[SunRescue] Hi!

[Peter Koch]

Hi!

>For 5MHz synchronous single-ended busses the SCSI-I hardware is
>sufficient and is a legal "alternative" in the SCSI-II spec.

Ah, i see. But the Sun3 only does async. 5MB/sec.

On the other hand i know that all Sun3 can disconnect/reconnect
and do it with harddisks that claim to be capable of dis/re.

>I wouldn't even backup that disk to an MT02 connected tape on the
>same bus off-hours as it'll wear the tape and drive out long before
>its time, not to mention probably take an order of magnitude longer.

It is not THAT bad. Maybe the drivers and the software can compensate
for the silly hardware of the "sc" and the MT-02. I have done such
backups and never noticed too much delays with my 3/160. Of course
things can change with a faster machine, say a SPARC 4/110 or so.

>It is definitely a problem on 3/80's from what I've read
>and heard second-hand.
>
>I've seen a 3/260 with "si" controller completely fry the TERMPWR line
>on a ribbon cable when the drive was accidentally rested live on the
>chassis thus grounding the line.

So this is maybe a problem of all Sun3, VME or not?!?

>...a volt/ohm meter and a bit of testing
>should ascertain whether or not any given SCSI port is conforming or not.

I'll investigate in this... fortunately i have some (hehe) Sun3's
here...

>> High quality cables and connectors? Flat ribbon cables
>> with pressed connectors work fine. Sun used them for
>> the internal cabling of the 3/xxx chassis, so it must
>> be ok!
>
>For internal cabling, yes, this is probably OK. 

I use flat ribbon cable with either DB-50 or Centronics connectors
all the time. I cannot say that high quality cables work better.
The high quality cables are usually 1m or 1.5m long, much longer
than necessary. I use them where i need the length and otherwise
i use flat ribbon cables which i can make as short as possible.

>SCSI will often work outside the spec.

Yeah! Sometimes even Y-cabling works. That is definately no longer
a bus! But i have found such things in my Sun3 servers...

>The speed of the data transfer is only barely related to the speed of
>the signals on the bus.

Yes, i know. It's the controller. The hard disks are definately
faster. The Sun3 SCSI controllers are slow beasts but despite
this they are troublesome too.

>However it worked pretty good for the time
>and was a lot cheaper than all of the alternatives then available. 

The alternative was the XY451 which was in no way better.
That was the reason why people bought third party controllers like
the Ciprico Rimfire and such. Here the expensive SMD disks could
show their muscles.

>Closed inspection showed that the problem is the internal SCSI cable.
>Consider those cables to be like belts in a car.  They may fail randomly
>and without warning.  Preventative maintenance is cheap and can save hours
>of frustration.

Maybe that's a good idea! I have a source of flat ribbon cable and
connectors of various kinds and it is not too expensive.

Tschuess

Peter