Minutes of X3T10 SPI-2 LVDS Study Group Meeting - 10/13/95

Lohmeyer, John JLOHMEYE at cosmpdaero.ftcollinsco.ncr.com
Sat Oct 14 17:12:00 PDT 1995


Minutes of SPI-2 LVDS Study Group Meeting 
                      X3T10/95-337r0

Accredited Standards Committee*
X3, Information Technology
                                                    Doc. No.: X3T10/95-337r0
                                                        Date: October 14, 
1995
                                                     Project:
                                                   Ref. Doc.:
                                                    Reply to: John Lohmeyer

To:         Membership of X3T10

From:       John Lohmeyer, Chair X3T10

Subject:    Minutes of X3T10 SPI-2 LVDS Study Group Meeting
            Denver, CO -- October 13, 1995


                                       Agenda

1. Opening Remarks

2. Approval of Agenda

3. Attendance and Membership

4. LVDS Terminator Specifications [Aloisi / Ham]
   4.1 DIFFSENS Operation [Ham]

5. LVDS Driver/Receiver Specifications
   5.1 Review of Power Budget Documentation

6. LVDS Cable Specifications
   6.1 Crosstalk testing [Spitler]
   6.2 The Case for 50 Ohm Cables (95-308r0) [McCall]

7. LVDS Testing Results [Ham]

8. Other Topics
   8.1 Hybrid Signaling? [McGrath]
   8.2 Should RST Signal Remain Single-Ended? [Gardner]
   8.3 Document review (95-315) [Ham]

9. Meeting Schedule

10. Adjournment


                              Results of Meeting

1.    Opening Remarks

John Lohmeyer, the X3T10 Chair, called the meeting to order at 9:00 a.m.,
Friday October 13, 1995.  He thanked Symbios Logic for hosting the meeting.

As is customary, the people attending introduced themselves and a copy of 
the
attendance list was circulated.

2.    Approval of Agenda

The draft agenda was approved with the addition of the following items:

      4.1 DIFFSENS Operation Review [Ham]
      8.3 Document Review (95-315) [Ham]

3.    Attendance and Membership

Attendance at working group meetings does not count toward minimum 
attendance
requirements for X3T10 membership.  Working group meetings are open to any
person or organization directly and materially affected by X3T10's scope of
work.  The following people attended the meeting:

         Name          S        Organization         Electronic Mail Address
 ---------------------- -- ------------------------- 
 -------------------------
Mr. Norm Harris        P  Adaptec, Inc.             nharris at eng.adaptec.com
Mr. Dennis R. Haynes   P  Burr-Brown Corp.          haynes_dennis at bbrown.com
Mr. Louis Grantham     P  Dallas Semiconductor      grantham at dalsemi.com
Dr. William Ham        A# Digital Equipment Corp.   ham at subsys.enet.dec.com
Mr. Andy Chen          A# Fujitsu Computer Prods    achen at fcpa.fujitsu.com
                          Amer
Mr. Richard Greenberg  V  IBM Corp.
Mr. Dean Wallace       P  Linfinity Micro 
          75671.3443 at compuserve.com
Mr. Edward A. Gardner  V  Ophidian Designs          gardner at acm.org
Mr. Ting Li Chan       A  QLogic Corp.
Mr. Clinton Wong       V  Quantum Corp.             cwong at asic.qntm.com
Mr. Richard Uber       V  Quantum Corp.             duber at tdh.qntm.com
Mr. Henry Wong         V  Quantum Corp.             hwong at asic.qntm.com
Mr. Brian N. Davis     A# Seagate Technology 
       brian_davis at notes.seagate
                                                    .com
Mr. Erich Oetting      P  Storage Technology Corp. 
 Erich_Oetting at Stortek.com
Mr. John Lohmeyer      P  Symbios Logic Inc. 
       john.lohmeyer at symbios.com
Mr. Tracy Spitler      V  Symbios Logic Inc. 
       tracy.spitler at symbios.com
Mr. Kevin Bruno        V  Symbios Logic Inc.        kevin.bruno at symbios.com
Mr. Richard Mourn      V  Symbios Logic Inc. 
       richard.mourn at symbios.com
Mr. Kevin Gingerick    V  Texas Instruments         4307725 at mcimail.com
Mr. Paul D. Aloisi     P  Unitrode Integrated       Aloisi at uicc.com
                          Circuits
Mr. Tak Asami          A  Western Digital           asami at dt.wdc.com
                          Corporation

