[SunRescue] sun3/80 scsi

Greg A. Woods rescue at sunhelp.org
Thu Dec 28 19:49:31 CST 2000


[ On Thursday, December 28, 2000 at 14:46:51 (-0500), jeff borisch wrote: ]
> Subject: Re: [SunRescue] sun3/80 scsi
>
> one other thing...
> > Also if you're making up your own external components, make sure you're
> > supplying termination power to the bus from the most distant device too
> > (and of course make sure it's the correct polarity, or else you'll find
> > that one of the SCSI cable wires lets the smoke out real quick!).
> > 
> 
> Sorry to be dense but what is the polarity are you referring to? Does it
> have something to do with connecting the ribbon cable backwards? You know,
> the mistake that is easy to make if you have the cheezy connectors without
> the the key molded into it.

No, not the connector orientation -- just the polarity of the voltage
supplied for TERMPWR.

In fact the pin assignments for single-ended 50-pin cables have been
designed such that if you reverse the orientation of a connection
(eg. on a 50-pin two-row header connector without a molded key) nothing
bad can happen -- TERMPWR is on conductor #26, and it's oppositely
paired conductor (#25) is supposed to always be left unconnected.
(68-pin cables are similar, but they have all of pins 33-36 listed as
TERMPWRB, so reversing the header would leave them still all connected
to TERMPWRB, though I've never seen a non-D-shell 68-pin connector)
According to the standard all full-time initiator devices must supply
terminator power.

All you need to worry about is whether or not your system has the
correct polarity power out on TERMPWR, and that it's not connected to
GROUND.

Now that I re-read the old Sun hardware-ref FAQ, I'm pretty sure you
shouldn't have to worry -- only a few models of 3/60 are noted as having
their TERMPWR lines connected to GROUND.  [part numbers s 501-1205-09 or
lower, 501-1322-01, 501-1334-01, and 501-1345-01.]

As the FAQ says:  "This can cause a short if another device in the chain
is providing termination power!".  In fact with some older, larger,
drives, ones that have no current limiter (i.e. fuse), and with a
sufficiently beefy external power supply, it's possible to completely
vapourise conductor #26 in the average ribbon cable (and sometimes a
power trace on the drive controller board).

Oh, and one model, 530-1874-02, of the Sparc-10 motherboard has the
TERMPWR line grounded on the internal cables too, but enabling TERMPWR
on an internal drive will supposedly "only" trip the PS current limiter....

Of course you should never need TERMPWR enabled on an internal drive
anyway, since there's no terminator to power on a stub.  :-)

However as I mentioned you do probably want to provide termination power
from at least the drive at the far end of the external bus -- just to be
sure that the voltage available at the terminator way out there is
within spec.  Assuming you have no out-of-spec power supplies, etc., it
should even be OK to have TERMPWR enbled on all external drives.

-- 
							Greg A. Woods

+1 416 218-0098      VE3TCP      <gwoods at acm.org>      <robohack!woods>
Planix, Inc. <woods at planix.com>; Secrets of the Weird <woods at weird.com>



More information about the rescue mailing list