p1394b Working Group Meeting,
April 30, 1998,
Rockwell facility,
Newport Beach, California,

chairman:       David Wooten,
		david.wooten@compaq.com

vice-chairman:	Michael D. Johas Teener
                mike@zayante.com

editor:         Eric Hannah,
		ehannah@mipos2.sc.intel.com

secretary:      Richard Churchill,
		richard.churchill@compaq.com

 1. Introductions
 2. Review of Agenda
 3. Review of Minutes
 4. Procedures
 4.1  Voting
 4.2  Prices and Pricing
 4.3  Call for Patents
 5. Task Group Reports
 5.1  UTP - Colin Whitby-Strevens
 5.2  Upstarts - Colin Whitby-Strevens
 5.3  BPort - Alistair Coles
 5.4  Standard Electrical - Eric Hannah
 5.5  Accelerations - Dave LaFollette
 5.6  Cable/connector - Max Bassler (by Bill Northey)
 5.7  Simulation - Prashant Kanhere (by Jerry Hauck)
 5.8  PHY-LINK - D.C. Sessions
 6. Additional topics
 6.1  Isochronous Resource Allocation and Reallocation
 6.2  1394a Power Distribution
 7. Liaison
 7.1  Fiber Channel - Colin Whitby-Strevens
 7.2  DAVID - David Wooten
 8. Future meetings
 8.1  June, St. Petersburg Beach, Florida
 8.2  July, Bath, UK
 8.3  August, Portland, Oregon
 8.4  September, Chicago, Illinois
 8.5  October, Maui, Hawaii
 8.6  December, ??, ??
 8.7  January, ??, ??
 9. Action Items
10. Adjournment

 1. Introductions

    The meeting was convened at 1:20 pm, and all attendees introduced

 2. Review of Agenda
 3. Review of Minutes
 4. Procedures
 4.1  Voting
 4.2  Prices and Pricing
 4.3  Call for Patents

 5.1  UTP - Colin Whitby-Strevens

    P1394b UTP5 Task Group April 98
    1. Loss budget calculations (Alistair Coles)
      - allows reach of 57m (see next item)
      - propose to lower launch and sensitivity (50mV) to assist Class B
        emissions requirements
    2. Reach issue
      - may not have a good solution for installed base or expectations of
        cable installation specification groups
      - limitation is both loss budget and passive equalization
      - propose to liaise with cable wiring specification groups
    3. Crosstalk measurements (Bill Northey)
      - pin choices look good
      - more testing promised
    4. Review of the current specification and open issues:-
      - Eye diagram mask for the transmitter signal at the RJ45 connector
        - based on shorthaul elec spec with an allowance for transformer, etc.
        - min rise time - based on short haul elec (still tbd)
      - Jitter budget
        - need simulations and proposal
      - Testing
        - informative spec on the electrical spec at various reference points
        - normative spec at the connector (TX and RX)
        - diagrams from teh short-haul electrical section
      - How to spec equalization - pic+em
        - in the system (after the connector), not in the cable
        - informative section on equalizer spec/example design - action AC/EH
      - Twist in a twisted pair
        - solved logically after scramber synch
      - Problem may be constrained by installed base
        - Lengh Demands
        - may have to liaise on cost/length/emissions trade-offs
        - may not have a solution which works well in installed base

    [Should we raise the issues with regard to cost and distance with VESA
    and DAVIC?  For the home market, we must meeting FCC class B requirements
    and not the class A ones for commercial, and so it is easier for them to
    reach 100 m.  VESA and DAVIC might remove 1394 from there standards if we
    cannot reach something close to 100 m, but we don't know.  They want the
    maximum reach of the commercial solution, and may not be fully aware of
    the problems associated.  We need to understand their process of arriving
    at the 100 m limit, if we are to make an informed decision.  Perhaps we
    need to send a letter to Bob Jansen about the point?  We might decide to
    not do CAT-5, if we have problems passing emissions requirements. ... We
    may lose something if we don't support what people expect ... The same
    debate went on in VESA HN ... One person surveyed over 2000 homes in
    North Carolina, and one new home designs, and found no runs of greater
    than 50 m.  Thus setting a limit of 50 m may not be a problem.  We can't
    cover everyone ... Teener - This is not an appropriate question.  We have
    to tell them what we have, and the response from them is likely to be
    either accepting, or they might tell us that X has a 100 m solution that
    passes class A.  If we don't say anything, we won't know, so let's just
    talk to them.  Might we have two standards -- one for class A, and one
    for class B?  No-one has told these groups that there are trade-offs ...
    Folks doing apartment houses will claim they need more than 50 m, but ...
    If you want fixed equalization, you can't do 0 to 100 m.  What if we did
    not require fixed equalization?  Wooten -- Perhaps we can have a variable
    equalization and a kit ...]

