Common RF Test On ATE

Size: px

Start display at page:

Download "Common RF Test On ATE"

Sara Powell
5 years ago
Views:

4 Typical PLL tests Transceiver devices contain a PLL which is used to create a local l oscillator frequency. The LO frequency is used to either: Beat with incoming RF signal to extract the baseband signal Combine with baseband signal to produce modulated RF output Typical tests on the PLL include Charge Pump Current Prescaler Selectivity 2017/12/15 All Rights Reserved - ADVANTEST CORPORATION 4

5 What is a charge pump circuit? Phase Locked Loops are Voltage or Current controlled oscillators. The charge pump is the circuit which increases or decreases the voltage or current to the VCO. Prescale This causes the frequency of Counter the oscillator to increase or decrease in proportion to the voltage change. Typically the charge pump is controlled by a frequency or phase detector which produces a voltage change proportional p to the difference from a reference signal. Phase / Freq detector Feedback Divider Charge Pump VCO 2017/12/15 All Rights Reserved - ADVANTEST CORPORATION 5

6 Transceiver Block Set voltage to CP, and measure current LNA / Mixer Base band Gain / Filter PLL / Local Charge Oscillator Pump DC Transmitter / Baseband modulator Filter instruments Control DUT to switch mode and band Digital Control 2017/12/15 All Rights Reserved - ADVANTEST CORPORATION 6

7 Goal of Prescaler Sensitivity Test Ensure that the PLL is stable in all operating modes of the device. In each mode the frequency of the recovered output after the prescale divider ider should be constant Across all modes the frequency of the recovered signal should not vary by more than a certain amount 2017/12/15 All Rights Reserved - ADVANTEST CORPORATION 7

8 Internal PLL structure Digitizer Measure Prescaler_out RF Port Source LO IN Prescale Counter Phase / Freq detector Charge Pump VCO TX LO RX LO Feedback Divider 2017/12/15 All Rights Reserved - ADVANTEST CORPORATION 8

9 Test Technique PLL is programmed to given mode using digital control RF source is used to provide input frequency to PLL Digitizer is used to capture the output frequency divided by the prescaler Time domain averaging of capture used to reduce noise Windowed fft built-in function used to determine frequency of signal after prescaler Difference between the highest and lowest extremes tested 2017/12/15 All Rights Reserved - ADVANTEST CORPORATION 9

10 Time domain and frequency domain (phase noise performance) IDEAL SIGNAL REAL WORLD SIGNAL V(t) = A o sin 2π f ot V (t) = [A o + E(t)] sin [2 π f ot + φ(t)] E(t) φ(t) 2017/12/15 All Rights Reserved - ADVANTEST CORPORATION 10

11 RF transceiver block DPS module TX TEST DUT Analog AWG Diff. IQ input Diff. RF output RF measure 90 DPS module Digital module RF Pure clock RF source RF Input Diff. Diff. IQ output Analog DGT 90 RX TEST Digital module RF Pure clock DUT 2017/12/15 All Rights Reserved - ADVANTEST CORPORATION 11

12 Common Transmitter Tests DC Tests Current consumption Supplies and internal references RF Tests Measure Output Power TX -ACPR TX EVM TX Compression point 2017/12/15 All Rights Reserved - ADVANTEST CORPORATION 12

13 RFTX - Output Power Power Test Description RF measure RF Freq Baseband signal (modulated signal or single tone) is added to IQ pin of DUT. RF measure Analog AWG DUT The Output Power of RF signal is measured. I Q I Voltage Time Q Voltage RF : Radio Frequency CW: Continuous Wave Time 2017/12/15 All Rights Reserved - ADVANTEST CORPORATION 13

14 RFTX - ACPR Power Channel Power Adjacent Channel Test Description Baseband signal (modulated signal ) is added to IQ pin of DUT. The Adjacent Channel Power of RF signal is measured. RF measure RF measure Analog AWG RF DUT Freq I Q I Voltage Time Q Voltage ACPR : Adjacent Channel Power ACLR : Adjacent Channel Leakage Power Ratio Time 2017/12/15 All Rights Reserved - ADVANTEST CORPORATION 14

15 RFTX - EVM EVM: Error Vector Magnitude (I0, Q0) : Ideal Symbol (I, Q) : Measured Symbol Q I Q Magnitude error (I,Q) Measured symbol RF measure RF Ideal symbol measured symbol Error vector DUT Phase error (I 0,Q 0 ) Ideal Symbol RF measure φ Analog AWG I I Test Description Q I Voltage Baseband signal (modulated signal) is added to IQ pin of DUT. Q Voltage Time The error vector magnitude of RF signal is measured. Time 2017/12/15 All Rights Reserved - ADVANTEST CORPORATION 15

16 Transmitter Compression Goal of the test: To determine the amount of compression seen on the I and Q signals at the RF output when the level of modulating I and Q signals are increased The I and Q signal should not compress by an amount greater than specified 2017/12/15 All Rights Reserved - ADVANTEST CORPORATION 16

17 Gain Comperssion-P1dB RF Signal DPS Linear area Non-linear area Amplitud de Amplitu ude RF IN RF OUT Amplitu ude Amplitu ude Time Time Measurement Details Definition of Term In a low input level area (a small signal area), the output level for an amp increases at a specific slope as the input level increases (linear area). With further increases in the input level, the amp begins to show non-linear characteristics, and the actual output level comes to be lower than that expected from the gain in the linear area. There would not be any problem with this if only the output level were the decrease to less than the expected value, but the effect of this non-linear behavior appears as a distortion of the amplified signal. If this value exceeds P1dB, the gain will drop rapidly, and the output level will come to saturation. The output power for when gain decreases by 1 db as compared to the gain in a linear area where the signal input is low is defined as output power when gain is compressed to 1 db (P1 db). The unit of measurement is [dbm]. Time Time Purpose This measurement is taken in order to evaluate the usable input/output range (linear area) of an amp. P1dBout=P1dBinput + G 1dB 2017/12/15 All Rights Reserved - ADVANTEST CORPORATION 17

18 Gain Comperssion-P1dB Output Pow wer (dbm m) 1dB compression Sweep Input power and measure output power Plot Output Power Vs Input Power Find the point where actual Output Power is 1dB less than projected Output Power Input Power when this Output Power shrink by 1dB is P1dB Input Power (dbm) 2017/12/15 All Rights Reserved - ADVANTEST CORPORATION 18

20 Intercept Point Third Harmonics-IP3 Bm d RF Signal Δf RF IN DPS RF OUT dbm f 1 f 2 Frequency 2f 1 - f 2 f 1 f 2 2f 2 - f 1 Frequencyenc Third-order distortion Measurement Details Definition of Term Refers to two-signal third-order distortion. Input two signals of the same level (possibly with some Third-order distortion difference between the levels of the two) and with a frequency difference of Δf into the DUT, measure the third-order distortion that appears in a side band of the DUT output, and calculate the IP3 of the DUT output power (OIP3) and the IP3 of the DUT input power (IIP3) from the point of intersection in the figure shown on the right. The unit of measurement is [dbm]. Purpose Unlike a distant distortion as in harmonics, adjacent distortion appearing in modulated wave exerts undesirable effects on adjacent frequency bands. Therefore, it must be evaluated. Cause adjacent distortion to be generated in output by inputting two signals, and measure adjacent distortion characteristics of the DUT. OIP3=Pout + IM/2 2017/12/15 All Rights Reserved - ADVANTEST CORPORATION 20

21 Third Order Intercept - Derivation For any harmonic we have : HD n = ( n 1)(HIP n Po ) where n = harmonic number. Consider the second-order products resulting from two input signals having the same magnitudes. The frequencies are f 1 + f 2 and f 1 f 2. Both components are proportional p to V 1V 2 = V 12 = V 22. If both input signals are increased x dbs, then both secondorder products increase by 2x dbs. 2017/12/15 All Rights Reserved - ADVANTEST CORPORATION 21

22 Third Order Intercept Consider the third-order products resulting from two input signals having the same magnitudes. The components at frequencies 3f 1, 3f 2, and 3f 1 -f 2 are normally out of band and are not considered here. The components at frequencies 2f 1 - f 2 and 2f 2 - f 1 are normally of interest to the RF engineer. These components have magnitudes proportional p to V 12 V 2 and V 1V 22 respectively. If both signals are increased by x dbs, then both third-order components increase by 3x dbs. Thus we have: IMD 3 = 2( IP 3 P0 ) 2017/12/15 All Rights Reserved - ADVANTEST CORPORATION 22

23 RFTX - IP3 Power IM Test Description Baseband signal (two tones) is added to IQ pin of DUT. RF measure RF measure RF DUT (2f1-f2) f2) f1 f2 (2f2-f1) f1) Freq The 3 rd intermodulation distortion (IM3) which is generated by output RF signal is measured. The 3 rd order Intercept Point (IP3) is calculated by IM3. Analog AWG I Q OIP3 = Power(output) + IM/2 IM = Power(output) Power(IM3) I Q Voltage Voltage Time Time IP3 : 3 rd order Intercept Point IIP3/OIP3: Input IP3 / Output IP3 IM3: 3 rd order Intermodulation ti Distortion ti 2017/12/15 All Rights Reserved - ADVANTEST CORPORATION 23

