Simulation of EMI in Hybrid Cabling for Combining Power and Control Signaling

Simulation of EMI in Hybrid Cabling for Combining Power and Control Signaling

Outline Motivation Hybrid cable design Cable parameters, impedance EMI and screening Connector modeling Cascaded analysis Transient co-simulation (ESD) Validation study Conclusions

SMPS/cable/sensor system Over-braid V+ IGBT block V- GN Screen Load RS485 RS485 Comm. Physical Layer Comm. Physical Layer Programmable Switched Mode Power Supply Hybrid Cable Hybrid connector Intelligent Sensor Need for EMI analysis

SMPS Emissions o PWM IGBT switching may be relatively low frequency o Fast edge rise time will generate harmonics that may cause EMI o RS 485 control signal may be corrupted by power switching noise Example τ r =1ns, τ=10us F 1 = 31.8 KHz, F 2 = 318 MHz A τ r 2Aτ/T 0 db/decade -20 db/decade A/2 τ -40 db/decade t F 1 =1/πτ F 2 =1/πτ r f

Simulation methods Static field solvers 3D full wave field solver 2D (+TL) cable solver 3D full wave field solver Circuit solver DC Parameters Impedance Equivalent Circuit S Parameters System Analysis

Cable Parameters (Capacitance) Pin Potentials Curved Tetrahedral Mesh Electrostatic Field Solution

Cable Parameters (Inductance) Current Paths Curved Tetrahedral Mesh Magnetostatic Field Solution

Cable Parameters (Impedance) Impedances from L/C values (ignoring losses) Z oo = (L 11 L 12 / C 11 + 2C 12 ) Z oe = (L 11 + L 12 / C 11 ) Z c = Z oe /2 Z d = 2 Z oo Z c = 44.5 Ohm Z d = 124.4 Ohms Electro-static and Magneto-static solvers directly calculate line parameters. Differential and common mode impedances can be calculated.

RS485 Impedance (Full-Wave FEM) Multi-pin differential port Curved Tetrahedral Mesh Port Mode Field Solution Port Impedance

Impedance Optimization RS485 insulator radii varied to optimize differential impedance. Initial design 127.6 Ohms Impedance tuned to 120 Ohms.

S Parameters and Cable Loss 20m length S parameters can be calculated to understand cable losses. Reference plane can be shifted to account for a longer length of cable. Avoids having to mesh a long cable.

Eye Diagram 10 Mbps Eye diagram for 10 Mbps; 100 ns pulse with 10 ns rise/fall time. RS485 protocol recommends maximum of approx 12 m for 10 Mbps.

Eye Diagram 100m cable Eye diagram for 10 Mbps; 100 ns pulse with 10ns rise/fall time. 100m long cable suffers from significant pulse distortion.

Cable Studio 2D (+TL) Model Aluminum Foil Copper Braid RS485 Twisted Pair (3 in. twist rate) 36 AWG conductors FEP insulation Power Wires 14 AWG PVC insulation Cable cross-section input using library parts and user-defined groups In-built calculators for foil and braid screens 2D Method of Moments solver used to calculate field in cross section Cable parameters automatically extracted from field solution TL modeling used to simulate a length of cable

EMI Crosstalk Analysis Transient task in Design Studio used for cable/circuit simulation. 20m length of cable modeled to analyze EMI crosstalk. SMPS transient voltage sources (650 V peak, 100µS period, 0.1µS edges). Coupling to RS485 line. Initial design without RS485 shielding.

EMI Crosstalk Analysis DM and CM Currents coupled into RS485 Switching voltage at SMPS 1 Amp common mode noise induced Will be filtered at receiver Induced differential mode must be kept below a few tens of mv

Impact of Twist Rate Initial study with RS485 shield removed. Induced differential current monitored at RS485 encoder. Twist rate of RS485 varied to see impact on EMI rejection.

Effect of Imbalance Imbalance introduced to investigate impact on EMI crosstalk. Circuit parameters asymmetric for RS485 line.