21 People Present

Status Key:  P    -  Principal
             A,A# -  Alternate
             O    -  Observer
             L    -  Liaison
             V    -  Visitor

4.    LVDS Terminator Specifications [Aloisi / Ham]

Paul presented the terminator circuit from 95-269r5.  He pointed out that 
the
circuit shown is ideal and actual implementations would be slightly 
different.
He said that he sees no problems in implementing the terminator and expects 
to
have samples in Spring 1996.  Paul's terminator circuit included the ability
to switch between single-ended, LVDS, and off (high impedance).

There was some discussion about whether a simpler three-resistor circuit 
might
be adequate for many applications.  Dennis Haynes presented a three-resistor
terminator circuit that used 1.14k ohm bias resistors from +3 volt and 
ground
to a 120 ohm shunt resistor.  Several refinements were suggested, including
the need to regulate the 3 volt input.

Paul noted that a three-resistor circuit would use more steady-state power 
and
would be difficult to make in a switchable variety.  There was another 
concern
about the constant bias current fighting the drivers.  Kevin Gingerick 
pointed
out that the constant bias current does not factor into the initial
transition; it only affects steady-state signal levels.

Eventually, the two terminator approaches were labeled the non-linear and
linear alternatives.  It was agreed that both alternatives should be
permitted.  Dennis Haynes presented several foils describing a general 
circuit
model of a terminator that accommodated both alternatives.  John Lohmeyer
suggested that the draft standard should document a set of V-I curves that
would contain all legal terminators.  The discussion was interrupted to re-
visit the signal and noise budget after it was noticed that there had been 
an
error in our previous calculations (see item 5.1).

After resolving the above budget issue, Dean Wallace sketched the V-I curves
that bound the allowable terminator characteristics (a modified bowtie 
shape).
He accepted an action item to create an electronic version of the V-I 
diagram.
It was agreed that the terminator common mode voltage reference needs to be
1.25 volts +/- 0.05 volts through a 150 Ohm +/- 20% resistor.

      4.1   DIFFSENS Operation [Ham]

Bill Ham reviewed the DIFFSENS circuit to clear up any misconceptions that 
may
have existed from the recent reflector messages. Single-ended devices always
ground the DIFFSENS line (and do not examine it). High-powered differential
devices (RS-485) pull up the DIFFSENS line with 1k Ohm resistors and disable
their drivers if the DIFFSENS line stays low (that is, one or more single-
ended devices are connected).

LVDS terminators attempt to hold the DIFFSENS line at 1.3 volts.  If the
DIFFSENS line stays within the voltage range of 0.7 volts to 1.9 volts, then
there are no single-ended or high-power differential devices on the bus; the
LVDS devices and terminators are enabled to operate in LVDS mode.  If the
DIFFSENS line is above 1.9 volts, then all LVDS devices and terminators are
disabled.  If the DIFFSENS line is below 0.7 volts, then the LVDS devices 
and
terminators either disable themselves or switch to single-ended mode,
depending on their capabilities.

The above DIFFSENS rules are robust, avoiding physical damage and permitting
devices to be implemented that autoswitch between single-ended and LVDS
operation.

