SPI-2/Fast-40 pins

Vit Novak Vit.Novak at Eng.Sun.COM
Tue May 23 12:58:28 PDT 1995

>From the timing point of view ( the sum of skews and delays ), the 
single ended and differential SCSI work the same and the  bus could 
be of the same length.

Bill Spence and Bill Ham proved that we can go 20m with the SE SCSI 
if only two devices are connected. It is the waveform degradation due
to the capacitive loads along the SE bus that constrains us  to < 6m
because of the 'sensitive' lines.

If we take care of the 'sensitive' lines = REQ/ACK/DATA/PARITY then
the remaining lines should not be a limiting factor to go > 6m in my
opinion because of their slow speed. Or, am I wrong?


----- Begin Included Message -----

>From @ncrwic.WichitaKS.NCR.COM:jmcgrath at qntm.com Tue May 23 12:13:25 1995
Date: 23 May 1995 10:57:44 U
Subject: Re: SPI-2/Fast-40 pins
To: "SCSI" <SCSI at symbios.com>,
        "Walter Bridgewater x2371" <wally at eng.adaptec.com>

        Reply to:   RE>>SPI-2/Fast-40 pins

Why over 3 meters?   Single ended SCSI is defined for 6 m.  Fast SCSI
was restricted to 3, and Fast-20 to 1.5 to 3, but this was only for
data transfer operations - since we never changed anything having
to do with non-DATA phase operations, the original 6 m should still
hold.  If we enhance REQ/ACK/DATA/PARITY to go > 6 m, then 6 m 
due to the other lines beomes the limiting factor.

Actually, the original 6 m single ended SCSI limit was probably a
REQ/ACK type limit as well, so maybe we can go longer.  Does anyone
remember (John?) and/or have done empirical tests?


Date: 5/22/95 1:02 PM
To: Jim McGrath
From: Walter Bridgewater x2371
Received: by qm_smtpgw.qntm.com with SMTP;22 May 1995 13:02:11 U
Received: from worf.qntm.com (worf-gw.qntm.com) by mail.qntm.com with ESMTP
	( id AA168522361; Mon, 22 May 1995 12:52:42 -0700
Received: from mpdgw2.hmpd.com by worf.qntm.com with ESMTP
	( id AA203123018; Mon, 22 May 1995 13:03:38 -0700
Received: (from root at localhost) by mpdgw2.hmpd.com ( id
OAA11795; Mon, 22 May 1995 14:02:35 -0600
Received: from aztec.ncrmicro.ncr.com( by mpdgw2.hmpd.com via
smap (V1.3)
	id sma011729; Mon May 22 14:01:41 1995
Received: from ncrwic.WichitaKS.NCR.COM (ncrwic.wichitaks.ncr.com
[]) by Symbios.COM ( with SMTP id OAA07619; Mon, 22
May 1995 14:00:49 -0600
Received: by ncrwic.WichitaKS.NCR.COM; 22 May 95 14:57:46 CDT
Received: by ncrcetc.WichitaKS.NCR.COM; 22 May 95 15:42:26 CDT
Received: by ncrhub4.ATTGIS.COM; 22 May 95 15:33:10 EDT
Received: by ncrgw1.ATTGIS.COM; 22 May 95 15:29:43 EDT
Received: from eng.adaptec.com ([]) by milpitas.adaptec.com
	id AA05690; Mon, 22 May 95 12:22:24 PDT
Received: from atlantis by eng.adaptec.com (4.1/SMI-4.1)
	id AA19208; Mon, 22 May 95 12:20:59 PDT
Date: Mon, 22 May 95 12:20:59 PDT
From: wally at eng.adaptec.com (Walter Bridgewater x2371)
Message-Id: <9505221920.AA19208 at eng.adaptec.com>
To: scsi at wichitaks.ncr.com
Subject: RE: SPI-2/Fast-40 pins

SCSI Reflector,

I agree that at least ack, req, data & parity have to be differential
for the reasons Tak suggest.

If the other 7 signals are single ended, 
Then their slew rates should be slowed down proportional to any
length increase beyond 3 meters.


Wally Bridgewater

----- End Included Message -----

More information about the T10 mailing list