Impact of Imbalance Differential mode current significantly increases with imbalance. 2 ma peak differential mode current. (0.1 ma for balanced case)

Shield Transfer Impedance Z T is the intrinsic parameter of a cable shield, characterizes its shielding effectiveness (Schelkunoff, 1930s) Z T = (1/I O )*(dv/dx) where I O is current flowing on one side of the shield and dv/dx is the voltage per unit length on the other side Kley (1991) proposed a model for a braided shield, including various contributions Diffusion of E and H fields through the shield material Penetration of fields through the small apertures in the braid Induction phenomena due to overlapping of strand wires ( porpoising effect)

Kley s Model for Over-Braid Z T Braid porpoising aperture diffusion

Kley s Model for RS485 Foil Z T Foil

Cable Studio Screen Definitions

EMI Crosstalk Analysis (Screened) DM and CM Currents coupled into RS485 Switching voltage at SMPS 20 ma common mode noise induced Almost 1 A in unshielded case

Common Mode Current Waveform Common current pulse on edge of IGBT voltage. Ringing due to reflections and ring frequency related to length of cable. In air, pulse takes 0.13µS to travel 40m round trip.

Improving the Shield In some cases, multiple layers of shielding are required Foil layers, foil with braid, multiple braids, etc Braids are good barriers at low frequencies, poor barriers at high frequencies, whereas foils are the opposite (usually very thin)

Model for Combined Z T Vance s model (1972) for equivalent transfer impedance of two shields: Outer shield Z T1 = transfer impedance of outer shield Z T2 = transfer impedance of inner shield Z S1 = internal impedance of outer shield Z S2 = internal impedance of inner shield L 12 = inductance of the shield to shield line Inside cable where b 1 is the outside radius of the inner shield and a 2 is the inside radius of the outer shield

Predicted Foil/Braid Screen Z T Braid/Foil

Imported Transfer Impedance

Comparison of Results Combined foil/braid screen significantly reduces CM current.

Connector S parameters Multi-pin waveguide port. Single-ended. Distributed computing used to solve for various port excitations. Transmission losses relatively small due to short length

Cascaded System Simulation Cable Connector Cable and connector blocks cascaded for full system response. Transient task used to simulate crosstalk.

Common mode current Cascaded system increases peak CM current to 30 ma. Without connector 20 ma.

Transient Co-Simulation Transient Co-Simulation Hybrid field/cable/circuit solution Full bi-directional coupling between cable and environment Ideal for susceptibility and emissions analysis 2D (TL) cable solver Circuit solver 3D Transient (TLM) full wave field solver

Cable Susceptibility (ESD) ESD pulse 10m cable routed between SMPS and sensor. ESD event occurs at SMPS end. Human body contact model (HBM) used in the simulation.

ESD Transient Co-Simulation ESD current ESD Generator Circuit

Induced Currents Over-Braid Copper Braid

Induced Currents Power Wires Power Wires

Induced Currents RS485 Line Aluminum Foil RS485 Twisted Pair

Time Animation

Cable Studio Validation Study 4 different shielded cables RG58 (braid) RG6 (foil and braid) Twisted pair (spirally wrapped foil/drain) STP Shielded Twisted pair (braid) Test the transfer impedance modeling Excite source wire loop, measure signal received on signal wires with and without the shield Compare measured and simulated results Thanks to Jeffrey Viel at NTS for providing EMC test facilities and my colleague Patrick DeRoy for his work on the validation study.

Twisted pair with drain wire Drain wire in contact with foil Spiraling seam applied to foil wrap

Z T contributions Adding a spiraling seam to the foil increases the inductive component of the transfer impedance. The foil material is aluminum-polyester-aluminum. It is not unreasonable to assume that this spiraling seam would exist.

Cable Studio setup, TNC connectors Cable Studio model uses different routes and cable cross-sections to model the shielded section between boxes and unshielded sections inside the boxes.

Drain wire cable and pigtail connectors

Cable Studio setup, pigtail connectors Cable Studio model uses different routes and cable cross-sections to model the shielded section between boxes and unshielded sections inside the boxes.

Coupling results, measured vs. simulated Coupling / db Unshielded results Spiral/drain pigtail Spiral/drain TNC STP TNC RG58 TNC RG6 TNC Frequency / MHz Good correlation between simulation (solid) and measurements (dotted).

Summary Hybrid cable design using various simulation workflows Impedance calculation and optimization EMI crosstalk analyzed for different configurations Effect of twist rate, imbalance and screening investigated Transient field/cable/circuit co-simulation ESD susceptibility example Validation results for shielded cables