Weird EyeQ Output from Arria 10 Transceivers

Question

Using TTK on an Arria 10 GX, I am noticing some vertically asymmetrical eye diagram output from EyeQ. I am stumped as to what could be causing this? Anyone seen something similar before?

http://www.alteraforum.com/forum/attachment.php?attachmentid=11083&stc=1

This seems to happen to any transceiver I use.

If I use Serial Loopback, the eye still seems asymmetrical.

http://www.alteraforum.com/forum/attachment.php?attachmentid=11085&stc=1

altera_forum · Answer

Generally the EyeQ would show the shape of the received eye diagram. Probably the actual eye is asymmetrical. Have you probe the eye diagram with scope to compare?

altera_forum · Answer

Does anyone know if there was a problem with the ODI/EyeQ circuitry in Arria 10? All mention of OID/EyeQ was scrubbed from the latest release of the Arria 10 Transceiver PHY User Guide.

Carsona - Did you ever figure out what was wrong with your eye diagrams above?

Thanks,

Bob

altera_forum · Answer

Yeh turned out it was a ES2 silicon bug as informed by an Altera engineer. Using production parts along with Quartus 16.0 solved this - I get proper eye diagrams now.

altera_forum · Answer

Thanks for the response. If it was just an ES2 problem why did they eliminate all mention of it in the Arria 10 docs and Arria 10 is not even listed as a supported family for EyeQ in the TTK:

https://www.altera.com/products/design-software/fpga-design/quartus-prime/features/swf-transceiver-toolkit.html

Trying to get an official answer from our local Altera folks.

altera_forum · Answer

I've asked my local FAEs too. Strange.

Forum Discussion

Weird EyeQ Output from Arria 10 Transceivers

6 Replies

Recent Discussions

Life time of EPM3128ATI100

LVDS support on Agilex 7

EMIF Pin Assignment for Agilex 7 FPGA I-Series DevKit (DK-DEV-AGI027R1BES)

Netlength Agilex5 028B

Agilex5 - Timings configuration