[etherlab-users] Link down: Hardly to recognize

Florian Pose fp at igh-essen.com
Tue Mar 24 14:09:13 CET 2009


Hello,

On Thu, Mar 12, 2009 at 05:59:13PM +0100, Erwin Burgstaller wrote:
> If the EtherCAT cable is pulled out, the link goes down and as long as
> the master module isn't unloaded, the master works like every thing's
> fine.
> 
> "ethercat master" looks like this:
> 
> Master0
>   Phase: Operation
>   Slaves: 3
>   Device0: 00:d0:c9:ae:49:8d (attached)
>     Tx count: 14388072
>     Rx count: 14352228
>   Device1: None.
> 
> Only the fact, Tx count and Rx count are the same on subsequent calls
> might be a sign that some thing's wrong, beside from some log file
> entries (e.g. Link down).
> 
> "ethercat slaves" and "ethercat sdo" are too operating normal, but it
> would be much better, if at least the slaves states are shown as OFFLINE
> or something like that.

You're right, this is a point on my TODO list, the link state should be
displayed in 'ethercat master' and the slave list should be cleared on
link down. I'll fix that in the near future.

> Unloading and reloading the realtime module works surprisingly too and
> just for simple digital IO's there's no difference in having a link or
> not.

You should always check your process data transfer with
ecrt_domain_state(). This the realtime way of detecting a link loss.

-- 
Best regards,
Florian Pose

http://etherlab.org



More information about the Etherlab-users mailing list