802.3z PMD Sub committee Chairman: Del Hanson Meeting was called to order 10 Nov 97, Monday, at 8:30am in Montreal with 41 members attending. Steve Brewer Atikem Haile-Marian Jahan Lotfi Bob Payne Dan Brown Del Hanson Andy Luque Petar Pepeljugoski Gadi Lahat Claude Hayek Arlen Martin Bill Quackenbush Edmund Chen Adam Healey Bob Mayer Mark Sankey Yuval Cohen Todd Hudson Vince Melendy Rich Seifert Jay Garcia Howie Johnson Larry Miller Steve Swanson Joel Goergen Chris Karaguleff Shimon Muller Wen-Tsang Tang Thomas Goerrgensan Don Knasel Bob Musk Ken Taylor Moty Goldis Paul Kolesar Mark Nowell Jonathan Thatcher Stephen Haddock Dan Kuchta Bhavesh Patel Schelto Van Doorn Theresa Watkins A secretary was chosen: Joel Goergen **Actually, the guy just didn't know any better and no one stopped him from volunteering. Del Hanson: There are 30 comments to resolve. His idea is to chose a group of ten people to review the proposed comment resolutions, modify, etc, and then review to the sub committee. Jonathan Thatcher was chosen as the chair of that group. The following people were to attend: 1. Jim Tatum, 2. Bob Mayer, 3. Dan Brown, 4. Schelto Van Doorn, 5. Vince Melemdy, 6. Steve Swanson. ** There were more, but I just couldn't get all the names, and, of course, when they met, I didn't do a sign up sheet, either ..... Del Hanson: Proceed through with Technical Resolution comments and have a sub group review these lessor comments off line. No objections - Motion carried. It was not decided when the sub group would meet. The TR comments to resolve are 187, 186, 88, and 86. It sounded like DMD and Jitter were the main points along with 100meters for SX and 200meters for LX was not acceptable. Jonathan Thatcher wanted to establish objectives, etc, of the group. Del proceeded to discuss there are at least two proposals that would address the above listed TR comments. These comments, that if resolved, could keep 802.3z on track. If unresolved, we would slip by 6 months. a. Publish the document b. No new negative objections c. Standards Board By Wednesday at midnight, if we document , we can move forward. We must have 75% agreement of PMD, 75% agreement of 802.3z, and 75% agreement or 802.3 So, we would allow individual proposals to be summarized here - three a, b, and c. a. Chris b. Jonathan / Del c. Jim Tatum Proposal A: Chris with Digital Equipment Corp cover : eye width measurements, 1300nm problem, bit error eyescan **I was not very fast at capturing what he was saying here. **His proposal went out on the Reflector, but will be distributed later. Test Diagram used: 10:15 Break for 15 minutes and then discuss the conclusions from Chris. 10:35 Resume * Scale Factors for 200meters * Restored eye to acceptable parameters * Typical fibers (62.5) 200 meters can be obtained Conclusion: If we write a spec with 100meter distance, then DEC's comment would be resolved. Discuss option B: Del Hanson and Jonathan Thatcher Del distributed the proposal, and stated it was available on the reflector. INSERT PROPOSAL Name -> This proposal indicates we have a problem - requires more scientific data. Howie -> I was too interested in what he had to say to write it down, sorry. Rich-> End user might want conditional launch unless expensive with the longer distance. Name -> columns should read "Laser Certified System" and not "cabling". Del -> We need to look at all parameters ..... Name ->260 and 440 meters, won't these numbers go to the old #'s??? Jonathan -> Get caught up in the concept, not in the #'s. Chris -> Laser certified cabling could be used to determine if your equipment works - we need to use caution here. Howie -> SX and LX approach clarification. SX LX Alternate or SI MMF x x SMF / Offset x Source Conditioning x Chris -> Requires we have a standard certification method to move forward with this as a solution. Name -> Link penalties based on Bandwidth? No- these are based on what we have already decided. Point is longest link lengths are already in document. Jonathan -> If a conditional launch is used, interface spec at the output of each ........ Name -> TIA uses numerical aperture and spot widths - insufficient for this. Proposal C: Jim Tatum [forgot to sign in on the attendance sheet, but was there the whole time] Offset patch cord -30 micron spot size +/- xxx on one meter patch cord. Howie - > Do we need to do anything with conditional/unconditional launches? What do we do on the standard? Del -> Process to continue * Group meets from 1-3 to resolve comments. * Meeting to continue at 7:30pm in room yet to be defined. Adjourn at 12 noon. 1:10pm Task group to discuss comment resolution with Jonathan Thatcher as chair. Bob Mayer, Jonathan Thatcher, Steve Swanson, Bob Musk, Shelto Van Doorn, Mark Nowell, Jim Tatum, Vince Melemdy. 165 - accept 166 - accept 167 - accept 79 - accept 103 - Steve Swanson will write this up and present it later to the pmd group 174 - accept 175 - a. okay b. INS?? Installer Madatory - two pics c. okay d. okay e. okay f. okay, but should be 26db rather then 20db. 176 - accept 177 - accept 81 - accept 168 - accept 169 - accept 185/186/88 ???? 102 - accept 170 - accept 171 - accept 77 - accept 172 - accept 173 - delete 20db 178 - accept 179 - accept 180 - accept 83 - accept 82 - accept 84 - accept 80 - accept Finished at 2:30pm Monday 10 Nov 97 7:30pm PMD continued Steve Brewer Joel Goergen Jahan Lotfi Mark Sankey Dan Brown Ed Grivna Arlen Martin Steve Swanson Richard Greiner Del Hansen Jeff Martin Tad Szostak David Cunningham Todd Hudson Bob Mayer Wen-Tsang Tang Simon Cushin Dave Hyer Vince Melendy Jim Tatum Bob Dahlgren Howie Johnson Larry Miller Pat Thaler Chris Diminico Chris Karaguleff Bob Musk Jonathan Thatcher Thomas Dineem Don Knasel Mark Nowell Shelto Van Doorn George Eisler Paul Kolesar Paul Pace Jeff Warren Howard Frazier Dan Kuchta Petar Pepeljugoski Theresa Watkins Jay Garcia Bruce LaVigne John Ritger Micheal Yam Howard Frazier -> Draft standard 802.3z is in good shape. A small number of comments. We need to get to a sponsor ballot. We have to end this week with a consensus. Then, pending circulation, we go for the sponsor ballot. Question: How many people will look at an item and reject it? If yes, there is a six month slip. If we pick an option that will be largely confirmed, we will be on schedule. Option A - Dramatically de-rate fiber length. Option B - Like A, but laser certified cable. Gets longer multi-mode distance. Howard believes either is less then 20% successful. Option C - Pushes burden back on system suppliers and adds conditional luanch. Question - Does option C require changing jitter specs? * Not sure yet. * Jim needs to detail more stuff because this is a lot of work. Option - Sufficient data to support 200M 62.5 and 400m 50 MMF. Then tweak the jitter numbers. Pat - Installed base is 62.5 fiber, would like to see 300 plus meters Howard: - Has copies of surveys for Tuesday. - Has EIA/TIA letter: 300M on 62.5 fiber. Del: - Decision reached on which option to focus on. - Work on that option. Q: Could we proceed with a de-rated link and proceed in parallel with PAR? Keeps 802.3z on track. Howard: Doesn't think it would help anyone to have a half-baked fiber spec. Jim: Further discuss option 3 or C. Del: Talk more about option 3 or C. Jonathan: What is really meant by CL/UL for MMF? Location of launch condition is not defined on pmd. Jim: Doesn't think SX is as bad off as LX. Normative part of conditioned and unconditioned launches - still need to support stuff out there. Howard: Process issue - Mandate to get 260 meters to 300 meters. David: - LX: Not risk we can do this. - SX: We have good talent in the room to solve this. Pat: We should not have comments listed as Del has on C in spec. Also, we should not use 'UL' in our spec. Tom: We need to have compromise. The customer may have to accept 98% or something of cables. Howard: 1. Can not compromise on Rules. 2. Can not compromise on PAR. 3. Can not compromise integrity as engineers. Del requests special session with the following: David Cunningham Chris Diminico Doug Harshberger Paul Kolasar Mark Nowell Jonathan Thatcher Dan Kuchta Dan Knasel Del Hanson Jim Tatum Don Brown Petar Pepelugoski Finished at xxxxx. 