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

Re: [802.3_25GSMF] 802.3cc 25GSMF AdHoc Meeting Series - meeting reminder for June 7th

Hi David, Pete,

 

I have put a few slides together to get a discussion going.  Just waiting for review and approval.

 

Thanks,

 

Mark

 

From: David Lewis [mailto:David.Lewis@xxxxxxxxxxxx]
Sent: 05 June 2017 02:11
To: Mark Kimber; Peter Jones (petejone)
Cc: STDS-802-3-25GSMF@xxxxxxxxxxxxxxxxx; 'kohichi.tamura (kohichi.tamura@xxxxxxxxxx)'; 'Jonathan King'
Subject: RE: [802.3_25GSMF] 802.3cc 25GSMF AdHoc Meeting Series - meeting reminder for June 7th

 

Hi Mark,

A couple of comments on your message:

1.       In previous NRZ SMF MPIs there was no explicit MPI penalty, just a small reflection penalty derived from the IEEE link spreadsheet model.  That reflection penalty only considered the reflectance of the Tx and Rx at maximum channel insertion loss.  We are talking about it now because of new knowledge gained when considering PAM4 PMDs, where MPI is a critical impairment that needs to be well controlled.  As shown by Jonathan’s ad hoc presentation in king_01_25gsmf_051017, MPI is also a penalty for NRZ links and can be significant for foreseeable numbers of connectors.

2.       I think the concern for MPI is at maximum loss, rather than minimum, because that is where the penalty affects the overall link budget.  Jonathan’s calculations have the maximum insertion loss lumped just before the receiver.

3.       For both 25GBASE-LR and -ER, we have a transmitter reflectance of -12 dB and a receiver reflectance of -26 dB.  This is in line with other NRZ serial SMF PMDs, but is not what is assumed for any of the PAM4 PMDs.  All of those have -26 dB at both ends in order to control the MPI penalty.

 

So, in addition to considering the link budget with N x -35 dB reflectances in the channel, I think we should consider requiring a -26 dB reflectance at the transmitter.  In practice this would mean having fiber stubs in the Tx and Rx ferrules and would rule out free-space designs with an air gap between the laser and the output fiber.

 

If transceiver makers want to take keep the -12 dB Tx reflectance, we will need to add an additional MPI penalty into the -LR and -ER power budgets. I think it’s appropriate to specify maximum discrete reflectance limits using a table, similar to the one below, copied from 802.3bs D3.1.

 

David

 

 

From: Mark Kimber [mailto:MKimber@xxxxxxxxxxx]
Sent: Sunday, June 04, 2017 3:00 PM
To: Peter Jones (petejone) <petejone@xxxxxxxxx>
Cc: STDS-802-3-25GSMF@xxxxxxxxxxxxxxxxx; David Lewis <David.Lewis@xxxxxxxxxxxx>; 'kohichi.tamura (kohichi.tamura@xxxxxxxxxx)' <kohichi.tamura@xxxxxxxxxx>; 'Jonathan King' <jonathan.king@xxxxxxxxxxx>
Subject: RE: [802.3_25GSMF] 802.3cc 25GSMF AdHoc Meeting Series - meeting reminder for June 7th

 

Hi Peter,

 

I am on business travel at the minute so will be difficult by Wednesday.  I have Jonathan’s spreadsheet downloaded but have used it in anger yet so that might need a couple of attempts to get right.

 

What I could do is outline potential use cases for discussion so we can get that agreed to start with. So how many reflections, etc.

 

Thanks,

 

Mark

 

From: Peter Jones (petejone) [mailto:petejone@xxxxxxxxx]
Sent: 04 June 2017 20:32
To: Mark Kimber
Cc: STDS-802-3-25GSMF@xxxxxxxxxxxxxxxxx; David Lewis; 'kohichi.tamura (kohichi.tamura@xxxxxxxxxx)'; 'Jonathan King'
Subject: RE: [802.3_25GSMF] 802.3cc 25GSMF AdHoc Meeting Series - meeting reminder for June 7th

 

Mark,

 

This feels like an offer to put together a presentation.

Can you have something ready for Wednesday?

 

Regards

Peter

 

_______________________________________________

Peter Jones            Cisco Systems

Principal Engineer     560 McCarthy Blvd.

