TDR, S-Parameters & Differential Measurements. TDR, S-Parameters & Differential Measurements

Transcription

1 DR, -Parameters & Differential Measurements Agenda Introduction DR theory of operation Real Example: amtec Golden tandard Board ingle Ended Impedance Coupling Effects Odd & Even Modes Differential Impedance Relating DR/D & -Parameters DR more than reflections 4-Port ingle Ended vs 2-Port Mixed Mode (Differential) Important Factors in DDR Accuracy Page 2 DR, -par, Differential Meas.

2 ignal Integrity Challenge Data Rates Increase >Gbps Risetimes become faster Reflections get larger Frequency Domain Data is Now Required Page 3 rend to Differential opologies Ideal differential devices Low voltage requirements Noise and EMI immunity Virtual grounding Non-ideal devices are not symmetric Can be identified by signalconversions Differential Common Common Differential Differential signal integrity design tools are needed Unintended signal conversion Page 4 DR, -par, Differential Meas. 2

3 What About Non-Ideal Devices? Undesirable signal conversions cause emission or susceptibility problems Differential-stimulus to common-response conversion + = EMI Generation Imperfectly matched lines mean the electromagnetic fields of the signals are not as well confined as they should be giving rise to generation of interference to neighboring circuits. Common-stimulus to differential-response conversion + = EMI usceptibility Imperfectly matched lines mean that interfering signals do not cancel out completely when subtraction occurs at the receiver. Measured by stimulating common-signal to simulate interference. Page 5 heory of Operation - DR Impedance DR Receiver E i E r DU tep Generator ime Domain Reflectometer (DR) Page 6 DR, -par, Differential Meas. 3

4 Impedance Mismatch erms Z L = Z 0 ρ = + ρ ρ Vr V i Impedance Calculated from ource Impedance and Reflection Coefficient. Reflection Coefficient, rho: How much was reflected? E E/2 What is the value of Ζ load? he DCA automatically calculates this for us. (ΔV) V r Ω Z L =Z 0 Ω 0 ime V i V r Zero Ω Page 7 he amtec Golden tandard Reference Board 3 A B 2 4 Details on Page 8 DR, -par, Differential Meas. 4

5 amtec Board Cross ection op Ground Plane Conductor A Conductor B op Ground Plane Dielectric Lower Ground Plane Page 9 Basic Microstrip Example As Positive Voltage tep is launched into conductor, positive charge is added to the conductor creating a current. Impedance is defined by the geometry and material properties independent of the voltage applied. Wider conductor or thinner dielectric increases capacitance, Z decreases impedance If Voltage is doubled, current is doubled, A X X X X X X X X X X X X v Z = i X = Current into page = Current out of page C Page 0 DR, -par, Differential Meas. 5

6 Coupling Effect on ingle Ended Impedance Adding other nearby conductors causes field lines to couple to the other conductor instead of the return path Capacitance to ground is reduced slightly ingle Ended Impedance of primary conductor is increased slightly Induced Voltage & Currents on adjacent conductor create Near and Far End Crosstalk, NEX/FEX A X X X X X X X X X X X B X X Page ingle Ended Impedance, Z E of A timulus on Port only 3 V=22mV Vr=8mV, V i =200mV, Z 0 =50Ω Z A E =54Ω Page 2 DR, -par, Differential Meas. 6

7 ingle Ended Impedance of B Exactly same as A Can be measured with either positive or negative voltage step at various amplitudes (linear) A B X X X X X X X X X X X X X Page 3 ingle Ended Impedance, Z E of B V=23mV timulus on Port 3 only Vr=7mV, V i=200mv, Z 0=50Ω 33 Z B E =54Ω Page 4 DR, -par, Differential Meas. 7

8 Vector Addition of Field Lines + - Page 5 imultaneous timulation ince the voltages/currents add/cancel, results can be calculated from independent measurements he amount of voltage that might couple onto a quiet net from an active net is completely independent of the voltage that might already be present on the quiet net Dr. Eric Bogatin ignal Integrity implified Prentice Hall A B X X X X X X X X X X X X X X X X X X X X X X X X X X Page 6 DR, -par, Differential Meas. 8

