Minutes of X3T10 SPI-2 LVDS Study Group Meeting - 8/14/95
Lohmeyer, John
JLOHMEYE at cosmpdaero.ftcollinsco.ncr.com
Tue Aug 15 11:14:00 PDT 1995
Minutes of X3T10 SPI-2 LVDS Study Group Meeting
X3T10/95-309r0
Accredited Standards Committee*
X3, Information Technology
Doc. No.: X3T10/95-309r0
Date: August 15, 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 -- August 14, 1995
Agenda
1. Opening Remarks
2. Approval of Agenda
3. Attendance and Membership
4. Proposed Agreements [Aloisi]
5. Proposal for SPI-2 Electrical Interface (95-307r0) [Ham]
6. The Case for 50 Ohm Cables (95-308r0) [McCall]
7. LVDS (95-269r3) [Aloisi]
8. Single-ended Reset Signal [Gardner]
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.,
Monday August 14, 1995. He failed to thank the host (himself).
As is customary, the people attending introduced themselves and a copy of
the
attendance list was circulated.
2. Approval of Agenda
An agenda was prepared and approved.
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 Organization Email
John Lohmeyer Symbios Logic john.lohmeyer at symbios.com
Ed Suder Texas Instruments ROBO at msg.ti.com
Kevin Gingerick Texas Instruments 4307725 at mcimail.com
Todd Nelson National ctwnsc at tevm2.nsc.com
Semionductor
Dennis Haynes Burr-Brown haynes_dennis at bbrown.com
Paul Aloisi Unitrode Aloisi at uicc.com
Brian Day Symbios Logic brian.day at symbios.com
Tracy Spitler Symbios Logic tracy.spitler at symbios.com
Frank Gasparik Symbios Logic farnk.gasparik at symbios.com
Mark Jander Symbios Logic mark.jander at symbios.com
Brian Davis Seagate davis at cdg.seagate.com
Wayne Werner AT&T aluxpo!wew at att.com
Chris Millsaps Buslogic chrism at buslogic.com
Michael Smith Dallas msmith at dalsemi.com
Semiconductor
Louis Grantham Dallas grantham at dalsemi.com
Semiconductor
Richard Uber Quantum duber at tdh.qntm.com
David McCall Quantum dmccall at tdh.qntm.com
Steve Robalino DPT robalino at dpt.com
Wally Adaptec wally at eng.adaptec.com
Bridgewater
Bill Ham Digital ham at subsys.enet.dec.com
Edward A. Quantum gardner at acm.org
Gardner
Vit Novak Sun Microsystems vit.novak at sun.com
4. Proposed Agreements [Aloisi]
John Lohmeyer presented Paul Aloisi's foil outlining the proposed agreements
to date on Low Voltage Differential Signaling. The foil is paraphrased
here:
* A single chip can work for both single-ended (Async and Fast-5
through
Fast-20) or LVDS (Async and Fast-5 through Fast-80)
* Fast-40, -80 will not be defined for single-ended
* Single-ended and LVDS signals will not be mixed (that is, hybrid buses
are not defined)
* The Single-ended/LVDS/RS-485 differential sensing circuit defined in
95-
269r3 is acceptable
* The Fast-40 and Fast-80 skew budget defined in 95-307r0 is acceptable
* The Single-ended/LVDS/RS-485 differential sensing circuit defined in
95-
269r3 permits automatic switching between single-ended and LVDS
drivers/receivers and terminators
* When operating in single-ended mode (the sense line is low), a device
capable of both single-ended and LVDS shall enable its `ground
drivers'
continuously.
* The pin-outs developed at previous meetings are acceptable (see 95-
307r0)
* Maximum node capacitance = 15 pF
* Use current mode drivers, no driver output impedance specification
* Terminators bias signal lines with 100 mV, limited to 100 uA maximum
current
* Receivers not required to have hysteresis
5. Proposal for SPI-2 Electrical Interface (95-307r0) [Ham]
Bill Ham reviewed the subject document. As expected, it provoked discussion
on several topics -- not all of which are captured here. See 95-307r1, when
available, for additional changes that resulted from the discussion.
