[etherlab-users] Invalid watchdog configuration

[Gavin Lambert]

On 20 October 2016 02:57, quoth Mathias Thiele:

<OpMode>

<Name>DcOn</Name>

<Desc>DC active</Desc>

<AssignActivate>#x0300</AssignActivate>

<CycleTimeSync0 Factor="1">0</CycleTimeSync0>

<ShiftTimeSync0>0</ShiftTimeSync0>

</OpMode>

…

    ecrt_slave_config_dc(sc, 0x0300, 0, 0, 0, 0);

 

This is not a correct translation of that configuration.

 

“Factor=1” means that the DC cycle time is expected to match your SM cycle time, ie. the rate at which you will be posting domain datagrams, which is typically the cycle rate of your real-time task.

 

For example, if you have a 1ms (1,000,000ns, 1kHz) realtime cycle rate, then your config should be:

                ecrt_slave_config_dc(sc, 0x0300, 1000000, 0, 0, 0);

 

Some slaves are less forgiving than others about the DC and SM cycles getting out of sync once in OP, so if you have further issues then you might need to adjust the way you sleep in your realtime task, slow your cycle rate, or use a more latency-guaranteed realtime framework.

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.etherlab.org/pipermail/etherlab-users/attachments/20161020/818d123c/attachment-0004.htm>