Network Analysis Seminar. Cables measurement

Transcription

1 Network Analysis Seminar Cables measurement

2 Agenda 1. Device Under Test: Cables & Connectors 2. Instrument for cables testing: Network Analyzer 3. Measurement: Frequency Domain 4. Measurement: Time Domain 5. Measurement: Enhanced Time Domain (TDR) 6. Labs 2

3 Agenda (Device Under Test) Specific Connectors (Fakra) Standard Connector (SMA) Coaxial Cable No Connector Balanced Cable 3

4 Agenda (Measurement modes) Frequency Domain Time Domain Enhanced Time Domain 4

5 Agenda (Measurement accuracy) No Correction Uncharacterized Adapter Port Extension Full2Port (SOLT or TRL) 5

6 1. Device Under Test: Cables & Connectors - Cable Equivalent Circuit - Coaxial Cables and Connectors - Balanced Cables and Connectors - Cable Parameters - Cable Measurement Issues 6

7 Cable is a Transmission Line Transmission lines are needed to convey RF and microwave energy from one point to another with minimal loss. The transmission line can be represented as an infinite series of twoport elementary components. Resistance, Inductance, Capacitance and Admittance define Characteristic Impedance Z 0. 7

8 Transmission Line Zo Zo determines relationship between voltage and current waves Zo is a function of physical dimensions and r Zo is usually a real impedance (e.g. 50 or 75 ohms) 8

9 a. Coaxial 138 Z0 log10 b a 2a 2b 1 velocity 1 for 0.65 v vacuum (air) for c factor teflon (PTFE) Frequency limit of coax High order modes may occur if: a b 9

10 RF Coaxial Cables RF Double Shield 50Ohm cable DC Resistance (shield) at 20 C Screening attenuation Capacitance (1kHz) Characteristic impedance < 21.5Ohm/km > 75dB <80pF/m (50+/-3)Ohm Velocity ratio 80% Attenuation at 20 C Return Loss at 1GHz Operating Temperature 20.5dB/100m at 100MHz 25.9dB/100m at 200MHz 53.5dB/100m at 1000MHz 76.7dB/100m at 2000MHz 98.1dB/100m at 3000MHz > 20dB -40 / +85 C 10

11 The Type N Connector DC to 18GHz N type 75ohm IS NOT compatible with N Type 50ohm 11

12 The Precision 3.5 mm Connector Air dielectric: is stable with temperature DC to 34GHz 12

13 The SMA Connector Usually teflon: this expands with temperature DC to 22GHz This pin is often the center wire of the semi-rigid cable 13

14 Standard Connector Summary Connector Metrology Instrument Production Cutoff Freq (GHz) Type F(75) N N Y 1 Y BNC (50 & 75) N N Y 2 Y SMC N Y N 7 Y Type-N (50 & 75) Y Y Y 18 Y APC-7 or 7 mm Y Y Y 18 N SMA (4.14mm) N N Y 22 Y 3.5 mm Y Y Y 34 Y mm or "K" N Y Y 44 Y 2.4 mm 2 Y Y Y 52 Y 1.85 mm 2, 3 N Y Y 70 Y 1.0 mm N Y Y 110 Y Sexed 1. Compatible with SMA and 3.5 mm connectors. 2. Not compatible with SMA, 3.5 or 2.92 mm connectors 3. Compatible with 2.4 mm connector Reference: Agilent Microwave Test Accessories Catalog, pp. 14,

15 The Automotive Fakra Connector DC to 6GHz FachKReis Automobil (Automobile Expert Group) 15

16 Connection Techniques Good connection techniques are required to produce reliable measurements. Before each connection: A. Inspect B. Clean C. Gage D. Use Torque Wrench PRACTICE MAKES PERFECT. 16

17 A. Inspect (and replace damaged connectors) 17

18 B. Clean WRONG Circular strokes leave torn fibers snagged on edges of center collet CORRECT Radial strokes do not leave fibers Use circular strokes for outer conductor face only 18

