Subject: Re: general sun3 question about scsi
To: None <ltost@erols.com>
From: Gordon W. Ross <gwr@mc.com>
List: port-sun3
Date: 04/30/1996 16:15:03
> Date: Mon, 29 Apr 1996 23:05:35 -0400 (EDT)
> From: LANCE TOST <ltost@erols.com>

> Hi there.. ihave a Sun 3/60 that i recently acquired and am attempting to 
> install netbsd on.  Unfortunately, i can't seem to get my scsi hard drive 
> to work properly.  Here's what i had to do:  Bought an internal scsi 
> drive and an external case.  So as it is, it looks like this:
> 
> Sun DB50 -> Centronics -> into case -> drive -> out of case -> centronics
> 
> My question is, where do i terminate?  On the drive with the terminating 
> resistors i received with it, or on the last centronics port?  

Either one should be OK if there is a very short cable between the
drive and the last connector, but the best is termination at the
extreme ends of the cable.

> Regardless, i've tried almost all combinations and it still won't work.  
> The built-in diags report that the SBC wr/rd test fails (no reason) and 
> the external loopback fails (various reasons depending on termination 
> and/or cable directions ... generally it says "all signals should be off 
> (high)".

That probably means that nothing is providing termination power.
Maybe you have one of the old Sun3/60 (#@$!) boards that GROUNDS
the termination power (TPWR) pin in the SCSI connector...
(in which case you need a cable with N/C on pin 26)

> Do either of these symptoms suggest the onboard scsi is fried?  People 
> have also  mentioned to me that scsi needs to be terminated at both 
> ends--is this already done on the mainboard?  I can't find a jumper or 
> anything which hints at it.

Probably not fried, but there could be a blown fuse.

> Also, when netbooting the netbsd kernel, it finds si0 and reports "SCSI 
> busy, resetting..." or something of the sort.. and it does this forever. 

Well, if the SCSI bus has no power to its pull-up resistors, then
it would look like it was stuck busy, among other things...
(I suppose the driver should declare it dead and give up 8^)

Gordon