5.    LVDS Driver/Receiver Specifications

      5.1   Review of Power Budget Documentation

The previous signal budget calculations did not allow enough voltage for 100
mV receiver sensitivity.  Bill Ham presented several possible new budgets:

Min Signal       Driven     Xtalk         Term       Receiver     Net
Current (mA)     Signal     Budget        Bias       Sensitivity  Margin
(half driver)    Level(mV)  (mV)          Level(mV)  (mV)         (mV)
 ---------------  ---------  ------------  ---------  -----------  --------
  2                 240        0             150        100         -10
  2                 240        0             150         50         +40
  2                 240       50             150         50         -10
  3                 360       50             150         50        +110
  2.5               300       50             150         50         +50
  2.5               300       50             150         70         +30

  2                 240       50             100         50         +40
                              50             100         50           0

  2.2               264       50             120         50         +24
  2                 240       50             120         25 (h)     +25

  2.2               264       35 (driven)    120         70         +39
  3 (neighbor line) 360      n/a (undriven)  120         70         +50

  2.2               264       35 (driven)    120         70 (h)     +15
  3 (neighbor line) 360      n/a (undriven)  120         70 (h)     +75

  2.2               264       35 (driven)    130         70         +29
                               0 (undriven)  100         70         +30

(h)          =  25 mV hysteresis receiver.
(driven)     =  line is driven
(undriven)   =  line is not driven, except by the terminator.  (Such lines
                were also called 'victims' because they are vulnerable to
                crosstalk from neighboring lines.)

It was noted that the only victim line during normal SCSI protocol is RST 
(for
flat ribbon cables).  Since nearly all protocol chips have glitch filtering 
on
RST, we may not need to worry too much about the noise margin on victim 
lines.
The last two lines in the above table represent the eventual agreements
reached during the meeting.

Signal budget agreements du jour:

      Receiver sensitivity:  70 mV (no hysteresis)
      Driver current:  4.4 mA minimum, 6 mA maximum
      Terminator bias:  130 mV maximum, 100 mV minimum
      Cable characteristic impedance: 120 Ohm minimum, 150 Ohm minimum
      (We still need to create a cable crosstalk specification.)

6.    LVDS Cable Specifications

      6.1   Crosstalk testing [Spitler]

Tracy Spitler presented 95-336r0 which summarizes his crosstalk testing of
mixed single-ended and LVDS signals on various cables.  He also tested LVDS 
to
LVDS crosstalk on both flat and twisted cables.  The conclusion was that
hybrid signals would result in unacceptable crosstalk from single-ended
signals into adjacent LVDS lines, even with twisted-pair cables.  LVDS to 
LVDS
crosstalk on flat ribbon cables is moderately high.  Flat ribbon cables 
should
not be allowed in the SPI-2 standard for LVDS, but it may work in many
configurations.

      6.2   The Case for 50 Ohm Cables (95-308r0) [McCall]

It was noted that 50 Ohm cables would further aggravate the signal budget
problems we addressed under item 5.1.  Since David McCall was not present,
this item was deferred until a subsequent meeting.

7.    LVDS Testing Results [Ham]

Bill Ham briefly reviewed the testing results he had shown at the Bedford, 
NH
meeting in September.  Bill had some new data showing that he had achieved
good margins at Fast-100 speeds using a point-to-point configuration of over
27 meters cable length.

8.    Other Topics

      8.1   Hybrid Signaling? [McGrath]

Based on the crosstalk testing results in 95-336r0 (see item 6.1), the study
group recommends that hybrid signaling not be done.

      8.2   Should RST Signal Remain Single-Ended? [Gardner]

Ed Gardner asked that this agenda item be removed from consideration.

      8.3   Document review (95-315) [Ham]

Bill Ham reviewed 95-315r0 and plans to prepare a revision 1 for the Palm
Springs meeting.

9.    Meeting Schedule

The next meeting of SPI-2 Study Group will be Monday November 6, 1995, in 
Palm
Springs, CA at the Hyatt Regency Suites Palm Springs (619-322-9000), hosted 
by
Western Digital Corp.  Another SPI-2 Study Group meeting is scheduled for
December 15, 1995 in San Jose, CA hosted by Quantum.

10.   Adjournment

The meeting was adjourned at 4:15 p.m. on Friday October 13, 1995.




More information about the T10 mailing list