[comp.sys.amiga] A500 SCSI Hard drives anyone?

lfk@amdahl.uts.amdahl.com (Lynn Kerby) (06/30/88)

[Though this is technical in nature, I want to get more exposure
 with this.]

Is there anyone out on the net that is using an A2000 style DMA controller
on an Amiga 500?  I realize that the 500 was not intended to be the machine
designed with expansion in mind, but someone else out there has probably 
tried to do the same.

I have an A500, 3rd party 501 with clock, Subsystem 500 (for holding Zorro
2 boards), Micron 2M, and a Pacific Peripherials Overdrive board with a
ST138N hanging off it!  The ST138N is partitioned into 3 slices for ease
of backups and storage management (IF I COULD TRUST IT!).

Now the problem....

When attempting to fill the drive up with goodies, I slapped a floppy into 
each of my 2 drives, opened 2 cli's and started copying selected sections
of both floppies onto different directories on the same partition.  See
drawing below if you are confused.

	dh0:		dh1:			dh2:
	 |		 |			 |
       MISC...	       aztec			MISC...
		      /  |  \
		 lib	asm   include
		  ^		 ^
		  |		 |
		df1:lib	      df0:include

Shortly after starting the second copy command, I begin to get requestors
saying that volume dh1: has a read/write error.  This is easily reproducable
on my machine (where I run stuff like interlace, GOMF2.2, conman, Dmouse1.07,
ARP1.1, etc.), so I thought I should eliminate all the above and went back
to a "STOCK" system.  After getting over the culture shock :-), I got around
to doing the same thing that caused the failure before.  Failed solid.  I
then took my system apart and took the controller and drive down to the 
local dealer for closer examination.  We set it up in one of the 2000's on
the floor and everything ran fine.  Now I'm really confused!  I brought it
back home and was easily able to reproduce the failure.  I checked voltage
everywhere, and it is between +1% and +8%, grounds seem to be fine, and I
am not running any of the PD stuff that I was running before.

The failure DOES NOT OCCUR if I only do one write to the drive at a time,
but what good is a multitasking machine if I destroy my filesystems when
I do more than one write at a time to a disk.

The media has been checked for bad blocks several times through all of this,
and has been low-level formatted with the correct (hopefully) parms several
times as well.  The read/write errors are file system related, and are 
generally recoverable with DiskDoctor.  

I'm at a loss for things to try (except buying a 2000 which I have started
working on).  My current suspicion is that there is a problem in the 500
either in the DMA logic/control or timing problems on the A500 bus.  I have
no idea how to verify that the problem is on the A500 bus though, so that
doesn't do me a lot of good.  If anyone out there has seen something like
this I would appreciate hearing from you (esp. CBM types).

BTW - I have contacted the vendor of the expansion chassis/drive controller,
	and they have no idea what could be causing this.

Thanks for any help/suggestions or just confirmations that others have
heard of/seen problems like the above.
-- 
     Lynn Kerby  -  Amdahl Corporation
                    Sunnyvale, CA
                    ...amdahl!lfk
		    lfk@uts.amdahl.com

Disclaimer: Any and all opinions expressed herein are my own and do not
            necessarily represent the views of anyone, especially my
            employer.