RTOS/AM4377: EtherCAT MDIO issue

Yong-kyoung Kim

Intellectual 470 points

Part Number: AM4377
Other Parts Discussed in Thread: AM4379, CDCE913

Tool/software: TI-RTOS

Hi, I'm using custom board modified from IDK.

Implementing EtherCAT slave on my board, I found some problem when using example code.

When power-on my board, example code stuck in MDIOPhyRegRead() function.

It stuck at this:

while(MDIO_USERACCESS_GO_EN_0x1 ==
HW_RD_FIELD32(baseAddr + MDIO_USERACCESS(0U),
MDIO_USERACCESS_GO));

At the first attempt to read that address success sometimes, but second attempt fails.

I want to get some help for this problem.

Where should I review to solve it? Hardware or software? And PHY chip or clock synthesizer?

Please let me know....

Thank you.

over 8 years ago

0 Biser Gatchev-XID over 8 years ago

TI__Guru**** 393215 points

The EtherCAT experts have been notified. They will respond here.

0 Paula Carrillo over 8 years ago

TI__Mastermind 40580 points

Hi Yong-Kyoung Kim, which EtherCAT version are you using? are you using PRU-ICSS EtherCAT slave (TI-RTOS Processor SDK) ? or from Industrial SDK ?

Also, you mentioned you are working on a modified IDK board, could you please elaborate which changes?. On the other hand, I believe for Industrial Development Kit (IDK) we only have AM4379 IDK not AM4377 (just a comment)

Thank you,

Paula

0 Yong-kyoung Kim over 8 years ago in reply to Paula Carrillo

Intellectual 470 points

Hi, Paula.
I changed Xtal, but not CDCE913. And IDK is degined for AM4379, but your examples are works fine with AM4377(It's confirmed with another custom board.
I'm using Industrial SDK. And I'm not using PDK, Linux. I Only use SYS/BIOS 6.42.2.29.
Well, I couldn't be sure something changed, because I use IDK schematic as it is. (so, PHY is same.)
In software, I added PMIC control before ICSS initialization.
Clocks are same at 25MHz.
Are there somethings I could check?
Thank you.

0 Paula Carrillo over 8 years ago in reply to Yong-kyoung Kim

TI__Mastermind 40580 points

Hi Yong-Kyoung Kim, one suggestion could be start checking MDIO registers (Link and Alive) when this happens. I will also ask around to the experts for additional ideas

PRU_MDIO 0x4a332400

thank you,

Paula

0 David Zaucha over 8 years ago in reply to Paula Carrillo

TI__Expert 7570 points

Hi Yong-Kyoung Kim

Paula's suggestion is the best place to start the analysis.

On a related point - it sounds like you are using the SYSBIOSSDK-IND-SITARA - If so, we recommend moving to the PRU-ICSS-INDUSTRIAL-SW since TI will no longer provide SYSBIOSSDK-IND-SITARA after 30 June 2017. This is described on the web page http://www.ti.com/tool/sysbiossdk-ind-sitara.

David

0 Garrett Ding over 8 years ago in reply to David Zaucha

TI__Mastermind 43296 points

Hi Yong-Kyoung,

>>When power-on my board, example code stuck in MDIOPhyRegRead() function.

Make sure you have PHY reset properly and PHY address configured correctly. And as Paula pointed - you need to look into MDIO ALIVE and LINK registers. ALIVE bits shall be set for corresponding PHY addresses if it is initialized correctly.

Regards, Garrett

0 Yong-kyoung Kim over 8 years ago in reply to Garrett Ding

Intellectual 470 points

Thank you for your answer.
I solved the problem by some hardware debugging..
I connected the 5V of BiSS bus on AM4377. Getting rid of this, MDIO issue is solved..
The BiSS master is on PRU, did this affect to another PRU-ICSS?

Thank you.

0 Paula Carrillo over 8 years ago in reply to Yong-kyoung Kim

TI__Mastermind 40580 points

Hi Yong-Kyoung, We think it is OK your changes, however, in a quick check, we didn't see a 5v connected to BiSS. Maybe this was a mistake? or could you elaborate more the changes you did?. Please let us know if everything is working ok in order to close this thread or if you need further assistance I would ping our HW experts.

thank you,
Paula

0 Yong-kyoung Kim over 8 years ago in reply to Paula Carrillo

Intellectual 470 points

Hi Paula,

Yes, it is a mistake.. I had to use a level shifter on my board.
EtherCAT things are works fine, now.

Thank you.

Processors

Processors forum

RTOS/AM4377: EtherCAT MDIO issue