Campus Switching S/W   Milpitas, CA, 95035 USA

Wrk: +1 408 525 6952   Mob: +1 408 315 8024

Email:                 petejone at cisco.com

About.me:              about.me/petergjones

_____________________________________________

 

From: Mark Kimber [mailto:MKimber@xxxxxxxxxxx]
Sent: Sunday, June 4, 2017 5:05 AM
To: Peter Jones (petejone) <petejone@xxxxxxxxx>; STDS-802-3-25GSMF@xxxxxxxxxxxxxxxxx
Subject: RE: [802.3_25GSMF] 802.3cc 25GSMF AdHoc Meeting Series - meeting reminder for June 7th

 

Hi Peter,

 

My view is we should agree how many connector with appropriate return loss and also the loss and its’ placement in the link.  Run some MPI sims and come up with a number.

 

For instance, it would be reasonable to consider 2, 4 and 6 connectors all with -35dB ORL.  I think the minimum loss for ER is 10dB so we should use that as a minimum.  For LR we can use something like 4dB or 5dB (same idea as used for 802.3bs and 802.3cd for 10km).

 

Thanks,

 

Mark

 

From: Peter Jones (petejone) [mailto:petejone@xxxxxxxxx]
Sent: 04 June 2017 02:56
To: STDS-802-3-25GSMF@xxxxxxxxxxxxxxxxx
Subject: Re: [802.3_25GSMF] 802.3cc 25GSMF AdHoc Meeting Series - meeting reminder for June 7th

 

Folks,

 

We have a meeting scheduled for Wednesday, June 7, 2017 2:00 PM-3:30 PM (UTC-08:00) Pacific Time (US & Canada).

 

I think the major topic that is open is comment i-38 (see below)

 

 

 

Anyone got any good ideas how to progress this topic?

 

Please let me know by EOD Tuesday PT if you want to present on this, or any other topics.

 

Regards

Peter

 

_________________________________________________

Peter Jones 

802.3cc 25GSMF Task Force Ad Hoc Chair  

_________________________________________________

 

 

-----Original Appointment-----
From: Peter Jones (petejone)
Sent: Wednesday, July 6, 2016 3:41 PM
To: Peter Jones (petejone); 802. 3 25GSMF; Matt Traverso (mattrave); Ray Nering (rnering)
Cc: David Malicoat; Jeffery Maki; Jonathan King; Ed Ulrichs; derek.cassidy@xxxxxx; Dino Pozzebon; Carlo Tosetti (ctosetti); Gary Nicholl (gnicholl); mitch.h.fields@xxxxxxxxxxxx; Pirooz Tooyserkani (pirooz); Greg Bradburn (grbradbu); 有本英生 / ARIMOTO，HIDEO
Subject: 802.3cc 25GSMF AdHoc Meeting Series
When: Wednesday, June 7, 2017 2:00 PM-3:30 PM (UTC-08:00) Pacific Time (US & Canada).
Where: Webex

 

 

Extending series.

 

-------------------------------

 

As announced in the F2F at the March meeting, (see http://grouper.ieee.org/groups/802/3/25GSMF/public/1603_Macau/jones_25gesmf_01_0316.pdf ), the 802.3cc 25GSMF AdHoc will re-commence weekly meetings in the Wednesday 2:00pm-3:30pm PST slot.

 

The goal of these meetings is to hear contributions to progress the work of the Task Force.

 

Please send requests for presentation time at least 24 hours before the meeting time (preferably earlier), and send presentations at least 24 hours in advance.

 

If there are no requests for presentation time, the meetings will be cancelled.

 

Regards

Peter Jones

25GSMF 802.3cc AdHoc chair

 
 
-- Do not delete or change any of the following text. --  
 
 
Join WebEx meeting  
Meeting number: 203 644 669 
Meeting password: 25gsmf (254763 from phones) 
 

Join from a video conferencing system or application
Dial 203644669@xxxxxxxxxxxxxxx 
From the Cisco internal network, dial *267* and the 9-digit meeting number. If you are the host, enter your PIN when prompted.  
 
 
If you are a host, go here to view host information.

Join by phone 
+1-408-525-6800 Call-in toll number (US/Canada) 
+1-866-432-9903 Call-in toll-free number (US/Canada) 
Access code: 203 644 669 
Global call-in numbers  |  Toll-free calling restrictions  
 
 
Can't join the meeting? Contact support.  
 
IMPORTANT NOTICE: Please note that this WebEx service allows audio and other information sent during the session to be recorded, which may be discoverable in a legal matter. By joining this session, you automatically consent to such recordings. If you do not consent to being recorded, discuss your concerns with the host or do not join the session..