CAN Critical Parameters and Setup Considerations for Module Level EMC Testing Jim Lawlis Ford Motor Company Network Communications 1/19/2005
1/19/2005 Jim Lawlis - Ford Motor Company 2 Agenda CAN Protocol Critical Parameters CAN Bit Timing Load Box Termination, CAN Options Load Box Experiment Special Considerations for Optical Repeaters Diagnosing Issues CAN Emission Reduction
1/19/2005 Jim Lawlis - Ford Motor Company 3 CAN Protocol Critical for Communication Parameters Arbitration Oscillator ToleranceOscillator Tolerance All delays need to be counted twiceAll delays need to be counted twice Note: Choke Delays Count 4 Times Rx Tx Rx CAN H CAN L Transceiver ATransceiver B Choke AChoke B
1/19/2005 Jim Lawlis - Ford Motor Company 4 Propagation Delays Communication Media Number of NodesNumber of Nodes HS CAN – 16, MSCAN - 32 Capacitance on each NodeCapacitance on each Node 100pF to 130pF Wire LengthWire Length HS CAN 30 Meters MS CAN 50 Meters Bus SpeedBus Speed Twisted Pair 5.5nS/M Twisted pair With 16 Nodes 6.6nS/M Twisted Pair with 32 Nodes 8.0nS/M
1/19/2005 Jim Lawlis - Ford Motor Company 5 CAN Load Box Configuration CAN Bit Timing Settings Monitor Mode Load Box Termination Data Monitoring – Need to Check the Data, Not Just that Can Traffic is Present Monitor Error Frames/Bus off
1/19/2005 Jim Lawlis - Ford Motor Company 6 CAN Bit Timing Basics The CAN bit Measuring Total Delays J2284 equations The Bit Timing Spreadsheet
1/19/2005 Jim Lawlis - Ford Motor Company 7 The CAN Bit
1/19/2005 Jim Lawlis - Ford Motor Company 8 Measuring Total Delays Measure from the TX of one to the RX of an identical Module. Rx Tx Rx CAN H CAN L Transceiver ATransceiver B Choke AChoke B
1/19/2005 Jim Lawlis - Ford Motor Company 9 J2284 Equations Created by Karl Overberg of Ford and Dr. Dietmayer of Philips Bench Validated with Ford internally Designed Modules Vehicle Validated on a Fleet Consisting of Police Vehicles, High End Ford Vehicles, Heavy Truck
1/19/2005 Jim Lawlis - Ford Motor Company 10 The Bit Timing Spreadsheet
1/19/2005 Jim Lawlis - Ford Motor Company 11 CAN Monitor Only Load Boxes Note: CAN is an all Node ACK or NACK Protocol You may think you are only Monitoring. Unless the tool has a Monitor Only Option and it has been Enabled, the load box is Participating in Communication!!!
1/19/2005 Jim Lawlis - Ford Motor Company 12 Termination Methods Termination Methods Terminated NodesTerminated Nodes Standard Termination Split Termination Test Tool TerminationTest Tool Termination A/C Termination (ISO 15765, Part 4, section )
1/19/2005 Jim Lawlis - Ford Motor Company 13 Load Box Termination Terminated NodesNon-Terminated Nodes ~ 60 Ohm 1% 4.7nF
1/19/2005 Jim Lawlis - Ford Motor Company 14 Chamber Set-Up Tool is Configured Delays Have been identified/Checked How do We Connect the Load Box to the set up in the chamber?