19 C. Gage (pin recession) Damage from Protrusion 19

20 D. Use the Torque Wrench Good connection techniques are required to produce reliable measurements. 20

21 b. Balanced Cables and Connectors A balanced line or balanced signal pair is a transmission line consisting of two conductors of the same type, each of which have equal impedances along their lengths and equal impedances to ground and to other circuits. The main advantage of the balanced line format is good rejection of external noise. Main disadvantage is the limited frequency coverage. 21

22 Balanced Cables with LVDS connector Double pairs cable Conductor resistance at 20 C Shielding effectiveness Capacitance (1kHz) Characteristic impedance Skew conductor-conductor (pair) Skew pair to pair Attenuation at 20 C Operating Temperature < 125Ohm/km > 55dB at 20MHz > 40dB at 1GHz < 50pF/m (100+/-6)Ohm < 25ps/m < 25ps/m 4dB/100m at 1MHz 10dB/100m at 10MHz 23dB/100m at 50MHz 33dB/100m at 100MHz 71dB/100m at 400MHz 81dB/100m at 500MHz 124dB/100m at 1000MHz -40 / +105 C Automotive version Low-Voltage Differential Signaling 22

23 Balanced Cables and Connectors: LAN Screened / Unshielded Twisted Pair Screened / Shielded Twisted Pair Local Area Network 23

24 Balanced Cables and Connectors: USB From USB 2.0 to USB 3.0: more complex connector, cable and test (e.g. Cross Talk) Universal Serial Bus 24

25 Cable Parameters The two principle factors which cause Attenuation are: Loss of conductors (caused by high frequency film effect), Dielectric loss. The Capacitance (pf/m at 1kHz) of a cable is indicated by the properties of the dielectric (the amount of electric charge when a potential difference exists between the two. Is directly proportional to the regularity of the dielectric's properties (typical values is 67 pf/m for PE). In the case of coaxial cables it is: Propagation speed is the speed of which an electrical signal travels along a line of Transmission. Is the ratio between speed of propagation within the cable and the speed in open space (66% for PE, Solid PolyEthylene dielectric). Characteristic Impedance Zo has to be as uniform as possible. The quality of the conductor and the geometry of the cable are not constant, causing signal distortion and loss. Screening attenuation depends on the external conductor's characteristics, which prevents the exchange of electromagnetic waves between the cable and the external environment. The Return Loss or Structural Return Loss (SRL is a specialized measurement of return loss referenced to the cable impedance) parameter is the measurement of the cable's production accuracy (mainly: constant dielectric extrusion pressure and cooling control. 25

26 Cable Parameters using VNA - Attenuation is S21 FD - Return Loss is S11 FD. SRL. Characteristic Impedance Zo - Screening attenuation (coaxial) is S21 FD - Cross Talk (balanced) is S21 FD - Length, Propagation speed is S11 TD - 1kHz requires LCR Meter] 26

27 Cable Measurement Issues - Frequency Broadband - Long electrical length (from swept to stepped sweep) - Reflection path loss (reflection dynamic range) - Non-Insertable (requires adapters) - Non-standard impedances (requires conversion) - Balanced (requires phy/sim BalUn transformer) 27

28 2. Instrument for cables testing: Network Analyzer - Block Diagram (sources, signal separation devices, receivers, analysis) - S parameters - Magnitude and Phase - Calibration (insertable, not-insertable, ) - Fixture Simulator function - Differential and Common Parameters (dd, dc, cc) 28

29 29 Network Analyzer Block Diagram

30 S parameters Completely characterize a two port device with four S-parameters S11 = forward reflection coefficient (input match) S22 = reverse reflection coefficient (output match) S21 = forward transmission coefficient (gain or loss) S12 = reverse transmission coefficient (isolation) Remember, S-parameters are inherently complex, linear quantities. However, we often express them in a log - magnitude format 30

31 > Transistor S parameters 1 GHz S S S S 12 Poor match: require a matching networks to couple signals into and out of the device. 31

