X3T9.2 /93-149 Rev 3 for the mailing
John Lohmeyer
jlohmeye at ncr-mpd.FtCollinsCO.NCR.COM
Fri Jan 21 08:10:31 PST 1994
I am forwarding Paul's proposal regarding TERMPWR decoupling to the
SCSI Reflector for wider discussion. (It will be in the mailing, but
there are a lot of people who are more likely to actually read it on
the reflector.)
John Lohmeyer
------------------------------------------------------------------------------
"Paul D. Aloisi 603-429-8687 wrote:"
>From uicc.com!aloisi Fri Jan 21 06:14 MST 1994
Date: Fri, 21 Jan 1994 07:44:27 EST
From: "Paul D. Aloisi 603-429-8687" <aloisi at uicc.com>
Reply-To: aloisi at uicc.com
To: larry_lamers at maxtor.com, john.lohmeyer at FtCollinsCO.NCR.COM
Message-Id: <00978DB9.D1322660.3176 at hawk.uicc.com>
Subject: X3T9.2 /93-149 Rev 3 for the mailing
X3T9.2/93-149 Rev 3
To: John Lohmeyer
Chairman, ANSI X3T9.2 (SCSI)
From: Paul Aloisi, Unitrode
7 Continental Blvd
Merrimack, NH 03054
Phone: 603 429-8687
FAX: 603 429-3460
Mail: Aloisi at uicc.com
Date: 20-January-1994
Subj: Proposed changes to SPI
These proposed changes to SPI are derived from
discussions of document X3T9.2/93-149 in the SCSI working
group on November 10, 1993. Specific working is proposed
for the identified sections of SPI.
1) Change 6.2 3rd paragraph:
The TERMPWR and TERMPWRQ lines should be decoupled at
each terminator with a low frequency capacitor
(2.2 < C < 10.0 uF) for devices not at the end of the cable,
devices at the end of the bus (2.2 < C < 50 uF),
and a high-frequency capacitor (nominally 0,01 microfarad).
Rational:
1. There needs to be a limit to the maximum capacitance that can
appear on the termpwr line.
2. If the capacitor size is not limited it causes major
glitches on the Termpwr line during hot plugging.
2) Change to Figure 9
Figure 9 Termination for differential devices.
Change Equal to or greater than 2.2 microfarads to
(2.2 < C < 5.0 Microfarads)
3) Changes to Annex G section G.4 (relating to
device removal and insertion with current
I/O processes allowed)
Add clauses (g) and (h) to read:
g) The power cycling in (d) includes the power cycle
to the term power line on the device being removed
or inserted that naturally occurs.
Switchable terminators in disconnect mode (i.e. not
being used as terminators) must not allow the high
impedance state at the device connector to be
disturbed during this intrinsic power cycle on the
term power line.
h) Changes in TERMPWR voltage caused by charging the
capacitor on the TERMPWR line of the device being
removed or inserted shall not be transmitted to the
signal lines through the termination being used by
the active bus.
Implementor's note: This specifically requires
EITHER:
(1) that there be no capacitance greater than
25 picofarads on the term power line of
the device -- passive terminators pass
approximately 60% of the TERMPWR voltage
to the signal lines OR
(2) that regulated terminators be used on the
active bus --
Limiting the capacitance on the TERMPWR
line of the device to 5.0 microfarads is
sufficient to meet this requirement if
alternative-2 active terminators as
specified in SCSI-2 are used on the active
bus
4) Changes to Section 7.2.2
Change b) to read:
b) Shall conform to EIA RS-485 (ISO 8482-1982
TIA TR30.2) except that the common mode
voltage range shall be 0 to +5 volts DC and
-7 to +12 volts for transients less than 100 mS
instead of -7 to +12 volts.
Rational:
The common mode voltage requirements for EIA485
are not possible without distroying SCSI cables.
(The 7 ground wires in differential SCSI limits
the common mode to transients or low voltages.)
In addition these limits may cause excessive
power dissipation in the bus termination.
***
More discussion when the one suggesting the change is in the
meeting. Adding to section 7.1.1
f) The SCSI-2 220/330 terminator is not allowed for SPI
applications
More information about the T10
mailing list