Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

Re: [802.3_NGAUTO] variations in EEE times and Table 78-4 parameters



Thanks, Natalie.  I had checked the comments but not the presentations.  he and I had talked about it earlier, because of the different format, I missed it.

Saied – thank you – assuming the various equalities hold, your defining of Tw_sys_tx on slide 9 should be enough to fill out the table.

 

From: NATALIE WIENCKOWSKI <nwienckowski@xxxxxxx>
Sent: Friday, January 11, 2019 6:25 AM
To: George Zimmerman <george@xxxxxxxxxxxxxxxxxxxx>
Subject: Re: variations in EEE times and Table 78-4 parameters

 

George,

 

Saied has a proposal for Table 78-4 in his contribution:  http://www.ieee802.org/3/ch/public/jan19/Benyamin_3ch_01a_0119.pdf.

 

Natalie

 


From: George Zimmerman <george@xxxxxxxxxxxxxxxxxxxx>
Sent: Thursday, January 10, 2019 7:06 PM
To: STDS-802-3-NGAUTO@xxxxxxxxxxxxxxxxx
Subject: [802.3_NGAUTO] variations in EEE times and Table 78-4 parameters

 

I’ve been bugged thinking about comment 199 and the precise synchronization of EEE modes.

I agree with the editor, that the times Tq, Tr, Ts are as precise as the transmit clock of the PHY – they are referenced at the MDI and there is no variation in symbol length for these blocks.

 

However, what was bugging me is that it just seemed that something had to be missing, because the signaling at the xMII couldn’t be precisely synched to the PHY frame blocks.  While Tq, Ts and Tr can be synced (because they are all referenced to the MDI signaling), the times relative to xMII signaling can’t be precise.  Then it occurred to me.  It looks like we are missing entries for the shrinkage parameters (and the systems, and definition of the parameters in Table 78-4. (this isn’t in the draft and I don’t see a comment to add it).

 

For those looking at what the proper values might be, and maybe putting together a late comment to add Table 78-4 into the draft, here is what is needed:

 

Looking at the table, it contains: Tw_sys_tx(min), Tw_phy(min), Tphy_shrink_tx (max),

 

Tphy_shrink_rx(max) and Tw_sys_rx(min),

 

For 1000BASE-T1, which uses the same synchronization mechanism of phy frames to xMII signaling, the first 3 of these are identical (10.8 us) and the last 2 are zero, along with a note:

All data transmission in 1000BASE-T1 PHY is synchronized to the PHY frame boundary. As such, the EEE function in the 1000BASE-T1 PHY is expected to assert the wake signal only at specific moments of time, aligned to PHY frame boundaries, and no shrinkage or delay at the RX side is expected.

 

For 2.5G/5G/10GBASE-T1 is seems as well, Tphy_shrink_rx(max) and Tw_sys_rx(min) are zero.

 

This means, since Tw_sys_tx(min) = Tw_sys_rx(min) + Tphy_shrink_tx(max) + Tphy_shrink_rx(max), and T_phy(min) = Tphy_wak(min) + Tphy_shrink_tx, that Tw_sys_tx(min) = T_phy(min) = Tphy_shrink_tx(max) = Tphy_wake_tx(max) -Tphy_prop_tx(min), or, the max time to synch to the phy frame.

(This all refers to Figure 78-5, if you want a diagram)

 

So, as we work issues like alert and the final details of EEE, it would be good to keep these numbers in mind so that we can fill in the missing Table 78-4 EEE parameters as well…

 

-george


To unsubscribe from the STDS-802-3-NGAUTO list, click the following link: https://listserv.ieee.org/cgi-bin/wa?SUBED1=STDS-802-3-NGAUTO&A=1


To unsubscribe from the STDS-802-3-NGAUTO list, click the following link: https://listserv.ieee.org/cgi-bin/wa?SUBED1=STDS-802-3-NGAUTO&A=1