10 Nov 97 9pm Special Committee Jonathan: Number of people in room that are free thinking - we should let them go and work. Del: Suggest just writing down ideas. David: Food (a man after my own heart). Jim's approach 1. 30un Break - be back at 9:30pm 9:30 start: Summary: 1. Conditioned launch measurement technique. 2. Definition of conditional launch. 3. Definition of parameter space for measured values. 4. How to ensure launch Possible tests: 1. Coupled power ration. * May have repeatability problems WRT speckle. * May be appropriate for only LX. 2. Measured spot size with a CCD (beam view system). 3. Measured with an aperture-based system - TIA test methodology. What is conditional launch? - min and max amount of modes. - Mode group power distribution. [ Sufficient mode volume so that individual modes can not dominate fiber performance. * Need to excite azimuthal modes * This excludes ROFL type launches ] [ Discuss several ways to achieve conditional launch * CPR as verification??? SX -1-4, not 5 / LX-1-2 * Special patch cord not useful on SX ] Meet 8:30am 11 Nov 97 Tuesday Special Committee 8:30 - 11 am Jonathan: 260 meters is max we can get today. Jim: Can't achieve anything better then OFL. REQ: 1. Must achieve OFL bandwidth. 2. Min length = 260/440. Jim: What launch profile breaks the link? 1. ROFL 2. Split Link What launch conditions exclude this? What launches should you not do? Dan: Maybe we need to define what we shouldn't do? 1. Launch pure SMF into MMF. 2. ROFL 3. Etc Jonathan: No single mode group will have more then x% compared to average. Dan: Today, all LX parts are cat5 Excluding cat5 from LX, one launch condition is lost Excluding cat5 from SX, doesn't effect much. Jonathan: 1. No one disagrees. 2. ? Paul: Does this mean ROFL is no longer a need to test for? NO We could define CPR as a defined value and not by category. Try to test for CPR??? Not easy way yet. Peter: What difference between measuring spot size or CPR? Jim: Not sure, but is there a db range (cat3) within cat3 that won't work? Thinks maybe beam measurement will show failure. Jonathan: Is the methodology for specifying long and short wave LX and SX cases the same? David: CPR by itself may not be sufficient to determine if we met the definition. We may have to use mode field diameter. Peter: Both are the same, but [I think one has a standard test]. Finished 11am 11 Nov 97 Tuesday 11am meeting J Paul Benson Moty Goldis Gadi Lahat John Ritger Kirk Bovill Ed Grivna Arlen Martin Richard Seifert Dan Brown Atikem Haile-Mariam Jeff Martin Derek someda Ian Crayford Del Hansen Bob Mayer Steve Swanson David Cunningham Adam Healey Vince Melendy Tad Szostak Bob Dahlgren Todd Hudson Tremont Miao Jim Tatum Chris Diminico Dave Hyer Bob Musk Pat Thaler Alan Flatman Chris Karaguleff Mark Nowell Jonathan Thatcher Jay Garcia Don Knasel Paul Pace Todd VaFiades Mark Gerhold Paul Kolesar Jerry Pate Shelto Van Doorn Joel Goergen Dan Kuchta Bhavesh Patel Bill Verleggen Thomas Goergenson Gerard Kuyt Petar Pepeljugoski Theresa Watkins Peter Rautenberg Micheal Yam Don: Update on the special committee. Add meaning to conditional launch. 3.8.2 fiber type 62.5 2-260/440 50 2-550/550 What we owe is a more detailed overview to make this decision on what to do. ?: Is this a spec for manufacturing XCVR? Fiber? Patch Cords? Dan: We need to measure at TP2? Seems to be some debate here. ?: Someone selling 1x9 parts better meet either CPR and aperture based measurement (TIA defined). Jonathan: We should have unconditional launch and define it. Chris: Doesn't think we can get votes with it. Steve: Will we have to with LX xcvrs? If we defined conditional launch with LX that did not require a patch cord, does LX need to be two parts -> no. Jonathan: If SM launch is normative, and something has to be added for MM, the group would "go nuts" if it was normative. Del: Is this a motion? Break at 12 and resume at 1pm for comment resolution. Jonathan: We reviewed the comments, and with the exception of 2 resolutions, the group should accept. Note: The group accepted the committee's resolution except for #103 and #175. Comment #103: Del and Jonathan discussed the comment. Chris was