andrewl@hades.ausonics.oz.au (Andrew J Lockwood) (02/28/91)
I need a little help getting a Seagate ST225N 20MB SCSI hard disk working
with the A590. The problem is:
I can low level format the drive, partition and FFS format without to
much trouble. Then if I copy files to the Seagate I will eventually get
a R/W error no 42. The drive doc says:
"Defect map overflow: This error indicates that the total number
of defective sectors to be formatted exceeds the table size."
If I then run fixdisk and scan all blocks it finds that after some 30 or
so cylinders the following 600 cylinders all have sector 0 defective with
the error 42.
The verify option in HDToolbox sends commands the Seagate
thinks are invalid. The error explaination from the Seagate manual is:
"Invalid Command: This error indicates that the received
command byte is illegal or not implemented"
When I read the configuration from the drive I get:
Cyl : 613
Heads: 1
Blocks per Track: 68 ( Amiga assumes this for me !)
The drive doc says:
Cyl : 615
Heads: 4
Blocks per Track: 17 for 512 Byte block
If I use the default cyl/heads/blocks I end up with approx 600 bad
sectors ( fixdisk ). If I use the 4 heads 17 blocks configuration
fixdisk reports every combination of cyl 33-613 head 0-3 sec 0 as
bad. This amounts to over 2000 bad blocks for this config as opposed
to the default. What's going on?
I have run the Seagate "OFFLINE VERIFICATION PROGRAM" which is built into
the drive and tests:
1 Read/Write verification on special test cylinder
2 Media scan of user data area
3 Seek Test
These tests are cycled through ten times. The drive passes this test. I
phoned the Seagate Australian distributors (Y-Micro) who claim that the
drive should be O.K.
I have connected the Seagate to the A590 internal SCSI 50 pin connector. CBM
Sydney Consumer Support say this is O.K. to do. The 20MB 3.5 " ST506 disk
inside the A590 works fine.
The seagate has termination resistors installed and is the only device on the
bus apart from the A590.
The system is a normal A500 + A590 ( 20MB WD + 2MB RAM ) + A501 .5 MB RAM +
3.5" CA Ext Floppy + 5.25 " floppy + phillips monitor.
The Seagate is powered by a 288 Watt Switching power supply and +12V & +5V have
less than 10 mv ripple ( measured with a Fluke 77 DVM ). The Seagate was bought
second hand off the net from a guy (Paul are you listening 8-( ) who showed me
it working on a MAC. I don't know how old it is.
I have tried to map out bad blocks by hand ( since the verify doesn't work).
This can be used with limited success. After specifing 100 or so bad blocks
a limit is reached after which there is no more space for system files on the
Seagate. If I make a partition in the region covered by the bad blocks mapped
out, I can use this with no troubles. This seems to suggest that the bad blocks
are real and the seagate self test is lieing to me ( but then there is the
difference between the number of bad blocks with the different configurations
which suggests the Amiga is lieing to me????)
My questions are:
1/ What is the significance of the Seagate self test? Does it really indicate
that the drive is O.K.? Does it indicate that there are no bad block that
can't be slipped automatically? (Fixdisk is only READING blocks as this
self test claims to do).
2/ Why do the different head/cly/sector configurations work at all and why
does one result in 4 times the number of bad blocks.
3/ Does anyone use a Seagate ST225N connected to the A590? Any problems?
4/ If my hard disk really does have 4 radial scratches on every surface, is
it possible to map out more than 100 or so bad sectors. I need at least
6 times the normal space.
5/ Is there a program which will scan a filesystem and write a file containing
all the bad blocks effectively mapping them out?
6/ Do I need some special termination at the A590 end? The cable is 1m long.
Any comments please reply via EMail to andrewl@hades.ausonics.oz.au as
I don't receive many newsgroups anymore.
Thanks,
Andrew Lockwood
--
__ _____ __ Andrew Lockwood andrewl@witzend.ausonics.oz.au
/ ) / / Ausonics Pty. Ltd. Ph /61/2/428 6452
/--/ / / 16 Mars Rd, Lane Cove. NSW. 2066. Australia
/ ( o (__/ o _/___) o #define WEEK_DAY SATURDAYjesup@cbmvax.commodore.com (Randell Jesup) (03/01/91)
In article <1991Feb28.040101.18591@hades.ausonics.oz.au> andrewl@hades.ausonics.oz.au (Andrew J Lockwood) writes: >I need a little help getting a Seagate ST225N 20MB SCSI hard disk working >with the A590. The problem is: Note that the ST225N is seagate's oldest scsi disk, and is only marginally SCSI-1 compatible at best (for example, it ignores the LUN field - if you flip the LUN-enable dip-switch on the 590, it will show up as 8 drives). >I can low level format the drive, partition and FFS format without to >much trouble. Then if I copy files to the Seagate I will eventually get >a R/W error no 42. The drive doc says: > "Defect map overflow: This error indicates that the total number > of defective sectors to be formatted exceeds the table size." Kind of wierd. Where did you get the 42 number from? I read that as HFERR_Phase: illegal or unexpected scsi phase (note that scsi extended error codes are not the same as scsi.device io_Error codes). Do you have reselection turned on? It may well be that the seagate doesn't support it (the original reason for the flag to turn it off!) >The verify option in HDToolbox sends commands the Seagate >thinks are invalid. The error explaination from the Seagate manual is: > "Invalid Command: This error indicates that the received > command byte is illegal or not implemented" They apparently don't implement the scsi Verify command (very bad). >When I read the configuration from the drive I get: > > Cyl : 613 > Heads: 1 >Blocks per Track: 68 ( Amiga assumes this for me !) > >The drive doc says: > > Cyl : 615 > Heads: 4 >Blocks per Track: 17 for 512 Byte block Note: seagate reserved 1 sector per _cylinder_ for mapping, plus 2 cylinders for track/cylinder replacement. You sure that's not 17 blocks per track, 67 blocks per cylinder? Hmm, maybe the old HDToolbox with A590's didn't have that - you could try a more modern one off an A2091 or A3000 install disk. In any case, Seagate is talking hardware, not usable configuration. >I have connected the Seagate to the A590 internal SCSI 50 pin connector. CBM >Sydney Consumer Support say this is O.K. to do. The 20MB 3.5 " ST506 disk >inside the A590 works fine. Yup. >1/ What is the significance of the Seagate self test? Does it really indicate > that the drive is O.K.? Does it indicate that there are no bad block that > can't be slipped automatically? (Fixdisk is only READING blocks as this > self test claims to do). Dunno, sorry. >3/ Does anyone use a Seagate ST225N connected to the A590? Any problems? We did it while writing the drivers and hdtoolbox. Sort of worked, though it has some serious firmware bugs (like the lun problem). >4/ If my hard disk really does have 4 radial scratches on every surface, is > it possible to map out more than 100 or so bad sectors. I need at least > 6 times the normal space. HDToolbox won't do it at this point (it only reserves 2 cylinders for our own bad block mapping (plus partition/fs info)). However, the value is variable, and other RDB setup programs (like the one for the Hardframe) can also read/write RDB's, and may allow setting the number of cylinders. With scsi disks normally mapping is done by the drive with the scsi command Reassign Blocks. >6/ Do I need some special termination at the A590 end? The cable is 1m long. 6m is the scsi spec. _should_ be ok (it might not hurt to try a short cable). -- Randell Jesup, Keeper of AmigaDos, Commodore Engineering. {uunet|rutgers}!cbmvax!jesup, jesup@cbmvax.commodore.com BIX: rjesup The compiler runs Like a swift-flowing river I wait in silence. (From "The Zen of Programming") ;-)