HARD DRIVE HASSLES

                            by Derek Woodlands



    Ed: Derek describes some very interesting symptoms and fixes for a
  wayward hard drive - his experiences may be very useful to others in a
  similar fix.


    ## 25 ## 25 ## 25 ## 25 ## 25 ## 25 ## 25 ## 25 ## 25 ## 25 ## 25 ##


    Having already picked up an Amiga A2090 hard disk controller card at a
  major department store sale for the princely sum of $20 (some will say
  that's all they are worth!), the hunt was on for a suitable hard disk
  when I came across an advertisement in the Sydney Morning Herald for 100
  megabyte SCSI hard drives for only $650. Enquiries revealed that the
  units bore the CMS brand and were preformatted for use on a Mac Plus (or
  should that be minus?).

    With my wallet lighter by the prescribed amount I was left with a
  dilemma. The controller card was not suitable for the mounting of a
  hard disk and I already had two 3.5" floppy drives. Rather than remove
  one of the FDD's, I decided to mount the drive in the 5.25" drive bay.
  Mounting proved simple with the construction of "feet" made out of some
  old 25mm aluminium angle bar. The drive was bolted to the left side of
  the bay onto the existing 5.25" mount. So far so good.

    Commodore in their great wisdom provide too short a lead from the hard
  disk LED on the front of A2000's to suit their own connector on A2090's
  and, rather than lengthen the wires by splicing in an extension, I
  elected to go without the LED (my 2000 was still under warranty and I
  did not want obvious signs of "tampering" should a repair be required in
  the future).

    HD Problems
    Formatting the drive proved no problem until I rebooted only to be
  confronted with that nastiest of requesters "DISK NOT VALIDATED".
  Repeat attempts saw the same thing happen and so it was back to the hard
  disk supplier for another unit (which was provided happily). The new
  one gave the same result. Contacting CMS put me on to NEC, the makers
  of the drive. They were perplexed and the only suggestion they could
  come up with was that the drive was formatted for a Mac and maybe that
  was the problem. they suggested "attacking" (my word!) the drive with
  a bulk disk eraser, which I have heard can actually do more harm than
  good because of residual magnetism between the tracks of the hard disk.

    About this time I read MAST's advertisement for the Flash 2000 hard
  disk control card in ACAR (The Australian Commodore and Amiga Review)
  magazine. I thought that maybe a modern controller would solve the
  problem (the A2090 is as old as Moses). Was I ever wrong!

    Lotsa bad blocks
    Upon trying to perform a low level format, the software supplied with
  the controller card informed me that error were encountered and that I
  should "Map Bad Blocks". This I tried and the software informed me that
  this operation could take a LONG time without saying just HOW long.
  After leaving my machine running for a couple of DAYS (no, it had not
  crashed, it was still mapping bad blocks), I rang MAST to inform him of
  the problem. Daniel rang the designer of the Flash 2000 and I was
  subsequently informed that that was a bit too long and something may be
  wrong with my system.

    At this stage, I was unsure whether the fault lay in my 2000 or in the
  drive itself. Another call to MAST asking whether I could bring in the
  card and drive and try it on one of their machines met with a very
  helpful "yes, no problem". Expecting to confirm my suspicions that the
  hard drive was, indeed, at fault I was surprised to say the least when
  it performed flawlessly on the MAST machine. Maybe my 2000 was the
  culprit.

    Miracle
    I returned home and began readying the computer for a trip to the
  doctor (can you claim on Medicare?) when the phone rang. I had already
  removed the 8 Meg RAM expansion card and a friend wanted a telephone
  number from my database. I turned on my machine and guess what, the
  hard drive autobooted!!!. Further tests without the RAM card showed
  that everything was A-OK.

    I immediately rang MAST to report a possible incompatibility between
  the RAM card (an OCTAMEGS, which is really a Supra product) and the
  controller card. The employee at MAST was quite concerned and promised
  to contact the designer of the Flash card the next day (Monday). With
  further contemplation of the problem, and remembering that I had the
  same problem with my A2090, I consulted an electrical engineering
  friend. This led me to think along the lines that the RAM chips could
  be emitting RF (Radio Frequency), which could upset the chips on either
  the controller card or the hard drive. I replaced the RAM card in the
  leftmost slot (a bridgeboard slot) and placed the A2090 between the two
  cards to act as a shield. Testing revealed that the drive worked,
  except for the odd clunk.

    Quick Fix
    Further rearrangement saw the hard card in the left most slot and the
  Ram card in the rightmost slot and this seems to have fixed the problem.
  I discovered that the use of the A2090 as a shield was not necessary.
  It seems that the RF (or whatever) is emitted mainly from the top of the
  RAM chips and luckily, the FDDs and so on to the right of the RAM card
  are unaffected. Obviously the hard drive itself must be highly
  sensitive to RF or surely this problem would have surfaced in the past
  with another Amigan.

    I have decided to run with this configuration for the time being, but
  have left my options open to either place a shield over the RAM chips
  (which may cause overheating), or to sell the RAM card and install
  another 8 Meg on board the Flash 2000 (which may put me back to square
  one).


    ## 25 ## 25 ## 25 ## 25 ## 25 ## 25 ## 25 ## 25 ## 25 ## 25 ## 25 ##