MAPLD 2004

1

• MAPLD 2004
• Washington D.C.
• 8th 10th September 2004
• Application and Experience of CAN as a
• low cost OBDH bus system
• Adrian Woodroffe
• Engineer
• Surrey Satellite Technology Limited, Guildford,
  UK

2
Contents

• Introduction to Surrey Satellite Technology Ltd.
  (SSTL)
• SSTL Missions using CAN
• SSTL CAN bus topology
• SSTL CAN hardware solutions
• COTS CAN
• RadCAN
• SSTL CAN for Spacecraft Usage (CAN-SU) protocol
• Future development
• Conclusion

3
Surrey Satellite Technology Ltd

• Predominantly Low Earth Orbit (LEO) missions
• Extensive use of Commercial Off The Shelf (COTS)
  technology, plus
• Module level redundancy
• Passive fail safe system design
• Provides a low cost solution
• Rapid utilisation of new technology
• Now moving to small GEO platforms
• British National Space Centre funded GEMINI small
  GEO communications platform
• Deployable panels
• Harsher radiation environment compared with LEO
• Higher reliability requirements

4
SSTL CAN Flight History
No CAN Bus
Partial CAN Bus
CAN Bus for all TTC
5
CAN Bus System Topology

• Two sub-system categories are connected to the
  CAN bus
• Semi-smart modules that respond to tele-commands
  and provide telemetry
• For example a reaction wheel will have a
  tele-command to set it speed, and a telemetry
  value to read its speed. The microcontroller
  will provide the control loop to maintain the set
  speed.
• Data processing modules, onboard computers etc
  that control the satellite
• Generally based round a microprocessor, the OBC
  will control the attitude and payload functions
  etc by issuing a set of tele-commands on the CAN
  bus
• Redundant busses are selected via latching relay
• A module can be commanded to change to the
  secondary bus, or
• Will automatically switch buses if the module
  does not receive any messages for 5 minuets

6
The Environment

• Low Earth Orbit
• 500Km to 1000Km
• Relatively benign radiation environment
• lt1Krad TID per year at the component (gt2mm Alu
  shielding from module boxes structure)
• SEU rate of approximately 1 per MByte of SRAM per
  day
• SEL, only 4 confirmed cases in SSTLs flight
  history
• A COTS solution is therefore acceptable if
• The system design is passively fail safe for
  attitude, power generation and thermal control
• Module level cold redundancy
• Higher Earth Orbits
• MEO and GEO orbits
• gt10Krad TID per year depending on orbits
• More complex systems require guaranteed active
  attitude control for deployable panels etc
• A radiation tolerant (for both TID and SEU /
  SELs) must be used

7
COTS Solution

• Semi-Smart Modules
• Based round a 8051 microcontroller
• On Chip CAN controller
• Makes maximum use of microcontroller peripherals
• 8 channel, 10 bit ADC for telemetry gathering
• PWMs and URATS
• Physical CAN transceiver
• Latching relay to select bus
• Data Processing Units
• Microprocessor base
• CAN controller memory mapped
• Physical CAN transceiver
• Latching relay to select bus

8
COTS Components Flown
assuming 50 are in cold redundant modules
thus off
9
Radiation Hard Solution

• No Radiation Hard 8051 microcontroller with on
  chip CAN is currently available
• The COTS 8051 does not meet radiation for orbits
  higher then LEO
• Aurelia CAST A (8051 with the CASA2 CAN
  peripheral on the same die is presently under
  development (see www.caen.it/micro)
• Solution for Semi-Smart Modules
• Separate 8051 CPU, with an external CASA2 CAN
  controller (Aurelia design), external PROM, SRAM
  and ADC had to be provided
• Shown as a block diagram on the following slide
• Solution for Data Processing Modules
• The Existing architecture can be copied using
  available radiation hard components
• Processor Atmel SPARC 695
• CAN Peripheral Aurelia CASA2 (Based on ESA
  HurriCAN IP Core)
• CAN Physical
• No Radiation Hard device is currently Available
• The COTS Phillips device is manufactured Silicon
  on Insulator and has shown promising total dose
  performance
• Aurelia have EM samples of a Radiation Hard
  Device
• RS485 drives can be interface to the CAN bus with
  some additional pull-up / pull-down resistors.
  This option is currently under evaluation