32 > Device Characterization Incident (R) Reflected (A) REFLECTION DUT Transmitted (B) TRANSMISSION Reflected Incident = A R Transmitted Incident = B R VSWR (SWR) S-Parameters S 11, S 22 Reflection Coefficient (Linear or Polar) G, r Impedance (Smith) R+jX, G+jB Return Loss (LOG) Gain / Loss (LOG) S-Parameters S 21, S 12 Transmission Coefficient (Linear) T,t Insertion Phase (Phase) Group Delay (Delay) 32

33 Magnitude and Phase Signals are complex quantities Vector representation of a signal: Im A Phase Re RECTANGULAR format: A = Re + j Im POLAR format: A = Magnitude, Phase SMITH CHART format 33

34 Measurement Error Modeling Systematic errors due to imperfections in the analyzer and test setup assumed to be time invariant (predictable) Random errors vary with time in random fashion (unpredictable) main contributors: instrument noise, switch and connector repeatability Drift errors due to system performance changing after a calibration has been done primarily caused by temperature variation Measured Data Errors: SYSTEMATIC RANDOM DRIFT Unknown Device 34

35 Systematic Measurement Errors for Two Port Measurement R Directivity A Crosstalk B DUT Frequency response reflection tracking (A/R) transmission tracking (B/R) Source Mismatch Load Mismatch Six forward and six reverse error terms yields 12 error terms for two-port devices 35

36 Two-Port Error Correction a 1 b 1 E D E S Forward model Port 1 E X Port 2 S 21 A S 11 S A 22 A E TT E L a 2 b 2 a 1 b 1 E TT' E L' Reverse model Port 1 Port 2 S 11 A S 21 A S 12 A E RT' S 22 E S' E D' A b 2 a 2 E RT S 12 A E X' E D E S E RT E D' E S' E RT' = fwd directivity = fwd source match = fwd reflection tracking = rev directivity = rev source match = rev reflection tracking E L E TT E X E L' E TT' E X' = fwd load match = fwd transmission tracking = fwd isolation = rev load match = rev transmission tracking = rev isolation Each actual S-parameter is a function of all four measured S-parameters Analyzer must make forward and reverse sweep to update any one S-parameter Luckily, you don't need to know these equations to use network analyzers! S11m ED S11 a S m E D S E E RT E S E m E X S 1 22 ' 21 12m E X ' ( )( ' ) L ( )( ) RT ' E TT E TT ' S m E D' S E m E D S E S E RT E S E L E m E X S ' 21 12m E X ' ( )( ' ) ' L ( )( ) RT ' E TT E TT ' S m E X S21 a 21 S22m E D ' ( )( 1 ( E E TT E S ' E L )) RT ' S m E D S 1 11 E m E D E S 1 22 ' S ' ( )( E RT E S ' ) E L ' E ( 21 m E X S )( 12m E X ) RT ' L E TT E TT ' 12 S12 a S22a S E ' S E ( m X )( 1 11m D ( E ' )) E TT ' E S E L RT S ' ( m E D S ' E )( m E D S ' ) ' ( )( ) E S E RT E RT ' S E L E m E X S m E X L E TT E TT ' S22m ' ( E D S )( 11 m E D S ' E ) ' ( )( ) E RT ' E S E 21 m E X S 12 m E X 1 L RT E TT E TT ' S ( m E 1 11 D S E m E D ' S E S E RT E S E L E m E X S m E X ' )( 1 22 ' ) ' L ( 21 )( 12 ) RT ' E TT E TT ' 36

37 Return Loss (db) VSWR Before and After One-Port Calibration 0 20 Data Before Error Correction Data After Error Correction

38 Calibration method: THRU Response 38

39 Calibration method: Full-2-Port Forward and Reverse direction 39

40 Calibration Kit 40

41 1. Calibrating Insertable Devices ZERO-LENGTH THRU What is an insertable device? has same type of connector, but different sex on each port has same type of sexless connector on each port (e.g. APC-7) DUT When doing a through cal, normally test ports mate directly cables can be connected directly without an adapter result is a zero-length through >Cal >Meas DUT Zero-Length Thru Cal: OK 41

