Hello,<div><div><br></div><div>A coworker and I are picking up where others have left off in developing software for a robot using EtherCAT for communication.</div><div><br></div><div>We are working with four EtherCAT slaves connected to a master node running on a Zotac ZBOX nettop with a Realtek 8169 ethernet card. Our wiki page describes our setup: <a href="http://code.google.com/p/atrias/wiki/NettopBuild" target="_blank">http://code.google.com/p/atrias/wiki/NettopBuild</a>. Basically, we are running 32-bit Xubuntu 10.04, using kernel version 2.6.32.11 with an RTAI patch.</div>
<div><br></div><div>Our problem is that when we start EtherCAT master using /etc/init.d/ethercat start, the master never sees the slaves. This is the output in dmesg: <a href="http://pastebin.com/HyTMTWpP" target="_blank">http://pastebin.com/HyTMTWpP</a>. Here is the entire dmesg output: <a href="http://pastebin.com/YLXSMr9k" target="_blank">http://pastebin.com/YLXSMr9k</a></div>
<div><br></div><div>After "EtherCAT 0: Starting EtherCAT-IDLE thread.", nothing happens. Nothing I do with /opt/etherlab/bin/ethercat shows up in dmesg (e.g., rescan should show a message saying, "Scanning bus...", but doesn't). The green ethernet port activity indicator lights, which should blink rapidly, are idle.</div>
<div><br></div><div><br></div><div>We've tried a lot of things (shooting blindly half the time), including the following:</div><div><ul><li>Recompile kernel with various modules enabled/disabled.</li><li>Recompile RTAI after each kernel recompilation.</li>
<li>Recompile EtherLab master after each kernel recompilation, with the addition of "--enable-cycles", "--enable-generic", "--enable-eoe", "--enable-debug-if", and "--enable-hrtimer" in various combinations.</li>
<li>Set up everything (per wiki instructions) on a completely fresh machine, with the same results. (So it is possible that our wiki is missing a step or two.)</li></ul><div><br></div></div><div>If it helps any, we do have an MSI nettop with an EtherCAT master that can at least see the slaves, though it has problems involving timeouts (topic of another question I posted a few days ago). We've been trying vainly to figure out what makes the MSI machine work, because as far as we can tell, both machines have the same Realtek ethernet card and have had the same compilation options for RTAI and EtherLab.</div>
<div><br></div><div>Does this sound familiar to anyone? We are completely stumped on what to do next. What other information could I provide that would help you help us?</div><div><br></div><div>Soo-Hyun Yoo</div>
</div>