happy with the result. Comment #175: Del and Jonathan discussed this coment. It took a while to get the wording, but they eventually got it. Motion: Jonathan moves to accept all changes of comment resolution list (with exception of 4 TR comments). Shelto seconds. Motion passed - no negative votes. Motion: For SX and LX parts John Ritger seconds SX: 1. Add row to SX transmit characteristics table 38.3 Description MMF Unit CPR(c) x we are limiting the length to the unconditioned xcvr. Jonathan hopes we could remove 'condition" from sx when we have gathered data. Vince agrees with Jonathan Chris ????? Somebody: 100/200 not acceptable. Does not like adding two launch conditions to already confusing 1x9 xcvr. Bob: Agree with short wavelength not good. Support conditioned launch normative. David: Thinks conditioned launch is best way to go. Jonathan: Wants unconditioned launch normative. Chris: Move forward with existing stuff and get better over time. Pat: We can not put 'TBD' number in the spec. Jonathan: If we take 'unconditioned' launch out of spec, we may not have parts for 6 months. Call the Question - Rich Sefert Second Technical motion 8 yes 14 no 11 abstain Motion Failed Motion: Jonathan: Move that only conditioned launch be specified in clause 38 for SX. Definition of 'condition' is the same. Second by Gadi Chris ??? Pat: Doesn't think this enforceable and doesn't think we should put this in the spec. Chris ??? Pat: Original motion does not imply implementation. Note added to motion: [no specifications normative or informative of conditional.] Defn conditioned: same as before. Steve: Call the question. 30 yes 2 no 10 abstain Technical motion passed Motion: Mark-> original motion that was tabled. Second John Rutger 1. Reviewed SX as before 2. LX: ???? Chris: Difficult for end user to deal with extra cable if required. Would want some way to help end user determine if they need a cable or not. Friendly amendment to add labeling back in. Paul: Friendly amendment to split motion into SX and LX rather then both at once. Call the question: ???? Voting only on SX portion of Mark's motion. 36 yes 0 no 4 abstain Technical motion passed Mark's LX Motion Pat: Paragraph just indicates a belief and does not require the external conditioner if it can be put inside. Chris: Offered word changes to not require an external launch. Vote: 29 yes 5 no 3 abstain Technical motion passed Pat: Does not like wording of what just passed. Motion: Paul Kolesar: Move launch conditioning shall be sufficient to allow the jitter budget to remain unchanged from draft 3.2. Second: Gadi The Jitter budget has not changed and suggests we leave it as is and fix things by the conditional launch. Mode conditioning should not force a change in the jitter numbers. Jonathan agrees. Chris: friendly amendment: add [due to mitigation of split impulse response (DMD).] on the end. Jonathan: Possibly because of potential need to change numbers, we may want to leave tp2 and tp3 open. Friendly amendment on words: [total] and [columns]. Now reads: launch conditioning shall be sufficient to allow the total jitter budget columns to remain unchanged from draft 3.2 due to mitigation of split impulse response (DMD). Vote: 33 yes 0 no 3 abstain Technical Motion Passed Del: Process from here on Mark: How much do we have to develop new content? None, but a small group needs to fill in some numbers. Four TRs on the table - put resolution against one and then resolve other three. Howie: 1. Take written statements and show commenters and make sure okay. 2. Paste up response. 3. Paste the exact wording of response. Paul: Won't withdraw until proof ??? Howie: Conditional withdraw ??? Chris: Conditional withdraw ??? Del: 1. Ed will upgrade document. 2. Get special committee together to fill in numbers. 3. Plan for next few months 4. Address 15 minutes tomorrow on clause 39. At 8:30am. 5. David to run special committee to fill in numbers. Finished 11 Nov 97 Tuesday 4:30pm special committee David Cunningham Don Knasel Vince Melendy Steve Swanson Chris Diminico Paul Kolesar Mark Nowell Jim Tatum Joel Goergen Dan Kuchta Petar Pepeljugoski Jonathan Thatcher Del Hansen Bob Mayer John Ritger Shelto Van Doorn Del: 1. CPR table. 