42 2 Calibrating Non-Insertable Devices NO ZERO-LENGTH THRU What is a non-insertable device? has same connectors on each port (type and sex) has different type of connector on each port (e.g., waveguide on one port, coaxial on the other) DUT DUT What calibration choices do I have for non-insertable devices? Problem is with THRU cal, where: use an uncharacterized through adapter > Cal Adapter > Meas DUT But Adapter length isn t known measurement error! 42

43 Adapter Considerations reflection from adapter leakage signal desired signal r measured = Directivity + adapter + r r DUT Coupler directivity = 40 db Adapter DUT Termination DUT has SMA (f) connectors Worst-case System Directivity Adapting from APC-7 to SMA (m) APC-7 calibration done here 28 db APC-7 to SMA (m) SWR: db APC-7 to N (f) + N (m) to SMA (m) SWR:1.05 SWR: db APC-7 to N (m) + N (f) to SMA (f) + SMA (m) to (m) SWR:1.05 SWR:1.25 SWR:

44 Solution is... Characterized Adapter Port 1 Port 2 DUT DUT cannot be replaced by a Zero Length Thru during Calibration Characterized Port 1 Port 2 Thru adapter 1. Transmission cal using Characterized Thru (Thru characterized and CalKit definition changed) Port 1 Cal Std. Cal Std. Port 2 2. Reflection cal using CalKit(s) standards Port 1 Port 2 DUT 3. Measure DUT 2-Port Error corrected with uncertainty due to the Thru Std. 44

45 Calibration Kit Solutions Coaxial (*): - APC7 Agilent - N (50/75ohm) Agilent mm (SMA) Agilent mm Agilent - 1mm Agilent - BNC Maury Microwave - Automotive Fakra Rosenberger Balanced (*): - USB3.0 BitifEye - LAN - Automotive LVDS (adapters: Rosenberger (*) Full-2-Port (complete ad accurate) calibration procedure at Cal Plane 2 (DUT Plane) is possible only with a Calibration Kit with same mechanical configuration of the Device Under Test (eg. Fakra CalKit if DUT has Fakra connectors). Otherwise use Calibration Kit suitable for Cal Plane 1 and try to compensate the Adapters (good quality, low reflection is required) contribution between Cal Plane 1 and Cal Plane 2: - using De-Embedding, - using Port Extension. 45

46 Measurement Uncertainty I m able to: - calibrate the NA; - measure the DUT; - but what is the uncertainty of the value? After two-port calibration Uncorrected Value and Uncertainty START MHz STOP MHz 46

47 Uncertainties Calculator: UncertTest.xls - The Uncertainty Calculator helps you determine measurement uncertainty due to your vector network analyzer and calibration kit. - Inaccuracies introduced through cable movement, connector repeatability and temperature drift are not included. 47

48 48 S21 uncertainty (Magnitude)

49 49 S21 uncertainty (Phase)

50 S11 uncertainty (Magnitude) Linear2Log 50

51 LAB: Uncertainty value CH1 S 21 log MAG 10 db/ REF 0 db u 1 CH1 S 11 log MAG 5 db/ REF 0 db Cor u 4 u 2 u 3 START MHz STOP MHz CENTER MHz SPAN MHz Value and Uncertainty 51

52 52 Fixture Simulator function

53 > (Automatic) Port Extensions Apply both electrical delay and insertion loss to enhance port extensions First approach to give reasonable alternative to building in-fixture calibration standards or de-embedding fixture Only fixture mismatch remains as source of error (dominated by coaxial connector). Open or short placed at end of each transmission line Coaxial calibration reference planes Ports extended Measurement accuracy is increased by minimizing the reflection of the transition by using good quality connectors and having good 50-ohm transmission lines on the test fixture 53