 5.2  Upstarts - Colin Whitby-Strevens

    Upstarts - April 98
    1. Welcome, Introductions
    2. Review of D0.10
      - C code for modified tone proposal now in the draft
      - modification includes positive acknowledgement to complete the
        negotiation
    3. Tone frequency (NEC presentation)
      - optical transceivers probably need something above 30 MHz
      - go for 49 MHz - better for EMI- and related times
      - but is 61 MHz easier?
        - decided to specify as 61.44 MHz
      - may specify a range?
    4. Open Issues
      - may not need the bias filter (at least during start up)
      - may not need to do the TpBias phase repeatedly
      - need to check that TpBias interval does not cause false p1394a detect
      - Experience of Fiber Channel start-up
        - high impedence or static transmitter may produce valid-looking
          signal at the far end
      - Check the effect of the external p1394a electrical components
        - time to charge up the capacitor on TPA - may need to spec it
          carefully now thought to be OK
        - avoid false TpBias detect from the input biasing on an optical
          transceiver

    [How many meetings do you think you will need?  Certainly the next and
    the one after ...]

 5.3  BPort - Alistair Coles

    1. New synchronization procedure - explanation of how synchronization works
       when thenew control coding is used (i.e. the scheme presented by Eric
       Deliot at the last meeting).
    2. Do we adopt changes to control scheme:
       Packet delimiters -> Cz codewords
       Request types -> Dx.y symbols
    3. ***
    ***

 5.4  Standard Electrical - Eric Hannah

    Cable Interface
    *** diagram ***

    [ ... We need a bias point set, but we have no pin or bias generator.
    So how do we do it?]

    Toy Model
    *** diagram ***

    [ ... You see a mode conversion here, with good matching ... There is a
    second possibility -- Noise may be converted into a differential level.
    ... We can do it other ways.]

    DC Bias Parameters
    [The sense of the last p1394b meeting was that we should have only two
    biasing conditions.  The first requires 3.3V logic and is required for
    bi-lingual mode operations.  The second is for beta only operation, and
    allow interoperability ... ]

 5.5  Accelerations - Dave LaFollette

    Accelerations Task Group Summary 4/30/98
    C code review
      some code written
        to be added to spec
        more code to be written
        notation is not final
          coordinate with bport and upstarts
    FIFO centering, synchronoization
      to be handled in bport code?
      node arbitration code will all operate in one clock domain

    Mixed bus discussion
    *** diagram ***
    Any 'a' nodes hurt bus performance

    [Mixed environments will not operate as efficiently as 'b'-only ones.
    ... Aren't sure how much effort we will put into the acceleration of the
    mixed environment situation ... If you hook two B nodes together, and
    hang two A nodes on one of the B nodes,, what are the clouds? ...]

    Acceleration group direction
    Continue beta-only arbitration C code
    Plan for limited mixed bus support
    propose Tree-ID, Self-ID modifications to identify beta-only buses
    more reflector traffic
      mixed buses
      self-id ...
      ...

 5.6  Cable/connector - Max Bassler (by Bill Northey)

    Agenda
    - Review and approve of last meeting's minutes
    - Review and vote on Bulk Cable Methodlogy
    - Review of past presentations
    - New items
      - presentations
      - develop on-going action items
        - Mr. Saito - proposal for cable keying

    [What about cables at higher speeds? ... We may be able to get by with an
    S400 cable at S800, and keying only with S1600 and beyond.  How do we make
    sure we have a 'b' bus manager? ...]

 5.7  Simulation - Prashant Kanhere (by Jerry Hauck)

    (No slides, comments only.)

