[net.periphs] Strange Noises From My Hard Disk

cramer@kontron.UUCP (Clayton Cramer) (04/29/86)

Help!  I have a 10 Mb CMI hard disk on my PC, mounted in an external 
cabinet.  I also have a DMA controller chip that I don't trust, but
because it is soldered in, I haven't replaced it.  Anyway, I'm 
starting to have problems with my disk drive apparently "disappearing"
from view of PC-DOS.  When I power the system off and then back on,
the problem goes away.  Finally, last night I using the hard disk and
there was a loud "thunk" noise three times in rapid succession.  It
was loud enough that I jumped back in fear of something attacking me
out of the drive.  (Yes, I know that's not going to happen, but it 
was THAT 6oud.)  After powering everything off and on again, the disk
was still working, and none of my data was lost.  Any idea what the
"thunk" noise is, and could my flaky DMA controller chip be the
guilty party?

o68@psuvm.bitnet.UUCP (05/09/86)

     
Questions about "THUNK THUNK THUNK" on a flaky hard disk...
The "THUNK" noise was a very rapid head seek, probably to the
beginning or end of travel.  Why?  It could have several causes.
     
Usually, the heads get lost and can't find either the cylinder
marks (loud thunk) or the end-of-travel marks (louder thunk).
     
I think the heads got out of sync with the controller, i.e. heads
on cylinder 200, controller on cylinder 120.  I'm not too familiar
with how the recovery for this is set up, but some drives will do a
Return-To-Zero seek to settle the question.
     
If you have a few things dropping out at once, i.e. a flaky controller
board, then the RTZ sensor might not be picking up and you'll get a
head slam.  These noises are scary, but rather harmless as far as
permanent damage goes on a winchester.  (On the old CDC Hawk drives,
a head slam was sometimes catostrophic.  The Forward-End-Of-Travel
stop would break off and make many itty-bitty marks all over the disk
and anything else it would hit.  But I digress...)
     
Now - for the cause: Your DMA chip may be what's doing it, but don't
bet your life on it.  If the data gets mangled anywhere on its way from
disk to memory, it can give a bad cylinder number (most likely it'll say
it's on cylinder 0 when it's not, but that isn't the only possibility).
Hence, a quick seek to find the proper track again - "THUNK".
     
If the thunk is a head slam, your controller may be what's gone flaky.
Check it over for loose connections, push down on the chips to make
sure they're seated right.  If you are comfortable with socketing chips,
you might reseat them while you're at it.  Look at all the solder
connections real closely - one might not be done right.
     
It's hard to say just what's causing it.  Sorry I can't be more help
than this.  Rest assured, though - your disk probably isn't committing
suicide.
                                                     ---Duck
     
     

smithg@kcl-cs.UUCP (Lionel Smith-Gordon) (05/16/86)

In article <5474O68@PSUVMA> o68@psuvm.bitnet.UUCP writes:
>The "THUNK" noise was a very rapid head seek, probably to the
>beginning or end of travel.  Why?  It could have several causes.

I had these funny noises last week - my hard disk sounded if it had
to 'go home' between reading every track/sector/bit/byte(??).

I fiddled with the drives ribbon cables and the problem went away...


-- 
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
| Lionel Smith-Gordon         [ N E O N ]                smithg@kcl-cs.UUCP |
| Kings's College (KQC) Department of Computing                             |
| University of London                           {ucl-cs ukc}!kcl-cs!smithg |
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -