OVMS3/OVMS.V3/components/canopen/docs/Howto-detect-CANopen-nodes.txt

199 lines
7.7 KiB
Text

================================================================================
How to detect and identify CANopen nodes
================================================================================
So you've got a CAN bus with some devices, but you don't know which of them
speaks "CANopen", if any? The OVMS v3 will help you to detect them and open
their CANs ;)
CANopen basics
--------------
A CANopen network consists of "masters" and "slaves", masters are clients,
slaves are servers. At most one master may be active at a time.
CANopen supports addressing of up to 127 slaves on a bus using node IDs 1-127.
Node address 0 is used for broadcasts to all nodes.
The CANopen protocol mainly consists of…
- PDOs (process data objects)
- SDOs (service data objects)
- NMT (network management)
- SYNC (synchronisation)
- EMCY (emergency events)
PDOs are regular, normally periodical, status update frames, for example sensor
data. You can log them using "can … trace on". PDOs can be sent at any CAN ID
except those reserved for other CANopen services.
SDOs are memory registers of nodes that can be read and written by masters on
request. SDO requests are normally sent at ID 0x600 + nodeid, responses at ID
0x580 + nodeid. SDOs are addressed by a 16 bit index + 8 bit subindex. Registers
and data types for a given device are documented by the device specific object
dictionary, normally represented as an EDS (electronic data sheet) file.
NMTs are short datagrams to control node startup / shutdown. There's a special
node state "pre-operational" allowing access to all operation and communication
parameters of a node in a standardized way. NMT requests are sent at ID 0x000,
responses and unsolicited updates (aka heartbeats) are sent at ID 0x700 +
nodeid with length 1.
SYNC messages are datagrams of length 0, normally sent at ID 0x080.
EMCY messages are only sent if a node encounters some kind of alert or warning
condition, they are normally sent at ID 0x080 + nodeid with a length of 8 bytes.
So if any of these IDs look familiar to you, chances are you've got a CANopen
network.
Note: CANopen coexists nicely with OBD-II and often does in a vehicle (i.e.
Renault Twizy). OBD-II devices normally are addressed at IDs > 0x780 so are
outside the CANopen ID range. Even if they use non-standard IDs, the devices
normally will detect and ignore frames not matching their protocol.
But of course there may be non-CANopen devices on the bus producing regular data
frames at CANopen response IDs and/or reading and possibly misinterpreting
CANopen requests sent to node IDs not planned by the manufacturer. Chances are
low this triggers problems, but you should be ready to switch off the vehicle
when doing a scan -- just in case.
Before you begin
----------------
…you need to activate the CAN bus(es) you're going to use. As a CANopen master
needs to write to the network, you need to start the interfaces in active mode,
i.e. …
OVMS > can can1 start active 500000
OVMS > can can2 start active 125000
… and then start the CANopen master for the bus(es), i.e.:
OVMS > copen can1 start
OVMS > copen can2 start
Detecting CANopen nodes
-----------------------
The "open" in "CANopen" means any implementation can decide how much of the
standard it implements. There are some few mandatory features though, a CANopen
slave has to implement, if it wants to comply with the standard.
The mandatory features helping to detect and identify CANopen nodes on a CAN
bus are:
- NMT (network management), especially RESET and PREOP
- NMT bootup event messages
- Standard SDO access in pre-operational mode
If you've got CANopen nodes on a bus, even silent ones, issuing "copen … nmt
reset" will tell all of them to reboot, and as bootup messages are mandatory,
you will see them in the OVMS log output like this:
I (162904) canopen: can1 node 1 new state: Booting
The OVMS CANopen master continously monitors the network for NMT and EMCY
messages. After bootup of all nodes, you can get a list of all nodes that have
been detected by issuing "metrics list co.":
OVMS > metrics list co.
co.can1.nd1.emcy.code
co.can1.nd1.emcy.type
co.can1.nd1.state Operational
Note: if you request a reset, nodes may decide to boot into pre-operational
state. That may produce some error messages. Don't worry, you can resolve this
anytime by issuing "copen … nmt start".
Identifying CANopen nodes
-------------------------
In pre-operational state, a CANopen node must be accessible at the CANopen
default IDs. That means if the node supports SDO access, we can query some
standard attributes from it.
That's what "copen … info" and "copen … scan" do. "copen … scan" queries all 127 node
IDs for the SDO standard device attributes. A complete scan takes about 20
seconds. A typical scan could look like this:
OVMS > level debug canopen
OVMS > copen can1 scan
Scan #1-127...
Node #1:
Device type: 0x00000000
Error state: 0x00
Vendor ID: 0x0000001e
Product: 0x0712302d
Revision: 0x00010019
S/N: 0x47c5…………
Device name: Gen4 (Renault Twizy)11 November 2011 12
HW version: 0x00000003
SW version: 0712.0001
Node #23: SDO access failed
Node #25: SDO access failed
Node #27: SDO access failed
Node #30: SDO access failed
Node #87: SDO access failed
Done.
D (227994) canopen: ReadSDO #23 0x1000.00: InitUpload failed, CANopen error code 0xffffffff
D (228194) canopen: ReadSDO #25 0x1000.00: InitUpload failed, CANopen error code 0xffffffff
D (228444) canopen: ReadSDO #27 0x1000.00: InitUpload failed, CANopen error code 0xffffffff
D (228844) canopen: ReadSDO #30 0x1000.00: InitUpload failed, CANopen error code 0xffffffff
D (238384) canopen: ReadSDO #87 0x1000.00: InitUpload failed, CANopen error code 0xffffffff
This means one CANopen node was found, and some non-CANopen frames were
received on IDs 0x580 +23, +25, +27, +30 and +87.
Great! What now?
----------------
As you now know there's a CANopen node, you can look for documentation on it.
You can also try to access more default SDOs to see if you can control and
configure the node.
If you're lucky, the device will provide its own EDS file at SDO 1021.00. You
can check that by issuing…
copen … readsdo <nodeid> 1021 0
The SDO address space layout is standardized:
Index (hex) Object
0000 not used
0001-001F Static Data Types
0020-003F Complex Data Types
0040-005F Manufacturer Specific Complex Data Types
0060-007F Device Profile Specific Static Data Types
0080-009F Device Profile Specific Complex Data Types
00A0-0FFF Reserved for further use
1000-1FFF Communication Profile Area
2000-5FFF Manufacturer Specific Profile Area
6000-9FFF Standardised Device Profile Area
A000-BFFF Standardised Interface Profile Area
C000-FFFF Reserved for further use
See CiA DS301 for details on standard SDOs:
https://www.can-cia.org/standardization/specifications/
For example, a device shall tell about the PDOs it transmits or listens to,
their IDs, update frequency and content structure at SDO registers 1400-1BFF.
This is mandatory in theory but real devices may not fully comply to that rule.
CANopen compliant standard device types like motor controllers need to
implement some standard profile registers. See CiA DS401 for the generic I/O
device class definition or DS402 for motor controllers.
Most devices will require some kind of login before allowing you to change
operational parameters. This is also done using SDO writes, but there is no
standard for this, so you'll need to dig into the device documentation.
Of course there's a lot more on CANopen, but this should get you going.
Have fun!