54 > De-Embedding Exclude undesired 2-port network from measured S-parameter Measured S-parameter Port 1 Undesired Network DUT Undesired Network Undesired Network Port 2 Port 3 De-embedded Response De-embedding ON/OFF is applied to all ports. Each port can be chosen as None or User. Undesired network is specified by Touchstone file (.s2p). 54

55 > Embedding (Port Matching) Include matching network of each port into measured S-parameter Port 1 Matching Network Measured S-parameter DUT Matching Network Matching Network Port 2 Port 3 Embedded Response Port matching ON/OFF is applied to all ports. Matching network is defined by each port independently. Matching network is specified by pre-defined circuit models or Touchstone file (.s2p). 55

56 > Matching Circuit : Single-Ended models 0 for Series C means no capacitor. Touchstone file (.s2p) can be defined for User. 56

57 > Matching Circuit : Differential Measured S-parameter Port 2 Port 1 DUT Matching Network Embedded Response Port 3 57

58 > Characteristic Impedance Conversion Convert S-parameter measured with 50 ohms to arbitrary port characteristic impedance DUT DUT 50W 50W XW YW Port 1: 50W Port 2: 50W Port 1: 100W Port 2: 50W Impedance Conversion ON/OFF is applied to all ports. Port impedance can be specified at each port. Example: SAW filter Port 1 50W --> 100W Port 2 50W S11 on Log Mag & Smith 58

59 > Unbalanced to Balanced Conversion Convert single-ended S-parameter to mixed-mode S-parameter YW DUT XW XW YW Define device type Single-ended & Balanced Balanced & Balanced Single-ended, Single-ended & Balanced Assign physical ports to logical ports BalUn ON/OFF is applied to each trace. DUT Y/2W 2YW 59

60 60 Single-ended to Mixed-mode conversion

61 Mixed-Mode S-Parameters - Sdd11 is the differential Attenuation Sdd21 Scd21 Sdc21 Scc21 - Sdc11 is the LCL (Longitudinal Conversion Loss) Sdc11 S S S S DD11 DD21 CD11 CD21 S S S S DD12 DD22 CD12 CD22 S S S S DC11 DC 21 CC11 CC21 S S S S DC12 DC 22 CC12 CC22 61

62 Why Differential and Common Parameters? Needs filter 62

63 3. Measurement: Frequency Domain - Measurement Technique - Insertion Loss, Attenuation and Phase matching - Return Loss and Impedance - Cross Talk, FEXT, NEXT - Screening Attenuation 63

64 64 Measurement Technique: Frequency Sweep

65 65 Coaxial Cable Measurement: - Return Loss, Zin - Insertion Loss, Attenuation and Phase Matching

66 Coaxial Cable Measurement: > Return Loss Failure in manufacturing process LogMag format SWR format 66

67 Coaxial Cable Measurement: - Screening Attenuation To maximize Dynamic Range: - Decrease IFBW - Increase Source Power [external Amplifier] - If possible use Receiver s direct inputs Ref. Standards, Design & Installation of CATV-Cables, Bernhard Mund, bedea 67

68 68 Balanced Cable Measurement: - Insertion Loss, LCL - Return Loss, Zin

69 69 Balanced Cable Measurement: - CrossTalk, FEXT, NEXT

70 4. Measurement: Time Domain - Measurement Technique - Resolution and Range - Delay, Length and Velocity Factor - Gating 70

71 Time Domain on a Network Analyzer 71

72 TDR vs Network Analyzer sources 72

73 Why Time Domain? Frequency Domain Time Domain 73

74 Time Domain procedure 74

75 75 (1) Frequency Data: Resolution and Range

76 (1) Effect of Frequency Span on Resolution For Example (Return Loss meas, so /2): - Frequency Span = 1GHz - k =0.45 (TD mode and Window) - Velocity Factor = 0.66 (PE) Resolution (s) = 0.5ns Resolution (m) = 0.05m 76