24 Power Output Flatness Measurement of the output levels across the various modes or bands supported by the device. Since the previous test for carrier suppression, output level, and unwanted sideband suppression provide all the information. It is not necessary to repeat measurements. Simple calculations are used to determine the flatness of the output levels. Look for the difference between each measurement Determine if the extreme values exceed the specification for flatness 2017/12/15 All Rights Reserved - ADVANTEST CORPORATION 24

25 Typical Receiver Tests RX IP3 RX EVM RX Gain I/Q Phase and Gain AGC Tests Noise Figure Receiver Blockers Filter Bandwidth and Ripple Image Rejection 2017/12/15 All Rights Reserved - ADVANTEST CORPORATION 25

26 I/Q Phase and Gain Identify any mismatch in phase or gain of the I and Q demodulation. d I and Q signals should have the same amplitude I and Q signals should have the 90 degree phase shift preserved 2017/12/15 All Rights Reserved - ADVANTEST CORPORATION 26

Output Power RFRX - IP3 Power Δ f OIP3 Fundamental 3 rd order Intercept Point RF source Freq Test Description IIP3 3 rd order Intermodulation Distortion Input Power RF measure

The 3rd order intercept point (IP3) is calculated l by IM3.

27 Output Power RFRX - IP3 Power Δ f OIP3 Fundamental 3 rd order Intercept Point RF source Freq Test Description IIP3 3 rd order Intermodulation Distortion Input Power RF measure Analog DGT RF DUT RF signal (two tones) is added to RF input pin of DUT. The 3 rd intermodulation distortion (IM3) which is generated by analog IQ signal is measured. The 3rd order intercept point (IP3) is calculated l by IM3. IIP3 = Power(input) + IM/2 IM = Power(output) Power(IM3) IP3 : 3 rd order Intercept Point IIP3/OIP3: Input IP3 / Output IP3 IM3: 3 rd order Intermodulation Distortion I Q Power I DC (2f1-f2) Power Q DC (2f1-f2) f1 f1 f2 f2 Δ IM (2f2-f1) Freq Δ IM (2f2-f1) Freq 2017/12/15 All Rights Reserved - ADVANTEST CORPORATION 27

RFRX - EVM Q Magnitude error (I,Q) Measured symbol Power Modulated

Freq φ DUT Test Description I RF measure Analog DGT RF RF signal

The error vector magnitude of analog IQ signal is measured.

28 RFRX - EVM Q Magnitude error (I,Q) Measured symbol Power Modulated Signal Phase error Error vector (I 0,Q 0 ) Ideal Symbol RF source Freq φ DUT Test Description I RF measure Analog DGT RF RF signal (modulated signal) is added to RF input pin of DUT. The error vector magnitude of analog IQ signal is measured. I Q Q I Ideal symbol measured symbol 2017/12/15 All Rights Reserved - ADVANTEST CORPORATION 28

29 RFRX - Gain Power Test Description RF signal (single tone) is added to RF input pin of DUT and output IQ signal is measured. Gain is calculated l by RF input signal and analog output signal. RF source RF measure Analog DGT RF DUT Freq I Q Voltage I Voltage DC Freq Q DC Freq 2017/12/15 All Rights Reserved - ADVANTEST CORPORATION 29

30 Automatic Gain Control o (AGC) Tests Device gain should maintain baseband output level Device may have several gain levels l to be testedt Production testing may not require all ranges be tested Typical test: Low signal amplitude gain High signal amplitude attenuation May include AGC off test 2017/12/15 All Rights Reserved - ADVANTEST CORPORATION 30

31 Noise Figure Derived from Noise Factor: Defined as the ratio of the input signal-to-noise and the output signal-to-noise Noise Figure is Noise Factor in decibels Values for Noise Figure range from: 0 NF < Calculated by: NF = 10 log (F) Where F = (S in /N in )/ (S out /N out ) 2017/12/15 All Rights Reserved - ADVANTEST CORPORATION 31

32 Receiver Blocking Measurement of the ability of the receiver to capture the wanted signal in the presence of unwanted signals. The unwanted signals are those other than adjacent channel signals. Apply signal with a blocker as two-tone RF + blocker = tone1 RF + Baseband = tone 2 Apply signal without blocker as two-tone Measure the difference in the gain with and without blocker Subtract gain with blocker from gain without blocker 2017/12/15 All Rights Reserved - ADVANTEST CORPORATION 32

