TMDSIDK437X: EtherCAT Slave DC issue

Bernd Sprenger

Part Number: TMDSIDK437X
Other Parts Discussed in Thread: AM3359

Hello,

I successfully created and compiled the "ethercat_slave_full_AM437x_arm" demo according to the PRU-ICSS Slave User Guide. It works fine when configured with the "Distributed-Clock" set to FreeRun/SM-Synchron. However, as soon as a select the "Distribute-Clock" mode to "DC-Synchron" it can not switch from PreOp->SaveOp->Op. The other commercial device on the bus is able to do this switching (the AM473x is the second device on the bus). This effect can be discovered with TwinCAT, as well as with the Acontis EC-Engineer and the Acontis EtherCAT Master running on a TMDXIDK5728.

I even tried the MLO and app files from the e2e support thread (e2e.ti.com/.../704543

Can you give me a hint?

Have a nice day,

Bernd

over 5 years ago

0 Bernd Sprenger over 5 years ago

Intellectual 340 points

I forgot to add that the AM437x reports a Sychronization Error....

0 Garrett Ding over 5 years ago in reply to Bernd Sprenger

TI__Mastermind 43296 points

Hi Bernd,

Did you put the device description file TiEtherCATlib.xml file under examples\ethercat_slave\esi to your master?

Also, does the pre-built binary work for you: http://software-dl.ti.com/processor-industrial-sw/esd/PRU-ICSS-ETHERCAT-SLAVE/latest/index_FDS.html

Regards,
Garrett

0 Bernd Sprenger over 5 years ago in reply to Garrett Ding

Intellectual 340 points

Hello Garrett,

thanks a lot for your help. The issue with TwinCAT is solved. However, it still doesn't work with the Acontis EcEngineer and also not with Acontis EtherCAT Master when DC is enabled on the slave (Distributed-Clock). The other device on the bus (MaxPos from Maxon) works fine. I get a Synchronization Error when going into Operation-Mode

My setup with EcEngineer as well as the EtherCAT-Master resembles the one described in "TIDUBZ1B-EtherCAT® Master Reference Design for Sitara™ AM57x Gigabit Ethernet and PRU-ICSS with Time-Triggered Send". Same Idk, Acontis EcEngineer and Acontis MasterStack... The only difference is the additionial device on the bus.

I also just tried a AM3359 ICE V2 with the same result - it doesn't go into Operation-Mode with Distributed-Clock enabled on the slave with the Acontis EcEngineer and also not with the Acontis MasterStack.

Are the pre-built binaries for previous versions of the PRU-ICSS-ETHERCAT-SLAVE somewhere available (I couldn't find them on your website)? Then I could check, if it would work with a previous version.

Thanks a lot - have a nice day,

regards

Bernd

0 Garrett Ding over 5 years ago in reply to Bernd Sprenger

TI__Mastermind 43296 points

Hi Bernd,

I recently attended the EtherCAT plugfest and met Acontis. We were able to test DC mode without issue. Have you also contact Acontis for the issue you are seeing?

Previous EtherCAT pre-built binaries are here -

http://software-dl.ti.com/processor-industrial-sw/esd/PRU-ICSS-ETHERCAT-SLAVE/01_00_05_00/index_FDS.html

Regards,
Garrett

0 Bernd Sprenger over 5 years ago in reply to Garrett Ding

Intellectual 340 points

Hello Garrett,

thanks for your answer - I am also in contact with Acontis - I had the version V3.0.1.12 of the MasterStack and they provided me with an updated version (V3.0.3.05). With this application their demo app is now able to put the AM473x into operational state with the DC-modes "BusShift", "MasterShift" and "TTS with LinkLayerRefClock". The first two also work in my application now - I only have to find the reason for the third one to fail...

Thanks a lot for your help,

have a nice day,

Bernd

Processors

Processors forum

TMDSIDK437X: EtherCAT Slave DC issue