Hi,

I finally made it to work. The EL3104 use a DC sync mode with SYNC1 event 
(0x1C33:01 = 0x03 - DC mode with SYNC1 event). In other words, we must set the 
SYNC0 cycle time to the desired cycle time and SYNC1 cycle time is in fact the 
time until the "Start latch" time. Hence, SYNC1 cycle time is used to "shift" 
the input latch. The odd thing is that to have a proper synchronization, we 
found by trial and error that we have to respect the following contraint:

SYNC1 Cycle Time >= 2 * SYNC0 Cycle Time

Otherwise, the "Shift too short counter" (0x1C33:0D) isn't null and the slave 
stays in SAFEOP mode. According to the documentation, my first guest, which 
didn't work, was to set the SYNC0 cycle time to my master period and then set 
SYNC1 to zero. Also, as an example if we use the smallest allowed cycle time 
for SYNC0, ie:

SYNC0 Cycle Time = 100 000
SYNC1 Cycle Time = 200 000

Then, based on that the slave also compute these values:

0x1C33:05 Minimum Cycle Time = 537 800
0x1C33:06 Calc and Copy Time = 337 800

This is a bit odd to me and I can't see what's the logic behind that. Notably, 
I can't see how the Minimum Cycle Time can be larger than the actual cycle 
time. Does someone have a clue?

Philippe

-----Message d'origine-----
De : [email protected] 
[mailto:[email protected]] De la part de Gavin Lambert
Envoyé : 13 décembre 2012 01:27
À : 'Sven Albert'; [email protected]
Objet : Re: [etherlab-users] Beckhoff EL3104 - Problem while setting OP

Quoth Sven Albert:
> Giving sync0 and sync1 period shouldn't be a problem at all.

True, but it does have surprising behaviour if you haven't read the spec (or 
haven't read it carefully enough).  In particular, if you want the actual sync0 
and sync1 periods to be the same (and synchronous), then you have to
*configure* the sync0 period as the desired period and the sync1 period as 0.

And it's very easy to paint yourself into a corner where the master won't be 
able to establish a DC lock.


_______________________________________________
etherlab-users mailing list
[email protected]
http://lists.etherlab.org/mailman/listinfo/etherlab-users
_______________________________________________
etherlab-users mailing list
[email protected]
http://lists.etherlab.org/mailman/listinfo/etherlab-users

Reply via email to