2. Figure that needs to be sketched. 3. Motion put together. David: 1. CPR numbers 62, 50, SX, LX 2. Figures LX, SX mode-power distribution. 3. Receive filters. 4. Text for test methods. 5. Change fiber contributed jitter allocation DDJ[24ps], RJ[0ps]. Jonathan: Mark's motion only referenced CPR numbers and not cat 1 through cat 5 table. Informative statement to say we can lower filter bandwidth Jitter table 38-10 Ask we reconsider in deterministic jitter for tp2 - tp3. We put in 24ps rather then 0. This allows up to 24ps for deterministic, but is can be 0. Mark: CPR numbers: 50 MMF 62.5 MMF SX 9 - 24 9 - 29 LX 10 - 25 15 - 30 If jitter isn't sufficiently mitigated, the CPR numbers will have to shift up. Motion for test section: Petar: Specify informative upper limit for receive 3db electrical bandwidth to 1500mhz. [ What we would need to go into the standard: * Conditional launch measurement technique. * Definition of conditional launch * Defined parameter space for measured values. * How to ensure proper ??? in a MMF step index coupler. * Define the application space of conditioned vs. unconditioned launch. ] I had to leave. 12 Nov 97 PMD 8:30am Dan Brown Adam Healey Vince Melendy Walter Sotelo Phil Campbell Howie Johnson Larry Miller David Stacy Ian Crayford Myles Kimmitt Bob Musk Steve Swanson David Cunningham Don Knasel Mark Nowell Tad Szostak Chris Diminico Paul Kolesar Paul Pace Wen-Tsang Tang Jay Garcia Rick Lacerns Jerry Pate Jim Tatum Mark Gerhold Gadi Lahat Bhavesh Patel Jonathan Thatcher Joel Goergen Du-Tuan Lam Bob Payne Todd VaFiades Thomas Goergenson Tommy Leuing Petar Pepeljugoski Shelto Van Doorn Moty Goldis Andy LuQue John Ritger Bill Verheggen Ed Grivna Arlen Martin Mark Sankey Theresa Watkins Del Hansen Bob Mayer Derck Someda Micheal Yam Ed Grivna: Comment Resolution 192 non dis 194 non dis 193 non dis 191 non dis 21 non dis 22 non dis 20 non dis 23 non dis 24 non dis 25 non dis 182 non dis 183 non dis 6 discuss later - not part of motion 9 non dis 8 non dis 7 non dis clarification of 'loss' on 7 and 8 Motion: Del Hansen: To accept comment resolution Second: Ed Yes 22 0 no 5 abstain Motion passed Ed: Open TR in 3.1 comments. Documenting an assembly [Bob was okay with this]. Motion: Del Second: Ed 25 yes 0 no 0 abstain Motion passed Ed: Comment #6 Discussion on differential skew Motion: Del Second: Jonathan 10 yes 1 no 14 abstain Motion passed Ed is done Del: Present summary of special committee David: We should not put template in document Dan: FOTP-167 is single mode specific and we really don't have detail on MM. Motion: Jim: To remove graph from motion yesterday and replace with informative text [Mark's motion, #3] Second: Mark Does not agree to split into two parts. Dan: Amendment to make two issues. Second: Jonathan Withdraw amendment and Jim agrees to remove last line. Mark Agrees. 24 yes 1 no 7 abstain Motion passed Pause in note taking - Jonathan needed something from yesterday. Comment resolution #??? 38.9 Physical labeling SX supports MMF only [No Disagreement] Comment resolution #176: No objections to correcting by removing pics 6,7,8,9,10. Chris: What should go into the standard? - additional cabling section install done new installs - WCMB - ROFL - Launch as specified in 38.3 Del: Currently we have annex B. Chris can work this and get back to Del. Dan: We don't know today about ROFL in fiber. Motion: David: Add the following note below the CPR table specifications for conditional launch: Radial over filled launch should be avoided. Second: Petar Jim: Doesn't think it is enough - should add something that CPR may not find other bad launches Howie: ??? ??: Avoid using OFL comments at all. Maybe a far field spec rather then CPR. Jim: Forgot thought Jonathan: 'Breaks CPR'. There is possibly out there other ROFL that could break CPR tests. There is stuff that can break CPR that could cause problems in the right fiber. Jim: CPR may not be sufficient. Steve: Standard test methods for near field and far field. Friendly amendment: Howie: Add [while they may meet CPR test conditions] David agreed Friendly amendment: Chris Add Clause 38B.3 David agreed Jonathan: Why anyone feels there needs to be an empty box to accept information later? David: Identify what we are trying to catch: * small numbers of spots away from center * spots with significant power in center * small number of offset spots with significant power CPR test may not catch this. Motion reads: Add the following note below the CPR table specifications for conditional launch (clause 38b.3): Radial over filled launch, while they may meet CPR test conditions, should be avoided. 