Kevin Gingerick was concerned about the proposed terminator bias circuit.
It
includes a "magic" circuit to provided 100 mV bias at no more than 100 uA.
He
thought a simpler circuit, perhaps just using three resistors should be
permitted. However, Bill and Paul Aloisi objected to the higher power
levels
that would result. The issue was deferred until later in the meeting.
Bill included several tables with first cuts at proposed electrical
parameters. The group suggested several refinements to the specifications
that will be incorporated in revision 1.
After examining the leakage currents of CMOS driver/receivers, the group
agreed to increase the proposed leakage current specification from 10 uA to
20
uA. This also caused the terminator bias current specification to increase
|from 100 uA to 250 uA.
The minimum DIFFSENS low-pass filtering was changed from an RC value to a
minimum time constant. An example circuit will be included. The
specification on the input pull-down resistor was changed to be at least 80
K
Ohms and small enough to overcome any receiver leakage current. The idea is
to bias the circuit so that an open input will cause the receiver "to
detect"
single-ended state.
Bill's 35 meter maximum cable length specification invoked considerable
discussion. While SCSI timing specifications may work at 35 meters, it is
not clear that LVDS will support that long of a cable due to attenuation. A
related (and equally controversial) discussion ensued on how many devices
can
be supported on a LVDS bus. Ed Gardner wanted to permit more than 16
devices,
while Bill felt that this capability was outside of the scope of SPI-2.
Bill
plans to do some lab testing to determine the effects of cable length and
number of devices on signal attenuation.
After discovering that earlier power calculations had included all receivers
powered on even when the drivers were enabled (not necessary and wasteful of
power), Bill re-opened Pandora's box and we re-visited the power budget
assuming that receivers would be powered off when the corresponding drivers
are enabled. As a result of the re-negotiations, we ended up with:
Driver current: 4 mA minimum (up from 3 mA)
5 mA nominal (up from 3.5 mA)
6 mA maximum (up from 4 mA)
Receiver sensitivity: 100 mV minimum (up from 50 mV)
Terminator bias: 150 mV maximum at 250 uA maximum (up from 100
mV)
6. The Case for 50 Ohm Cables (95-308r0) [McCall]
David McCall presented 95-308r0, "Fast 40 Proposal", which proposes that an
option for 50 Ohm cables be included in SPI-2. He pointed out that many
customers want to configure systems with 50 Ohm backplanes with close stub
spacing. A 50 Ohm configuration option would facilitate such systems.
Bill Ham spoke against this proposal as it would invalidate the power budget
just agreed to. Furthermore, it would require about twice as much power
dissipation for the drivers/receivers. He preferred that we push backplane
designers to use higher impedance design techniques. No consensus was
reached.
7. LVDS (95-269r3) [Aloisi]
Paul distributed copies of revision 3 of his proposal. Most of his comments
had been covered under item 5, above.
8. Single-ended Reset Signal [Gardner]
Ed Gardner suggested that it might be desirable to keep the RST signal
single-
ended in a LVDS system. This would facilitate using the trailing edge of
the
RST signal to qualify whether the system is single-ended or LVDS. Since the
RST signal is highly filtered, it should not be seriously affected by
common-
mode noise. This topic needs further thought and discussion -- no consensus
was reached.
9. Meeting Schedule
The next meeting of SPI-2 Study Group will be September 11, 1995, in
Bedford,
NH at the Sheraton Tara Wayfarer Inn (603-622-3766), hosted by Digital
Equipment Corp..
10. Adjournment
The meeting was adjourned at 4:40 p.m. on Monday August 14, 1995.
--
John Lohmeyer E-Mail: john.lohmeyer at symbios.com
Symbios Logic Inc. Voice: 719-573-3362
1635 Aeroplaza Dr. Fax: 719-573-3037
Colo Spgs, CO 80916 SCSI BBS: 719-574-0424 300--14400 baud
More information about the T10
mailing list