    [Plan to start with the products of working groups that finish first.
    As the Startups plan to finish soon, they may be the first to be done,
    which seems appropriate ... Hope to have the first results back for the
    St. Petersburg meeting. ...]

 5.8  PHY-LINK - D.C. Sessions

    (No slides, comments only.)

    [We are looking to keep the interface narrow, upping the clock as we move
    to higher speeds.  Use reflected wave signaling, and allow the board
    implementor to choose the impedence ... Have some "help required" items
    for filling in numbers and tables. ... John F., how do you feel about
    taking on the changes in the PHY-LINK protocols to document the changes?
    Not sure.  Could you take a stab at documenting your discussions?  Yes ...
    Colin, you know what question you want to answer?  Everyone here has the
    responsibility to produce any questions they want answered regarding the
    PHY-LINK interface.

 6. Other topics

 6.1  Isochronous Resource Management Concerns - Richard Churchill

    <Richard Churchill lead a brief discussion of problems with the process 
    of allocating and re-allocating isochronous channels and bandwidth, with
    emphasis on behavior following a bus reset.  Though the present methods
    can result in very unfortunate behaviors, these are not very likely.  As
    they are non-deterministic, these behaviors are highly problematic for
    some applications.  The problems can be substantially reduced or eliminated
    in many or all cases by the changes recommended.  A paper explaining the
    nature of the problems, and the recommendations for resolving them, was
    promised for the reflector and group web and ftp sites.>

 6.2  1394a Power Distribution - Steve Bard

    <Steve went through the slides at a pace that made it impossible for the
    secretary to record their contents.>

    Power distribution

    1394 current limiting

    1394a current limiters

    1394a node types

    1394 power configuration (1)

    1394 power configuration (2)

    1394 power what UL sees:

    [The discussions were very interesting, with lots of interesting comments
    by the attendees.  We have a problem ... 90 Va with the majority of the
    world, and 50 Va in Scandinavia.  There are a number of 1394 device
    vendors seeking to be 'accessory class devices.'  There is a problem with
    "self extinguishing enclosures" and mixing AV equipment that is not UL
    etc. compliant, and PCs, etc., which are.]

 7. Liaison

 7.1  Fiber Channel - Colin Whitby-Strevens

    Fibre Channel Informal liaison
    T11
    1. Presented "modified 8B10B"
       to T11.2 Cu Working Group (22 Apr) 98-166.pdf
       at http://www.t11.org, go to 'docs' but avoid the link on left, and
       use the one on the right.
    2. Requested speed negotiation presentation next meeting
    3. T11.2 decided to start up a '10 bit interface' report group for
       anticipated 2 GBaud and 4 GBaud - They note our and GE activities

    [They are concerned about EMC and such, including signal integrity, etc.
    across their 10-bit interface. ... It is good that we will be co-locating.
    ...]

 7.2  DAVID - David Wooten

    [This liaison letter was received by me from DAVIC.]

	"SOURCE:	DAVIC Physical Layer Technical Committee
	TITLE:	Request for consideration on long distance 1394 breakage
					_______________

	"LIAISON STATEMENT

	"TO:		IEEE P1394b Working Group
	APPROVAL:	Agreed to at DAVIC Pleanry  meeting
	FOR:		Action
	CONTACT:	Larry  Behmer, PHY TC Chair	Tel:	+1-303-965-8089
	US West	email:	lbehmer@uswest.com

	"The DAVIC Physical Layer Technical Committee is currently planning to
	utilize IEEEP1394b technology for its Home Network.  During our delib-
	erations on this decision concern has been expressed regarding the 
	ability of the 8B10B coding to address our needs for long distance 
	multi-hop arrangements.  As a result DAVIC requests IEEEP1394b eval-
	uate whether the 1394 bus breaks at S100 with 8B10B coding by exceed-
	ing the gap count(max 63) under the following models in the figure.  
	DAVIC expects that long distance hops below should be accommodated by 
	a DAVIC Home Network.

	*** diagram with text showing assumed behavior of 8B/10B and 4B/5B ***
        *** encoding schemes in long-haul configurations 		   ***

	"We request you to consider these models and find a resolution for the 
	robustness of Home Network at S100.  We  look forward to ongoing con-
	tact with your Working Group and would like a response by June 1."

