Path: utzoo!attcan!uunet!wuarchive!kuhub.cc.ukans.edu!jli
From: JLI@kuhub.cc.ukans.edu
Newsgroups: comp.sys.ibm.pc
Subject: Re: SCSI hard drive controllers..
Message-ID: <12376@kuhub.cc.ukans.edu>
Date: 23 Sep 89 00:43:28 GMT
References: <1096@neptune.AMD.COM> <51@pldote.intel.com> <[251a3461:5678.2]comp.ibmpc;1@ddsw1.MCS.COM>
Organization: University of Kansas Academic Computing Services
Lines: 66

In article <[251a3461:5678.2]comp.ibmpc;1@ddsw1.MCS.COM>, karl@ddsw1.MCS.COM (Karl Denninger) writes:
>>Item 5678 (1 resps) by kevin at neptune.AMD.COM on Fri 22 Sep 89 04:35
>>[Kevin Tomasek]    Subject: SCSI hard drive controllers..
> 
> We have qualified the WD1006V/SR2 boards a while back (about 6-7 months),
> and we build PCs.  Out of some 60-70 boards, I have had one bad one.  It
> failed within the first two weeks, and was immediately replaced.

  We are using the WD1006V/SR2 a lot, mostly with the Mitsubishi MR535 (I
could be wrong about the model number), the 65MB(RLL)/40MB(MFM) drives. It
indeed has a very low failure rate, compare with the ones we used to use
(such as the DTC 7287).  However, we had a lot of problem with the combination
of WD1006V/SR2 and Miniscribe 8450 (40MB RLL, 771 Cyl. 4hd.).  Does anyone
encounter this problem?  The common problem is that the bad sectors keeps
increasing (slowly at first, and after a few weeks it grows faster and 
faster).  After reformating the drives, the newly appeared bad sectors
are gone, but they are coming back soon.  We have specially recorded a few
of such cases, the bad sector locations seemed to be complete random.  It
is very frustrating, almost 95% of such combinations we installed (in different
kind of machines, DTK, Zenith, IBM AT, etc.) were returned to us within
two months of installation with the same problem (if we only installed a
few of them, we may consider unlucky. But up to 200 units? It's just too
much!).  
 
> None of the others have given anyone trouble.  There are some drives which
> fail to work with these things; if you try using it with an MFM drive it may
> or may not work.  RLL-qualified drives have all worked to date, and none
> have failed (so far).
>

  We also use WD1006V/SR2 with Seagate drives (277R, 4144R), and have not
found any problems yet.
 
> The 1006 is also a track-cache board, and incredibly fast.  It will also
> tolerate a 16 Mhz bus clock, so you can tell your system to run the 16-bit
> bus at 0 wait states (if your BIOS supports this option).  The 1006 will
> also remap drive types internally so you don't have to fight the drive
> tables in your machine.  Note, however, that the remapping will NOT work for
> drives with > 1024 cylinders.  SCO Xenix doesn't care; simply turn off the
> bios and use the formatter we provide.  386/ix, on the other hand, will
> choke.  For MSDOS this is not much of an issue.
>
  The on-board BIOS is nice, and it also provide user defined drive 
configuration capability (Type 1), so you can use some odd drives if the
operating system is MS-DOS.  We haven't tried SCO Xenix with it yet, but
as far as we know, the Novell Network would not like the on-board BIOS.
 
> All in all, quite a nice board and cheap too.  We sell 'em for $179.  For
> the performance you get they're hard to beat; I/O speeds approach
> 700KB/second depending on your I/O bus capabilities (I've seen anywhere from
> 550KB/sec to 690KB/sec).
> 
  This board is fast.  I have one in my computer (386/25MHz) at home, with
a Mitsubishi 65MB drive. With the DOS 4.01's smartdrive installed, the
reported disk performance index (Norton 4.51) is 9.6.

  By the way, we also have an unsolved problem with Seagate SCSI drives.
Every one of these drives (3.5" 20MB) we installed in past 12 months had
the same problem, i.e. very slow access time and sometimes won't boot.
We have sent them back to Seagate for warranty repair, and some of them
have been sent for the third time.  Any suggestions and observations?

Thanks!

Q. Liu
U of Kansas