1/19/2005 Jim Lawlis - Ford Motor Company 15 CAN Load Box Experiment Parameters That Need to be Defined Board to Case (A/C, D/C, Open)Board to Case (A/C, D/C, Open) Case to ground plane (On Ground Plane, 5 cm Above ground plane)Case to ground plane (On Ground Plane, 5 cm Above ground plane) Network Load box to system Load box power wiring (On Ground Plane, 5 cm Above Ground Plane)Network Load box to system Load box power wiring (On Ground Plane, 5 cm Above Ground Plane) A full Factorial Experiment was run because it provides the most information and 12 runs were possible
1/19/2005 Jim Lawlis - Ford Motor Company 16 Experiment Setup Y Performance = X1 X2 X3 X1 = Routing (5cm or Ground Plane) X2 = Case to Table (5cm or Ground Plane) X3 = Board to Case (AC, DC, or Open) DUT Load Box Metal Case X2X2 Ground Plane CAN Tool X3X3 Power Ground CAN_H CAN_L Make these lines as short as possible X1X1
1/19/2005 Jim Lawlis - Ford Motor Company 17 Experiment Analysis Previous Experience determined the worst case Load Box Setup noise was observed in the Vertical BICON 30 to 200Mhz Area of violations above the limit line were measured and calculated for each test The Best Performance was Identified
1/19/2005 Jim Lawlis - Ford Motor Company 18 DOE Test Matrix Run OrderPower RoutingCase to TableBoard to Case 1GND A/C 25cmGNDA/C 3GND5cmA/C 45cm A/C 5GND D/C 65cmGNDD/C 7GND5cmD/C 85cm D/C 9GND Open 105cmGNDOpen 11GND5cmOpen 125cm Open
1/19/2005 Jim Lawlis - Ford Motor Company 19 Experiment Data
1/19/2005 Jim Lawlis - Ford Motor Company 20 Data Processing Normality of the Data and Residuals Minitab DOE Analysis Multi-Variable Chart Interaction Plot Main Effects Plot
1/19/2005 Jim Lawlis - Ford Motor Company 21 Normality of Data
1/19/2005 Jim Lawlis - Ford Motor Company 22 Normality of Residuals
1/19/2005 Jim Lawlis - Ford Motor Company 23 Minitab DOE Analysis General Linear Model: Data 2 versus Power Routin, Case to Tabl,... Factor Type Levels Values Power Ro fixed 2 GND 5cm Case to fixed 2 GND 5cm Board to fixed 3 AC DC Open Analysis of Variance for Data 2, using Adjusted SS for Tests Source DF Seq SS Adj SS Adj MS F P Power Ro Case to Board to Power Ro*Case to Power Ro*Board to Case to*Board to Error Total
1/19/2005 Jim Lawlis - Ford Motor Company 24 Multi-Variable Chart
1/19/2005 Jim Lawlis - Ford Motor Company 25 Interactions Plot
1/19/2005 Jim Lawlis - Ford Motor Company 26 Main Effects Plot
1/19/2005 Jim Lawlis - Ford Motor Company 27 Experiment Results DUT Load Box Metal Case GND Ground Plane See EMC ES-XW7T-1A278-A(X) (X): Recent Version CAN Tool Board to Case DC GND Power Ground CAN_H CAN_L Make these lines as short as possible
1/19/2005 Jim Lawlis - Ford Motor Company 28 Monitoring CAN through Optical Repeaters The Signal Delay of a Typical Optical Repeater (wire to optical to wire) will be ~300ns This does not include the signal delay through the Optical Cable ~5.0ns/meter.
1/19/2005 Jim Lawlis - Ford Motor Company 29 Network Length Penalty for Optical Repeaters Assume Tool uses a Crystal Fast Tool Transceiver (Combined RX and TX delays less than 255ns vs. 330ns) Communication and Arbitration can be Guaranteed through 24 meters Tool to Furthest Module
1/19/2005 Jim Lawlis - Ford Motor Company 30 Diagnosing Issues No Substitute for an Oscilloscope Just turning off CAN communication will not tell the story. You need to rerun with the TX pin shorted to the dominant state. This shows the noise on the 5 Volt feed as well as on the ground of the node. CAN tools need to make assumptions Advantages of monitoring the Common Mode and Differential Mode signals
1/19/2005 Jim Lawlis - Ford Motor Company 31 CAN Radiated Emission Reduction Methods Overall Goal Communication Bus Noise will always be Low Frequency. High Frequency Noise is due to Setup issues Ferrites Chokes Split Termination
1/19/2005 Jim Lawlis - Ford Motor Company 32 Base Line Rod
1/19/2005 Jim Lawlis - Ford Motor Company 33 Baseline BICON
1/19/2005 Jim Lawlis - Ford Motor Company 34 Ferrites Rod
1/19/2005 Jim Lawlis - Ford Motor Company 35 Ferrites BICON
1/19/2005 Jim Lawlis - Ford Motor Company 36 Choke 51uH Rod
1/19/2005 Jim Lawlis - Ford Motor Company 37 Choke 51uH BICON
1/19/2005 Jim Lawlis - Ford Motor Company 38 Split Termination Rod
1/19/2005 Jim Lawlis - Ford Motor Company 39 Split Termination BICON
1/19/2005 Jim Lawlis - Ford Motor Company 40 Next Issues Wake/Sleep Bus Recessive State Loading Not a lot of information availableNot a lot of information available Questions/Discussion