33 Filter Bandwidth and Ripple Measure the bandwidth of the receiver input Measure the power of a received signal in the center of the receiver s band. Measure the power of a signal of the same amplitude when the frequency is varied across the band of the receiver. Observe the difference in amplitude at specific frequencies and determine ripple. Observe the amplitude at bins outside the passband 2017/12/15 All Rights Reserved - ADVANTEST CORPORATION 33

34 Image Rejection Apply two inputs to the DUT Desired channel with Baseband signal Image offset from channel center frequency Device down-converts the RF to IF IF is down-converted to baseband Baseband spectrum is examined to see if Image channel present Rejection ratio = IF signal /IF image 2017/12/15 All Rights Reserved - ADVANTEST CORPORATION 34

35 Test Challenges Device high performance require high Spec for ATE instruments. - Bandwidth for modulation signals. LTE( 60MHz,120MHz) - Short settling time for DC, digitizer, RF instruments on ATE tester. - Better noise floor,better phase noise performance is needed. Low Cost - Sequence control to eliminate overhead. - In site parallel testing - Multi-site Interlacing - Smart-Calculation ( multi-thread handle data and calculation) 2017/12/15 All Rights Reserved - ADVANTEST CORPORATION 35

36 SmarTest 8 Instrument Instrument RF RF RF Meas Stim VNA Digitizer AWG DCVI diginout HW Cards WSRF WSMX DPS128 PS1600 A unified way to setup hardware resources. An abstract layer between SmarTest test programs and the hardware resources of V93000 system. 2017/11/ /12/15 All Rights Reserved - ADVANTEST CORPORATION 36

38 Instruments spec improvement benefit testing TX Test Results Summary WSRF/MX vs Previous Generation IQ Phase stability TX EVM & ACLR (Band 01, 2xCCA, 20MHz-20MHz), BW=120MHz TX EVM & ACLR (Band 39, 1xCCA, 20MHz), BW=60MHz CIM3 compared with LAB s MXA RF and Baseband IQ test Previous Generation Limited by digitizer's uncertainty 5ns Limited by instrument BW (~22MHz) 24 captures (6 captures/test mode x 4 modes) Limited by instrument BW (~22MHz) 12 captures (3 captures/test mode x 4 modes) Limited by digitizer noise floor Need to test RF and BB separately (resource limitation) WSRF/MX WSMX uncertainty is only 300ps Up to 200MHz BW 4 captures (1 captures/test mode x 4 modes) Up to 200MHz BW 4 captures (1 captures/test mode x 4 modes) Better noise floor Able to do RF and BB measurement in parallel 2017/12/15 All Rights Reserved - ADVANTEST CORPORATION 38

39 Operating Sequence An operating sequence is an arrangement of calls of patterns, actions, and transaction sequences to be executed. The arrangement can be serial, parallel, or a combination of both. sequential { actioncall DC; parallel { sequential {actioncall AWG; } sequential { transactioncall Condition1; parallel { sequential{actioncall meas1porta; } sequential{actioncall meas1portb; } } } } } DC Mixed PA RF RF Straightforward and descriptive All domains supported Precise synchronization 2017/12/15 All Rights Reserved - ADVANTEST CORPORATION 39

SmarTest 8 Key Features Operating Sequence Protocol DGT capture RF stim Easy to program with using Device Setup API In previous test program, there are 27 RX BASIC Test Suites (27 RX Path) In WSRF,

40 SmarTest 8 Key Features Operating Sequence Protocol DGT capture RF stim Easy to program with using Device Setup API In previous test program, there are 27 RX BASIC Test Suites (27 RX Path) In WSRF, we can test all 27 test suites in one test suite Multiple test suites can be combined with Operating Sequence to save test time Stop point can be set in the Operating Sequence during debugging Easy to debug with using Spectrum Analyzer or Oscilloscope No need to modify any codes 2017/12/15 All Rights Reserved - ADVANTEST CORPORATION 40

41 Site Interlacing In case of limit resources, Site Interlacing is automatically arranged Site Interlacing is based on DUT board description. PA PA PA PA RF power site 1&5 RF power RF power RF power site 2&6 site 3&7 RF power Independent resources site 4&8 RF power Shared resources 2017/12/15 All Rights Reserved - ADVANTEST CORPORATION 41

42 RF transceiver LTE CAT 10, traditional serial test flow (shared source architecture) BB measure IQ BB stim IQ BB measure IQ BB stim IQ Site 1 Site 8 A B C x M Y A B C x M Y RF stim RF measure RF stim RF measure A B C x M Y Test time 2017/12/15 All Rights Reserved - ADVANTEST CORPORATION 41