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

	"http://www.davic.org
	13 April 1998

	"To:	David Wooten
		Chair
		IEEE P1394b Working Group
		David.Wooten@COMPAQ.com

	"Dear Sir,

	"Attached please find a liaison statement from DAVIC which was approved 
	by the DAVIC meeting held in Milan during March 9-13 1998.

	"Your kind consideration is requested.

	"Yours sincerely,

	"H. Yasuda,
	President and Chairman of the Board, DAVIC"

    [Indications are that the delays claimed by the 4B/5B presentation can
    be matched or bettered by an 8B/10B implementation, and so if the 4B/5B
    solution was sufficient to DAVIC's needs, then p1394b with 8B/10B should
    work too.]

    Problems with 4B5B encoding
    - Runlength over 60 bits (due to scrambling after encoding)
      - synchronization more difficult
      - baseline wander (distortion)
      - receiver sensitivity penalty
    - Disparity not constrained
      - further aggravates baseline wander and receiver penalties
    - Inadequate control space
      - needs 10 bits just like 8B/10B
      - p1394a?
    - MLT-3 costly to implement and counter productive
      - have to increase launch voltage
    - Poor robustness for control codes
      (c/f 8B/10B Hamming distance properties)
    - No scalability to higher data rates
      - need common solution for all speeds
    - Not clear that 4B/5B can do all we need (overlapped arbitration, etc.)

 8. Future meetings

    Details about the various future meetings are posted as they are made
    available in various locations, including the 1394 TA Upcoming Events
    page (http://1394ta.org/upevents/upevents.html), and the p1394b group
    home page (http://www.zayante.com/p1394b/index.shtml).

 8.1  June, St. Petersburg Beach, Florida; host - Chuck Brill

    We will be meeting at the Best Western Serata Beach Resort, in St.
    Petersburg Beach, Florida, on Thursday and Friday June 11 and 12.
    Reservations may be made by calling 1-800-344-5999, or by calling
    the hotel directly at 813-367-2771.  Working group meetings will be
    held all day Thursday, plus Friday morning.  The working group plenary
    session will be held Friday afternoon.

    This meeting is being held in conjunction with the TA PCWG, p1394.1 and
    p1212r working groups, which will meet at the same site on Tuesday and
    Wednesday.  The June T11 meeting is being held one block away on Monday
    through Thursday.

 8.2  July, Bath, UK; hosts - Colin Whitby-Strevens and Alistair Coles

    July 13 and 14, Monday and Tuesday, we will meet at the Hilton National
    Hotel, in Bath, UK.  Reservations should be made directly with the hotel
    by calling +44 1225 463 411, or sending a fax to +44 1225 464 393.  You
    must identify yourself as attending the "IEEE Meetings" to obtain the
    preferred room rate (which is substantially lower than the normal rate)
    and to assure that attendee room-nights are created to the group.  Also,
    send confirmation of your participation dates, and your accommodation
    dates, to Colin (colinws@bristol.st.com), so that adequate catering may
    be provided and all attendee room-nights properly credited to the group.

    This meeting is being held in conjunction with the p1394.1 and p1212r
    working groups, which will meet on Wednesday, Thursday and Friday of
    this week.

    For further details, please see the note from Colin at the group web site
    (http://www.zayante.com/p1394b/meeting9807/cws980310-bathmeet.pdf).

 8.3  August, Portland, Oregon; host - Steve Finch

    This meeting will be held in Portland, Oregon, on August 18 and 19, in
    conjunction with the p1394.1 working group, which will meeting on the
    following two days.

    Details are not yet available.  Information will be posted as it is made
    available.

 8.4  September, Chicago, Illinois; host - Max Bassler

    This meeting will be held in Chicago, Illinois, on September 10 and 11, in
    conjunction with the p1394.1, (PCWG ?) and perhaps p1212r working groups,
    which will meet on the previous two days.

    Full details are not yet available, but some interesting possibilities for
    after-hours recreation have been mentioned, so watch for details as they
    become available.

 8.5  October, Maui, Hawaii; host - Farrukh Latif

    This meeting will be held at the Sheraton Maui resort on Kaanapali Beach,
    near Lahaina, Maui, Hawaii, on October 12 and 13.  The p1394.1 and p1212r
    working groups will also meet at the same site on October 15 and 16, with
    one day separating their meetings from the p1394b meetings.

