EtherCAT Star Topology

[Paul_Tyler]

I have 8 Micro-Epsilon confocalDT 2451 distance measurement probes that I want to connect to my PXI EtherCAT Master. The probe controllers only have one port so I cannot connect them in a line topology. A normal network switch didn't work.

I have a PXI real-time target as my EtherCAT master, and am trying to use a Beckhoff EK1100 EtherCAT Coupler connected to four EK1122 2-port EtherCAT Junctions to give me my star topology. The IFC2451s will then connect into the EK1122s. I have tried this setup using Beckhoff's TwinCAT (30-day evaluation) and that works as expected.

If I add my EtherCAT Master to my project I get all the other devices added as well (EK1100, EK1122's and IFC2451) but the topology is wrong. The attached view LV1.GIF shows what LabVIEW gives me. The IFC2451 slave device is shown as being in the top level device chain, but it should be attached to Device 2 (EK1122). When I run my VI I get this message "An error occurred while attempting to switch the I/O scan mode. The distributed clock could not be enabled because the master could not read the network topology. This could be caused by a slave device having an improper connection to the network."

The other symptom I have is that if I right click on any of the EK1122s there is no 'New...' option in the context menu.

If I plug an IFC2451 into the X2 port of the last EK1122, or into the Out (X2) port of the EK1100 I don't get the error switching the I/O scan mode. 

Is there a fix for using these Beckhoff couplers?

Is there any alternative hardware that I can use as an EtherCAT "switch" that has been used successfully with LabVIEW?

Thanks

Paul

[AudioVideoDisco]

Hi Paul,

You mention that a normal network switch didn't work, is there a reason why? What type of network switch did you use?

You can hook up 3rd party EtherCAT slaves behind a 3rd party switch in a star topology. However, using a National Instruments EtherCAT master requires Distributed Clocks (DC) to be disabled on all EtherCAT slaves, this may be why your problem is occurring.

The link below mainly discusses setting up NI EtherCAT RIO but has a small section on configuring 3rd party EtherCAT devices.

Getting started with EtherCAT on NI Programmable Automation Controllers: https://knowledge.ni.com/KnowledgeArticleDetails?id=kA03q000000YHbbCAG&l=en-US

I hope this helps.

Best Regards,

Paul

Applcations Engineer

NIUK

[AudioVideoDisco]

Hi Paul,

I've just stumbled across this comment thread that you've posted in recently and unfortunately you are correct in saying that you are encountering issue 236792.

However, it might be still worth trying to disable the distributed clocks on your slave devices and trying again.

Also in relation to the network switch you have tried; you won't be able to use normal routers switches or hubs with EtherCAT devices. This is because they use the EtherCAT® protocol, which is different than the TCP/IP protocol that typical routers, switches and hubs use.

Let me know how you get on with this issue.

Best Regards,

Paul

Applications Engineer

NIUK

[Paul_Tyler]

I thought it didn't work because of something fundamental with EtherCAT protocols - see crio 9074 ethercat with a switch/hub between the cRIO and the slaves.

The switch I was using was a TP-Link TL-SF1016DS, a straightforward 16-port unmanaged switch.

Is there a recommended method for disabling the Distributed Clocks?

Paul

[AudioVideoDisco]

Hi Paul,

In the manual for the confocal DT (http://www.micro-epsilon.com/download/manuals/man--confocalDT-2451-2471--en.pdf) it mentions a 'Free Run' mode under the 'Distributed Clock'. It might be worth trying this out, I'm not sure if it disables the distributed clock though. You may have to contact Micro-Epsilon to see if this is possible.

In relation to using a normal switch with EtherCAT, more information can be found at this link: Can Switches Hubs or Routers be used with the 9144?

Best Regards,

Paul

Applications Engineer

NIUK

[DirkW]

Hi Paul,

I think we do not support any port devices that change the topology away from a line. You can try disable Dc if it was enabled before but what you need I think is a patch for supporting port devices, which is available for LV 2012 and EtherCAT 2.4. Let me know if you have that version so I can give the patch to you.

DirkW

[Paul_Tyler]

The software version I am using is LabVIEW 2011 Service Pack 1, LabVIEW Real Time 11.0.1, and NI Industrial Communications for EtherCAT 2.4. However I noticed the screen attached to the PXI-8109 controller reports "NI Industrial Communications for EtherCAT 2.3 started successfully" even though NI-MAx tells me the EtherCAT version is 2.4.

Is there a patch available for this combination, or will I have to upgrade to LabVIEW 2012?

[DirkW]

Yes, In order to get the ethercat swirtch devices working over more then 2 ports, you need LV 2012 and ECAT 2.4 with a patch I can provide. We will release a 2.5 driver in end February that has full support for switches including DC, which will support LV 2012 and later.

The way our installer work is that a driver version is always bound to a specific LV/RIO version. So if you install 2.4 on a LV 2011 it will install the latest ECAT driver that had LV 2011 suppoort which is 2.3. LV 2012 sup[ports ECAT 2.4 and 2.5.

For you if you have only two ports to branch you could disable DC for the slaves connected to these branches and it should work as well.

DirkW

[Manucla]

Could you share LabVIEW Drivers of IFC2451?