Frequency Switching Power [dbm] Conventional Method Frequency Switching Around Fundamental HD3_3f1,3f2 Sampling Frequency_fs Frequency [khz] 41/50

42 Phase Frequency Switching Power [dbm] Conventional Method Phase Frequency Switching Around Fundamental HD3_3f1,3f2 Sampling Frequency_fs Frequency [khz] 42/50

43 Pre-Distortion Pre-Distortion Components Power [dbm] Conventional Method Around Fundamental Pre-Distortion HD3_3f1,3f2 Frequency [khz] 43/50

44 Outline Research Background Conventional Method Proposed Method Experimental Results Extension to ΔΣDAC Conclusion 44/50

45 Proposed Techniques using ΔΣDAC DSP ΔΣDAC Digital ΔΣ Modulation Nonlinearity DAC Conventional and Proposed digital input signal 45/50

46 Power [db] Power [db] Power [db] Power [db] 0 Phase Switching using ΔΣDAC Conventional Enlarge Normalized Frequency f/ fs 0 Phase Switching Enlarge Normalized Frequency f/ fs Normalized Frequency f/ fs Normalized Frequency f/ fs 46/50

47 Power [db] Power [db] Power [db] Power [db] 0 Frequency Switching using ΔΣDAC Conventional Enlarge Normalized Frequency f/ fs 0 Frequency Switching Enlarge Normalized Frequency f/ fs Normalized Frequency f/ fs Normalized Frequency f/ fs 47/50

48 Power [db] Power [db] Power [db] Power [db] Phase Frequency Switching using ΔΣDAC 0 Conventional Enlarge Normalized Frequency f/ fs 0 Phase Frequency Switching Enlarge Normalized Frequency f/ fs Normalized Frequency f/ fs Normalized Frequency f/ fs 48/50

49 Power [db] Power [db] Power [db] Power [db] 0 Pre-Distortion using ΔΣDAC Conventional Enlarge Normalized Frequency f/ fs 0 Pre-Distortion Enlarge Normalized Frequency f/ fs Normalized Frequency f/ fs Normalized Frequency f/ fs 49/50

50 Outline Research Background Conventional Method Proposed Method Experimental Results Extension to ΔΣDAC Conclusion 50/50

51 Conclusion Low IMD3 signal generation with low-cost AWG Only program change, No hardware change No need for calibration No need for AWG nonlinearity identification 4 proposed techniques cancel IMD3 Applicable to Nyquist-rate DAC and ΔΣDAC Low cost testing of communication application ADCs can be realized 51/50