77 (1) What is the Maximum Range that can be measured? For Example (Return Loss meas, so /2): - Frequency Span = 1GHz - # of points = Velocity Factor = 0.66 (PE) Range (s) = 100ns Range (m) = 19.8m 77

78 (2) Bandpass Mode 78

79 (2) Low Pass Mode 79

80 (2) Low Pass Step vs Impulse Inductor 80

81 (2) Low Pass modes with Impedance changes 81

82 (2) Summary on modes 82

83 (3) Window Windowing improves the dynamic range of the time domain measurement by modifying (filtering) the frequency domain data prior to conversion to the time domain. 83

84 (3) Effect of Windows on Resolution: traces Response changes as window shape changes: (a) minimum window is best when higher resolution is needed to resolve signals with equivalent magnitudes, (b) maximum window is ideal for best dynamic range if responses are very different in magnitude. 84

85 (4) Gating Operation S11 S11 1 IFFT 2 Original Frequency Response Time Domain 4 S11 FFT 3 S11 Filter Out this Response Frequency Response w/ Gate Time Domain w/ Gate 85

86 (4) Gating Example: Time Domain response Connector Cable Connector Termination Remove the effects of the Input Connector S11 LIN 200 mu / REF -400 mu Gate On Cor START -1 ns STOP 9 ns 86

87 (5) Frequency Domain response with Gate On Connector Cable Connector Termination Gate Off S11 LIN 200 mu / REF -400 mu Gate On Cor START -1 ns STOP 9 ns Use Gating to show the frequency response of the output connector & termination only 87

88 88 > Coaxial cables with adapter and N connector

89 89 > Coaxial cable with adapter and Fakra connector

90 90 > Balanced cable with adapter and LVDS connector

91 91 Delay, Length and Velocity Factor

92 5. Measurement: Enhanced Time Domain (TDR) - Measurement Technique - Eye Diagram and Mask - Jitter, Emphasis, Equalization 92

93 93 Measurement Technique: Inverse Fourier Transform

94 94 Eye Diagram and Mask

95 95 Jitter, Emphasis, Equalization

96 6. Labs - Frequency, Time and Enhanced Time with Coaxial cables and adapters - Frequency, Time and Enhanced Time with Balanced line 96

97 97 LAB1 Frequency, Time and Enhanced Time (TDR) with Coaxial cable and adapters

98 98 LAB2 Frequency, Time and Enhanced Time (TDR) with Balanced line

99 Q & A THE END 99

100 Dielectric 100

101 Balanced Cables and Connectors: LAN Name Type Bandwidth Applications Notes Cat3 UTP [6] 16 MHz [6] 10BASE-T and 100BASE- T4 Ethernet [6] Described in EIA/TIA-568. Unsuitable for speeds above 16 Mbit/s. Now mainly for telephone cables [6] Cat4 UTP [6] 20 MHz [6] 16 Mbit/s [6] Token Ring Not commonly used [6] Cat5 UTP [6] 100 MHz [6] Cat5e UTP [6] 100 MHz [6] 100BASE-TX & 1000BASE- T Ethernet [6] 100BASE-TX & 1000BASE- T Ethernet [6] Cat6 UTP [6] 250 MHz [6] 10GBASE-T Ethernet Common in most current LANs [6] Enhanced Cat5. Same construction as Cat5, but with better testing standards. Most commonly installed cable in Finland according to the 2002 standard. SFS-EN Cat6a 500 MHz 10GBASE-T Ethernet ISO/IEC 11801:2002 Amendment 2. Class F S/FTP [6] 600 MHz [6] Class Fa 1000 MHz Telephone, CCTV, 1000BASE-TX in the same cable. 10GBASE-T Ethernet. Telephone, CATV, 1000BASE-TX in the same cable. 10GBASE-T Ethernet. Four pairs, S/FTP (shielded pairs, braid-screened cable). Development complete - ISO/IEC nd Ed. Four pairs, S/FTP (shielded pairs, braid-screened cable). Development complete - ISO/IEC nd Ed. Am

102 Why VNA has Wider Dynamic Range? 102