9 Differential Circuit erminology Odd Mode Impedance - Impedance of a single line, while the pair is driven in the odd mode (only differential signal, no common signal). Even Mode Impedance - Impedance of a single line, while the pair is driven in the even mode (only common signal, no differential). Differential Impedance - Impedance the differential part of a signal will see. (um of the Odd Mode Impedances) Common Impedance - Impedance the common part of a signal will see. (um of Even Mode Impedances/4) Page 7 Odd Mode Impedance Defined as Impedance of a single line when voltage on A is opposite polarity of voltage on B Can be measured either by stimulating both lines simultaneously OR combining currents/voltages from independent measurements Z odd is less than Z E due to induced currents combining (more current = less impedance) A B X X X X X X X X X X X X X X X X X X X X X X X X X X Page 8 DR, -par, Differential Meas. 9

10 Even Mode Impedance Defined as Impedance of a single line when voltage on A is identical to voltage on B Can be measured either by stimulating both lines simultaneously OR combining currents/voltages from independent measurements Z even is more than Z E since induced currents oppose each other (less current = more impedance) A X X X X X X X X X X X B X X X X X X X X X X X X X X X Page 9 Odd Mode Impedances ( 33-3 )= B odd V r =5mV, V i =-200mV, Z 0 =50Ω Z B Odd =43Ω - 3 = A odd V r =-4mV, V i =200mV, Z 0 =50Ω Z A Odd =43Ω Page 20 DR, -par, Differential Meas. 0

11 Even Mode Impedances = A even = B even V r =30mV, V i =200mV, Z 0 =50Ω Z A Even =ZB Even =67Ω Page 2 Differential Impedance Z Diff =Z A odd+z B odd 2 ( A odd +B odd )/2= CD A odd -B odd = DD V r =-29mV, V i =400mV, Z 0 =00Ω Z Diff =86Ω amtec simulated Z Diff =86.6Ω Page 22 DR, -par, Differential Meas.

12 Common Impedance Z Comm =(Z A even+z B even)/4 2 V r =30mV, V i =200mV, Z 0 =25Ω Z Comm =33Ω ( A even +B even )/2= CC A even -B even = DC Page 23 ry using your single-ended DR to make any two of the following measurements. From these measurements, you can work out the true differential impedance: he impedance of one trace while the other is reasonably well-terminated at both ends. he impedance when both traces are ganged together in parallel. he near-end crosstalk induced on the second trace when the first is driven. he necessary equations are pretty hairy, but you don't have to know them. Howard Johnson, PhD EDN Article, 8/22/ Page 24 DR, -par, Differential Meas. 2

13 Understanding -Parameters Optical Device (Lens) Electrical Device (PCB) incident signal ransmitted signal 2 DU 22 Reflected signal Reflection, 22 ransmission 2, response stimulus Page 25 heory of Operation DR/VNA Impedance -parameters Receiver DR E i E t E r DU ine Wave Reflection A DU B ransmission Reference R Receiver Magnitude & Phase tep Generator Network Analyzer ime Domain Reflectometer (DR) Vector Network Analyzer (VNA) Page 26 DR, -par, Differential Meas. 3

14 cattering Parameters & Impulse Response Frequency Domain (Phase not shown) ime Domain ( jω)* H ( jω) = R( jω) FF iff FF iff ( t ) H ( t ) = R ( t ) 2 (phase not shown) Multiplication in the Frequency Domain = Convolution in the ime Domain Impulse Response Page 27 tep Response (DR/D) vs Impulse Response ( )= t ( t ) H ( t ) = R ( t ) A perfect impulse is equivalent to a differentiated perfect step herefore if we differentiate the tep Response we get the Impulse Response ( )= t FF iff -Parameter! ( 2 ) Page 28 DR, -par, Differential Meas. 4