43 RF transceiver LTE CAT 10, choice for max parallelism (independent RF subsystem architecture) BB measure IQ A B C RF stim BB stim IQ Site 1 x M Y RF measure BB measure IQ A B C RF stim BB stim IQ Site 8 x M Y RF measure Parallel RF test flow Y M x C B Test time benefit by architecture t Serial RF test flow A B C x M Y Test time 2017/12/15 All Rights Reserved - ADVANTEST CORPORATION 42

Architectural differences LTE-A CAT 10 Transceiver 3 DL / 2 UL Shared stim source vs independent RF subsystem A B C RF architecture based on a fanout architecture

architecture A B M Y x C WSRF Y M x C B A FDD Test time benefit by architecture t Engineering efficiency Faster TTM SmarTest 8 and WSRF architecture Improved fault

44 Architectural differences LTE-A CAT 10 Transceiver 3 DL / 2 UL Shared stim source vs independent RF subsystem A B C RF architecture based on a fanout architecture with shared stim sources RF architecture based on independent RF subsystems and true parallel stim / measure ports M Y X misc f Traditional ATE shared source architecture A B M Y x C WSRF Y M x C B A FDD Test time benefit by architecture t Engineering efficiency Faster TTM SmarTest 8 and WSRF architecture Improved fault coverage Mission mode parallelism spurs, harmonics,... Test time benefits Higher parallelism shorter test time 2017/12/15 All Rights Reserved - ADVANTEST CORPORATION 43

Performance RF Performance: Frequency coverage: 10 MHz 6 GHz RX Bandwidth: 200 MHz, 350 MHz (undersampling) TX Modulated Bandwidth: 200MHz Dynamic range: 145 db

), debug, & reuse Operating sequence: actions (synced with digital, DPS, MX, RF) Unified tools, GUIs and APIs for digital, DC, analog, and RF Auto handling of

46 Performance RF Performance: Frequency coverage: 10 MHz 6 GHz RX Bandwidth: 200 MHz, 350 MHz (undersampling) TX Modulated Bandwidth: 200MHz Dynamic range: 145 db Settling Time Frequency change us Power change - 80 us SmarTest 8: TTQ and early competitive throughput Consistent, intuitive, test oriented ( instruments ), debug, & reuse Operating sequence: actions (synced with digital, DPS, MX, RF) Unified tools, GUIs and APIs for digital, DC, analog, and RF Auto handling of Upload + Background processing ac 160MHz signal 5.8GHz 1024 QAM 0.55% EVM Up to 200 MHz bandwidth 2017/12/15 All Rights Reserved - ADVANTEST CORPORATION 45

48 Architecture A Wave Scale MX card consists of 32 instruments contained in 16 HW Units High Speed Unit: High Speed AWG High Speed Digitizer High Resolution Unit: High Resolution AWG High Resolution Digitizer PMU Per Pogo Test Processor Controlled Synchronization Large Memory Pool Hardware Signal Processing Unit (SPU) 32 instruments per card 16 Units per Wave Scale MX card All can be used in parallel, independent, full pattern controlled 64 bidirectional analog pogos incl. PMU 2017/12/15 All Rights Reserved - Advantest Corporation 48

49 Wave Scale MX Key Values 4x density, 32 parallel instruments All 32 instruments are fully independent All functionality is sequencer controlled, no trigger pins, exact repeatablity Flexibility via licensing of units Dramatically simplified use-model (SmarTest 8) ) Overall improved performance Larger and flexible memory 216 Msamples shared btw. 4 instruments (Digitizer) All pogos can be connected to either AWG or Digitizer (bidirectional) Flexible I/O Matrix for simplified loadboard design PMU per pogo No dedicated analog calibration equipment needed Except in case of tight IQ requirements 2017/12/15 All Rights Reserved - ADVANTEST CORPORATION 48

50 Summary Wider bandwidth on WSRF (200MHz) Better & more user friendly software (SmarTest 8) Easier to burst test together (Operating Sequence) Operating sequence debug with stop point Multisite data handling Measurement stability Very good SDEV across all tests Good repeatability No averaging needed WSMX better noise floor Minimum effort on test time reduction (TTR) Outstanding COT 2017/12/15 All Rights Reserved - ADVANTEST CORPORATION 50

ADI 2006 RF Seminar. Chapter II RF/IF Components and Specifications for Receivers

ADI 2006 RF Seminar. Chapter II RF/IF Components and Specifications for Receivers ADI 2006 RF Seminar Chapter II RF/IF Components and Specifications for Receivers 1 RF/IF Components and Specifications for Receivers Fixed Gain and Variable Gain Amplifiers IQ Demodulators Analog-to-Digital