[etherlab-users] Failed to read reg alias of slave (plain text)

[Florian Pose]

Hi,

On Tue, Mar 16, 2010 at 09:38:30AM +0100, Len Remmerswaal wrote:
> Hi all,

thanks for posting in plain text. ;-)

> I am trying the etherlab master, and I am getting the following message 
> when attaching a small ethercat segment and starting the Etherlab master 
> (v. 651e9cfb244b).
>
> EtherCAT DEBUG: Reading alias from register 0x0012 of slave 5.
> EtherCAT WARNING: Failed to read reg alias of slave 5.

This warning means, that the slave alias could not be read out of
register 0x0012 and is taken from the SII (EEPROM) instead. This is not
crucial. Maybe we can output it in debug mode only.

> Revision number: 0x00000000 ???
> Serial number: 0x00000000 ??? All 0?

This is read from the SII (EEPROM). These two are vendor specific. If
the vendor of your slave decides not to maintain them, that's the way it
is.

> Distributed clocks: no ??? Beckhoff's docs say yes

This is read out of the appropriate slave register, usually slaves do
not lie at this place. Maybe you got a very /old/ slave? I don't think
that Beckhof EL3162 in revision 0 supports DC. Cou can read out the
software/hardware version from the SDOs 0x1009 and 0x100a (see
dictionary). Maybe you can ask Beckhoff, if this version has to support
it...

> 0 N/A down open yes 4 ??? All ports N/A ?
> 1 N/A down open yes 6
> 2 N/A down closed no -
> 3 N/A down closed no -

This is also read from register 0x0110. You can check it again with the
following command: 'ethercat reg_r -pX -tuint16 0x0110.
The N/A means the port physics.

> When running a few reg_read commands I get this:
>
> # reg_read -p5 0x0012 2
> Failed to read register: Input/output error

see above.

-- 
Best regards,
Florian Pose

http://etherlab.org