32 yes 2 no 2 abstain Motion passed Del: * conditioned launch * notes to be placed above two tables in the text * remove last line on Del's proposal Motion: Steve Second: Rich S Presented by Del: Statements on conditioned launch to be introduced in 38.3.1 page 38.7, line 6 and 38.4, page 38.10, line 6. Conditioned launch (CL) produces sufficient mode volume so individual MMF modes do not dominate fiber performance. This reduces the effect of peak to peak differential mode delay (DMD) between the launched mode groups and diminishes the pulse splitting and nulls in the frequency response. A CL may be produced by using a mode-conditioning -hybrid patch cord inserted at one or both ends of a duplex link between the optical PMD MDI and the cable plant. These hybrid patch cords contain a fiber of the same type as the cable plant (ie 62.5 or 50um fiber) connected to the optical PMD receiver input MDI and a specialized fiber/connector assembly connected to the optical PMD transmitter output. Examples of specialized mode-conditioning fiber/connector assemblies include a special step-index MMF for use at either wavelength or a generic ISO 11801 SMF with offset ferrule launch into the MMF cable for use with 1000base-LX. Some sources may produce CL directly and thus not require the use of external mode-conditioning patch cords. 26 yes 0 no 3 abstain Motion Passed 38.6.10 Coupled power ratio: Coupled power ratio (CPR) is measured in accordance with EIA/TIA-526-14A. Measured CPR values are time averaged to eliminate variation from speckle fluctuations. Motion: Dan Second: Petar We agreed to use this statement as lead into receiver 38-32: In clause 38.3.2. page 38, line 22, add note: To limit jitter, the receiver upper electrical 3db bandwidth should be less than 1500mhz. In clause 38.4.2, page 38.10, line 5A add a note. CPR ranges: 50 MMF 62.5 MMF SX 9 - 24 9 - 29 LX 10 - 25 15 - 30 Passes - no objections Howie Johnson: We need to update a clause: Differential sensitivity: 1. Minimum differential sensitivity is the smallest value of peak to peak differential (ppd) amplitude at which a receiver is expected to operate, under worst case conditions, without exceeding the specified BER. 2. Maximum differential sensitivity is the largest value of peak to peak differential (ppd) amplitude at which a receiver is expected to operate, under worst case conditions, without exceeding the specified BER. Motion: Adopt all of motions up through 11am, 12 Nov 97 and updated comment data base as per resolving 4 TR comments. First: Paul Second: Don Passed - no objections Del: Completed task at hand. Motion: Chris: Provide note C to table 38-12- Modal Bandwidth ( ) C) Annex B. Provides additional information on modal bandwidth and launch condition. Chris will work with Howard and Geoff on this issue: Follow up letter to TIA regarding procedure for WCMB as described in annex B. - launch conditions as specified in 38.3 At present, there is no linkage to annex B Second: Shelto Technical motion: 28 yes 0 no 3 abstain Motion Passed Motion: Split spec LX- MMF / LX- SMF First: Arlen Martin Second: Dan Kuchta Howard suggests we table this for discussion at 802.3z this afternoon. Motion???? Table 7 yes 11 no 15 abstain Motion failed Howie: Wants to keep 2 parts to lower number of combinations that won't work when plugging in fiber. Bob: We could make cheaper parts if we split LX part. Dan Agrees. Howie: The group was directed to move ahead with two parts. Jonathan: Call the question. 18 yes 16 no 2 abstain Motion Failed. Meeting done