    This is an excellent hotel, with many exceptional features.  Look for full
    meeting details in the near future.

 8.6  December, ??, ??

    We have one tentative invitation for December, and thus have no firm
    location or dates.  Pending finalization, we are still soliciting further
    invitations.

 8.7  January, ??, ??

    We have one tentative invitation for January, and thus have no firm
    location or dates.  Pending finalization, we are still soliticing further
    invitations.

 9. Action Items

    One action item was taken by John Fuller regarding interface and protocol
    issues.

    David Wooten took an action time to respond to the DAVIC liaison letter
    on behalf of the group as a whole.

    Richard Churchill took an action item to publish a paper discussing the
    IRM behavior as it stands, and suggestions for changes.

    Steve Bard took an action item to provide more information regarding
    regulatory and safety issues related to the power bus.

10. Adjournment

    The meeting was adjourned at approximately 4:40 PM.

===============
Attendance List

Steve Bard
steve.bard@intel.com
503-264-2923

Max Bassler
mbassler@molex.com
630-527-4490

Hans Bjorklund
hans.bjorklund@sepow.mail.abb.com
+46 240 782563

Dave Brown
dave_brown@3com.com
408-764-6334

Don Chambers
chambersd@jae.com
714-753-2600

Dao-Long Chen
dao-long.chen@symbios.com
970-223-5100 x9461

Richard Churchill
richard.churchill@compaq.com
281-514-6984

Alistair Coles
anc@hplb.hpl.hp.com
+44 117 922 8750

Michele Dale
mzdale@lucent.com
610-712-3294

Firooz Farhoomand
firoozf@ix.netcom.com
408-653-4059

John Fuller
jfuller@microsoft.com
425-703-3863

Eric Hannah
ehannah@mipos2.sc.intel.com
408-765-4441

Jerry Hauck
jerry@macrodesigns.com
510-668-1006

Daisuke Hiraoka
dai@sm.sony.co.jp
+81 3 5448 4603

Ku Janam
kujanam@samsung.co.kr
+82 331 200 4755

David Johnson
dkjohnson@ti.com
972-480-3632

Akihito Kuwabara
akihito@sj-pceg.ccgw.nec.com
650-528-5905

Dave LaFollette
dlafolle@mipos2.sc.intel.com
408-765-2587

Farrukh Latif
flatif@lucent.com
610-712-7546

Paul Levy
paul.levy@vlsi.com
602-752-6382

Jun-ichi Matsuda
matsuda@ccm.cl.nec.co.jp
+81 44 856 2082

Reza Moattar
reza.moattar@tsc.tdk.com
714-508-8731

Jim Nelson
nelson@laya.nrtc.northrop.com
562-948-8357

Bill Northey
northewa@bergelect.com
717-938-2119

Ozay Oktay
ozay_oktay@cissc.canon.com
714-856-7180

James Piccione
james.piccione@smi.siemens.com
408-895-5136

Dennis Rehm
dennis.rehm@symbios.com
970-226-9602

Kyozo Saito
kyozo_s@gw3.alps.co.jp
+81 229 23 5111

Tomoki Saito
saito@ccm.cl.nec.co.jp
+81 44 856 2082

Brad Saunders
bradley.saunders@rss.rockwell.com
714-221-6513

D. C. Sessions
dc.sessions@vlsi.com
602-752-6545

Masood Shariff
mshariff@lucent.com
732-957-5479

John Smolka
jsmolka@ti.com
972-480-2094

Ron Soderstrom
rows@vnet.ibm.com
507-253-6290

John Ta
john.ta@tus.ssi1.com
714-573-6957

Ken Taylor
foken@aolcom
904-427-4311

Michael Johas Teener
mike@zayante.com
408-461-4901

Colin Whitby-Strevens
colinws@bristol.st.com
+44 1454 611500

Jonghwa Won
jhwon@secns.sec.samsung.co.kr
+82 331 200 3353

David Wooten
david.wooten@compaq.com
281-518-7231

Niwa Yoshikatsu
niwa@sm.sony.co.jp
+81 3 5448 4603

Patrick Yu
patrick_yu@el.nec.com
408-588-5436