LVD SCSI Driver Currents

bbrown!hpx400!HAYNES_DENNIS/error_bbc////////HPMEXT1/DENNIS#b#HAYNES#o#TUC#o#06%u2 at globe.indirect.com bbrown!hpx400!HAYNES_DENNIS/error_bbc////////HPMEXT1/DENNIS#b#HAYNES#o#TUC#o#06%u2 at globe.indirect.com
Fri Nov 10 11:11:00 PST 1995


I attended the SPI-2 Study Group working meeting on
Monday November 6.  Towards the end of that meeting,
which was addressing technical issues for LVD SCSI,
I think I heard a comment something like this:

"Maybe we should just spec the minimum value for
the driver current and let each company choose their
own maximum value.  Since the maximum driver
current is just a power dissipation consideration each
company can make their own tradeoffs between
increased circuit complexity to be able to hold a small
difference between the min and max driver
current vs. less circuit complexity and more power
dissipation."

[At least that s what I think I heard.  By then I was
pretty much brain dead from being in the meeting from
9:00 AM to 7:00 PM.]

This discussion was prompted by some material presented
by Richard Uber of Quantum that indicated that it is
desirable to 1) decrease the maximum value of differential
resistance of the terminator (from 130 ohms to 110 ohms
and  2) increase the minimum driver current from 4.4 mA
to 5.0 mA.   That material was from page 6 of his work and
was titled "FAST-40 NOISE MARGIN CASES, Incident wave
transitions, DRIVEN NEGATED ---> ASSERTED"


But another part of Richard s presentation indicates that
the Maximum driver current  IS  critical.  That is on page 2
and is titled "FAST-40 NOISE MARGIN CASES, Incident wave
transitions, DRIVEN NEGATED ---> HIGH IMPEDANCE
NEGATION".  It showed that the problem of state reversals
for bus release transitions is made worse by large maximum
driver currents.

I just wanted to point out that it appears that both a min
and a max values for the driver current will still need to
be specified.

So it looks like the driver designers are still
between-a-rock-and-a-hard-place (and the space between
them is getting smaller).  I m sure that sounds familiar to
most designers.


Dennis Haynes
Burr-Brown Corp.
Tucson, AZ
EMail: haynes_dennis at bbrown.com
Phone: 520-746-7262




More information about the T10 mailing list