10
Radiation Hard Solution
11
COTS / Radiation Hard comparison
COTS
Radiation Hard
12
Protocol Introduction

• Higher layer protocol for CAN developed by SSTL
  CAN for Spacecraft Usage (CAN-SU)
• Used for telemetry, tele-commands and data
  transfers
• Developed in 1995 before the extended 29
  identifier was available
• Driven by Requirements
• Wide range of required telemetry values from many
  modules
• Support multiple system configurations
• Keep it simple
• Limitations
• 11 bit CAN ID field dictates the need for
  extended addressing in the CAN data field

13
CAN-SU Protocol
A sequence number used in burst of 8 packets Not
discussed further here
Address of module or task Sending the
message i.e. the OBC
Various control bits as defined in the CAN
protocol Including data field Length
Defines message type Tele-command, telemetry
request etc
Address of module or task the message is being
sent to i.e. reaction wheel
CAN-SU Bit definitions
4 Bytes of Data
Telemetry or Telecommand Address
ID (8 bits)
Seq (3 bits)
Address (10bits)
From ID (8 bits)
Control (8 bit)
Data (32bits)
CAN Bit definitions
11 Bit CAN Identifier
CAN Control
8 bytes of data in each CAN message
14
Protocol Telecommand
Request Source
Request Sink
Tele-command request
Tele-command request
ID (8 Bits)
Seq 0
Len 7
From (8 Bits)
Address (10 Bits)
Data 32 Bits
Control TC-Req
Tele-command acknowledge
ID (8 Bits)
Seq 0
Len 7
From (8 Bits)
Address (10 Bits)
Data Response 32 Bits
Control TC-Ack
Tele-command ack
15
Protocol Telemetry
Data Sink
Data Source
Telemetry request
Telemetry request
ID (8 Bits)
Seq 0
Len 3
From (8 Bits)
Address (10 Bits)
Control TL-Req
Telemetry response
ID (8 Bits)
Seq 0
Len 7
From (8 Bits)
Address (10 Bits)
Data Response 32 Bits
Control TL-Res
Telemetry response
16
Future Developments (1)

• COTS
• Microchip PIC 3.3V microcontroller with on chip
  CAN running at 3.5 MIPS (compared with 1 MIP
  8051).
• Lower power
• FLASH based program code for in-circuit
  programming (last resort in orbit)
• RAD Hard
• Evaluate Aurelia CASTA
• RadCAN next generation will be a System on a Chip
  (SoC) solution.
• Currently targeted at an Actel RTAX1000
• Combines an open source 8051 IP core with the ESA
  licensed HurriCAN core, uses FPGA SRAM with EDAC
  and hard code the 8051 program code, to produce a
  single chip solution (minus the ADC)

17
Future Development (2)

• RadCAN Next Generation

18
Future Developments (3)

• Protocol
• The use of the 11 bit identifier is severely
  limiting to CAN-SU
• SSTL is investigating a re-work of the CAN-SU
  protocol to use the 29 bit identifier
• The 29 bit identifier could incorporate both CAN
  task ID and telemetry / tele-command channel
  number
• 29 bit identifier could also mean use of the
  Remote Frame is possible
• CANOpen
• SSTL is also investigation the option to migrate
  to the industry standard CANOpen
• http//www.can-cia.de/canopen/
• Heavier protocol than CAN-SU but selected parts
  could be used to replace and improve on CAN-SU
• Provides more complex synchronisation factions if
  required

19
Conclusions

• COTS components
• 10s of orbit years use on SSTL satellites in LEO
• Suitable for radiation benign LEO missions
• Sub-system redundancy and fault tolerant system
  design has resulted in no observed failures of
  the CAN bus or its constituent components.
• Radiation Hard CAN solutions
• Not many components are available at the present
  time but this is changing
• Hurricane (ESA IP core)
• Aurelia CASA 2 (CAN peripheral)
• Aurelia CAST A (8051 micro CAN peripheral), in
  development
• Aurelia CAN Physical driver, in development