15 ingle-ended -Parameters and DR/D Port Port 3 Four-port single-ended device Port 2 Port 4 Return Loss or DR Insertion Loss or D Near End Crosstalk (NEX) Far End Crosstalk (FEX) Frequency Domain Parameters ime Domain Parameters FF or IFF Page 29 ingle-ended to Differential -Parameters Network Analyzers typically use this method ingle-ended Balanced Port Port 3 Port 2 Port 4 Balanced port Balanced port 2 timulus Differential timulus Common timulus Port Port 2 Port Port 2 Response E Differential Response Common Response Port Port 2 Port Port 2 DD DD2 CD CD2 Naming Convention: mode res., mode stim., port res., port stim. DD2 DD22 CD2 CD22 MM DC DC2 CC CC2 DC2 DC22 CC2 CC22 Page 30 DR, -par, Differential Meas. 5

16 ingle-ended to Differential DR/D Parameters Agilent DR uses this method Balanced ingle-ended Port Port 3 Port 2 Port 4 port port 2 timulus Differential timulus Common timulus Port Port 2 Port Port 2 Response Differential Response Common Response Port Port 2 Port Port 2 DD DD2 CD CD2 DD2 DD22 CD2 CD22 DC DC2 CC CC2 DC2 DC22 CC2 CC22 E Naming Convention: mode res., mode stim., port res., port stim. MM Page 3 ingle-ended to Differential DR/D Parameters [ ] = [ M ]*[ ]*[ M ] MM E timulus Differential timulus Common timulus Port Port 2 Port Port 2 Response Differential Response Common Response Port Port 2 Port Port 2 DD DD2 CD CD2 DD2 DD22 CD2 CD22 DC DC2 CC CC2 DC2 DC22 CC2 CC22 [ ] = [ M ] *[ ] [ M ] E MM * Conversion Matrix M = Page 32 DR, -par, Differential Meas. 6

17 he 4 Matrices of a 4-Port Device DD DD2 CD CD2 DD2 DD22 CD2 CD22 DC DC2 CC CC2 DC2 DC22 CC2 CC22 All 4 Matrices can be calculated from any of the 4 Matrices DD DD2 CD CD2 DD2 DD22 CD2 CD22 DC DC2 CC CC2 DC2 DC22 CC2 CC22 Page 33 DCAj -Parameters Calibrated -Parameters Real-time Update Built-in to the GUI Page 34 DR, -par, Differential Meas. 7

18 Correlation with Network Analyzer Blue = 20GHz PNA, Red = DCAj, Data compared with PL Page 35 Correlation with Network Analyzer Blue = 20GHz PNA, Red = DCAj, Data compared with PL Page 36 DR, -par, Differential Meas. 8

19 Differential Circuit esting What s Important? Differential teps must be well matched Differential Channels must be de-skewed - use the DR De-kew procedure prior to making measurements Page 37 Matching the teps When the DR outputs are terminated with 50 Ω standards, there is no coupling so the Differential Impedance should be 2 x 50 Ω = 00 Ω Mismatch or overshoot will affect the resolution of the Differential Impedance measurement DCAj has ~0 Ω (0%) variation ek CA has ~7 Ω (7%) variation Page 38 DR, -par, Differential Meas. 9

20 Agilent Calibration Agilent s built-in Calibration removes the effects of imperfect steps AND fixturing to give the ruest Differential Impedance. DCAj has < Ω (<%) variation Page 39 More Information DCAj - High Precision ime Domain Reflectometry (AN 304-7) Measuring Differential Impedance with DR to Improve High-peed Bus Designs, (AN 382-5) Improving DR/D Measurements Using Normalization (AN 304-5) PL - ignal Integrity Analysis eries Part : ingle-port DR, DR/D and 2-Port DR EN Part 2: 4-Port DR/VNA/PL EN Part 3: he ABCs of De-Embedding EN Page 40 DR, -par, Differential Meas. 20