TMS320F28388S: Ethercat error counter register issue

Hi,

When you say bad cable, what does that imply? A physically damaged cable or a straight vs cross cable? In the second case, PHY's with auto mdix capability will work with no issue. Also, do you know what is the exact error reported by the other device?

Thanks,

Hi Lohithakshan,

it is not physically damaged cable, just may not be good, other device will report error from both register 0x300 and 0x309

Hi Strong,

OK. If there is error in received frame, it should be flagged. This is completely handled by hardware. Can we get the error register dump from the other device to investigate? Also, can we know what EtherCAT master is used in this setup?

Thanks,

Hi  Lohithakshan,

please see the register from other device, and Ethercat master is elmo P-MAS2N

Hi Strong,

Thanks for the data.

After looking at the error counter and referring to the error counter interpretation provided in section 3.12.4 of

https://download.beckhoff.com/download/document/io/ethercat-development-products/an_esc_faq_v1i2.pdf

One reason could be PHY to PHY interoperability differences among various devices. As the PHY is external and can be from different vendors, It is good to check if all the devices in network are using a PHY as per the selection guideline documented at

https://download.beckhoff.com/download/Document/io/ethercat-development-products/an_phy_selection_guidev2.6.pdf 

For F2838x, are they using a TI ControlCARD? What is the application running? Is it a TI provided EtherCAT application or something that is newly developed? Is the master able to scan and identify F2838x when directly connected to it with the suspected cable?

Thanks

Lohithakshan,

customer are using dp83822 for phy, it is customer's servo product based on F28388 which are ethercat slave, and the code is developed based on TI demo example, master is PLC elmo P-MAS2N, the master is able to scan and identify F2838x. 

Hi Strong,

As master is able to scan and identify F2838x, there is no major physical layer interop issue between them. Will check if there is any setting that can impact. Do we know the details of other device - make and functionality, PHY involved etc.? Also, is maser reporting any error on its side - error counters for master?

Thanks,

Lohithakshan,

Other device have the same function as F28388, Phy is ksz, the master report error from other device , not F28388 device. 

Strong,

As discussed, you could disable the odd nibble detection in PHY and see if that solves the issue.

Thanks,

Lohithakshan,

Thanks for your suggestion, would you suggest how to disable the odd nibble detection in PHY? Thanks.