Re: [10GMMF] 62MMF connector loss and draft test link config

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

Thread Links

Date Links

Thread Prev

Paul,

I don't think any of us would recommend this as a good practice. The question is, should we condemn it as bad practice? As you are probably aware, the vast majority of 1G and 10G links work just fine with mixed 50/62.6 cable plant (as few operate at the extreme corners of the specification). I'm dubious about 10GMMF "working just fine."

-----Original Message-----
From: owner-stds-802-3-10gmmf@LISTSERV.IEEE.ORG [mailto:owner-stds-802-3-10gmmf@LISTSERV.IEEE.ORG]On Behalf Of Paul Kolesar
Sent: Friday, August 20, 2004 8:39 AM
To: STDS-802-3-10GMMF@LISTSERV.IEEE.ORG
Subject: Re: [10GMMF] 62MMF connector loss and draft test link config

Jonathan,
Systimax Solutions has never recommended mixing core sizes within a channel. While it may be tolerated by LED-based applications by accounting for the extra insertion loss, laser-based applications have the potential to generate more modal noise. Given that the loss between 62 and 50 um can range upwards to over 4 dB, depending on the mode fill of the 62 um fiber, the modal noise effects could be significant. This is an area in need of study if mixing of core sizes is deemed desirable. But up to now, Systimax Solutions has simply not supported it.

Regards,
Paul Kolesar
SYSTIMAXŽ SOLUTIONS
1300 East Lookout Drive
Richardson, TX 75082
Phone: 972.792.3155
Fax: 972.792.3111
eMail: pkolesar@systimax.com

Jonathan Thatcher <jonathan.thatcher@IEEE.ORG>
Sent by: owner-stds-802-3-10gmmf@listserv.ieee.org

08/20/2004 09:51 AM
Please respond to jonathan.thatcher

To: STDS-802-3-10GMMF@listserv.ieee.org
cc: (bcc: Paul F Kolesar/CommScope)
Subject: Re: [10GMMF] 62MMF connector loss and draft test link config

This thread brings up an interesting question. In the past, we have never written into the standard any requirements, or even guidance regarding a mix of 50 and 62.5 micron fiber. Need we now? jonathan -----Original Message----- From: owner-stds-802-3-10gmmf@listserv.ieee.org [mailto:owner-stds-802-3-10gmmf@listserv.ieee.org]On Behalf Of David Cunningham Sent: Friday, August 20, 2004 4:45 AM To: STDS-802-3-10GMMF@listserv.ieee.org Subject: Re: [10GMMF] 62MMF connector loss and draft test link config Paul, RE: 70/70 launch In the literature there are many versions of the so called steady state launch mode power distribution (MPD) or the equilibrium mode power distribution. As far as I can ascertain the 70/70 launch method is a simple method to approximate the steady state mode power distribution condition. So on that basis I believe the calculations are consistent with the 70/70 launch condition. But I did not specifically model 70/70 launch. For your information, from the various forms of steady state MPD equations, I used the one that was simplest for me to add to my calculation. It is the (1 - N/Nmax) version of Olshanksy: Olshanky, R (1975). Applied. Opt. 14: 935-45. As I said there are others I could have used but with little real difference in the results. RE: 7 or 9 um The range of MPD distributions that could have be used to measure the installed connectors span the range from OFL to the Steady State. I agree that if Steady State is assumed then the offset could be as high as 9 um for 0.5 dB loss. However, the discussion concluded that 9 um did not appear to be a believable value to use - it's an awfully big offset even for MMF connectors. So, as I recorded we decided to use the value of 7 um. We also placed two connectors at the 7 um offset. The good news is that we done enough work to narrow the range of offset to (7 to 9) um. This is quite a small range of values and hopefully we can formally adopt a value in that range at the next meeting. I would suggest that those that have started testing or simulations might be wise to include both 7 um and 9 um as the maximum offset value. Regards, David -----Original Message----- From: owner-stds-802-3-10gmmf@listserv.ieee.org [mailto:owner-stds-802-3-10gmmf@listserv.ieee.org]On Behalf Of Paul Kolesar Sent: 20 August 2004 04:30 To: STDS-802-3-10GMMF@listserv.ieee.org Subject: Re: [10GMMF] 62MMF connector loss and draft test link config David, thanks for this work. The model I have for 62.5 agrees with your analysis for the steady state launch condition. Please confirm that in your model the steady state condition represents a 70/70 launch (i.e. 70% of maximum core diameter and 70% of maximum NA). I have a comment regarding the 7 um limit. The 7 um value was chosen for the OM3 work because it represented 0.5 dB for 50 um fiber under steady state conditions. The remaining 0.25 dB needed to reach the 0.75 dB limit was allotted to mechanisms other than lateral offset. The equivalent offset for 62 um fiber is 9 um from your graph. Thus connectors need to be of tighter tolerance for 50 um fiber than for 62 um fiber. As product lines of cabling vendors in North America and other parts of the world have shifted to include a larger mix of 50 um fiber, the tolerance for connectors needed to be tightened to maintain good performance. Fortunately progress in connector technology permitted this. The result is that over time connectors have gotten better. This leads to an interesting issue when dealing with legacy plant. The older plant likely has poorer connectors due to both tolerance and wear. I think this leads to a need for considering different values for legacy 50 and 62 um plants. To address these considerations, I would suggest that the limit for 62 um fiber be set at 9 um and that for 50 set at 7 um. Regards, Paul Kolesar SYSTIMAXŽ SOLUTIONS 1300 East Lookout Drive Richardson, TX 75082 Phone: 972.792.3155 Fax: 972.792.3111 eMail: pkolesar@systimax.com David Cunningham <david_cunningham@AGILENT.COM> Sent by: owner-stds-802-3-10gmmf@listserv.ieee.org 08/18/2004 02:24 PM Please respond to "IEEE 802.3aq 10GBASE-LRM" To: STDS-802-3-10GMMF@listserv.ieee.org cc: (bcc: Paul F Kolesar/CommScope) Subject: [10GMMF] 62MMF connector loss and draft test link config Dear All, As promised on todays Fiber Channel Ad Hoc I attach a PDF file containing the following information: * My calculated graphs of connector loss versus offset for overfilled launch, steady state launch and 23 um offset single mode launch. * The agreed draft of the worst-case test configuration for channel time variation measurements with comments from our call to document our reasoning behind the chosen configuration. Regards, David