XK_ETH_XU316_DUAL_100M board release ? Topic is solved

[benbiles]

Hi , is there any news on the XK_ETH_XU316_DUAL_100M board release ?
I noticed it mentioned in lib_board_support.
Any chance of the design files so I can make it myself ? I can live with a dropped packet 1% of the time as its just for development.

[Ross]

Please contact xmos support directly and they should be able to provide the design files.

[Ross]

Just FYI that first packet being dropped 1% of the time looks to be a phy issue - we spotted it as our regression testing looks for 100% correct traffic. We are unaware of a work-around.

[benbiles]

Great thanks, I'll contact xmos support and see.

I just designed a 2 x XMOS XE216-512-TQ128-C20 board with hopefully 32 channels of USB <-> AVB ( 4 x TDM 8 channel data lines )

maybe that will be less trouble than merging the USB and AVB example code for XK_ETH_XU316_DUAL_100M for testing.
Unless someone already did that?

[Ross]

I think this has been done by a few people - partners, customers, and even an internal "fun" project some time ago - but nothing been made public AFAIK.

[fabra]

We have done an AVB to Milan Interface on the basis of a XE232. Also fully MIlan compliant.
See https://www.joyned.at/mu32-milan-to-usb

When it comes to dual PHY we usually use Marvell 88E6320 with one uplink to XMOS. I guess the 6320 is in the same ballpark as buying two separate PHYs and then there are more resources for processing available on the XMOS. What do you think?

[Ross]

An update on XK_ETH_XU316_DUAL_100M, it will remain an internal test platform due to its flexibility. Support will be removed from lib_board_support. It was a modular design (comprising 4 boards and some sizeable connectors) and not ideal for providing externally.

For anyone interested, I attach an image of what it looks like.

A new board (XK-ETH-316-DUAL) will be released with similar capabilities, those following lib_board_support will notice references appearing soon.