Alan,

On Wed, 17 Jan 2007 13:02:44 -0000, Alan B. Pearce wrote:

I'd said:
> >These days perhaps, but in the Good Old Days of 14" ten-disk platters on
> >mainframes, if the power failed they'd spin for about 15 minutes!  Very
> >eerie to be in a computer room illuminated by emergency lights, with no
> >sound but the very quiet whine from half a dozen disk drives.  Normally
> >they'd be braked to a halt, so it was only in a power failure that it
> >would happen.
> 
> Hmm, all the Winchester drives I dealt with had brakes that came on with 
> power drop. The platter drives I dealt with all did an auto retract by 
> dumping a charged cap across the voice coil on power down. I do remember 
> they took a while to spin down even on a normal "unload" operation.

Right, but I'm talking about pre-Winchester technology, where the heads are part of the drive and you change just the disk-pack, weighing about 
10kg - early 1970s.  IBM's model number was 2319 (they don't write numbers like that any more! :-) and a single drive/pack held 25MB (sic).  There 
were a pair of drives in a wardrobe-sized cabinet, and a typical string was 9 drives, one being an online spare, so in a set of cabinets that would fill a 
large living room, you got 1/5th of the amount of storage space that I can now hide under my watch!

They did do an emergency-retract of the heads on power loss as you say, but with only air and bearing friction to slow down the platters, they spun 
for a remarkably long time.  Normal time to stop was about 45 seconds, if I remember rightly.  Except when one of our operators was getting 
impatient, overrode the drawer-latch to pull it open, then used friction on the top platter (that wasn't used for data) to stop it more quickly - I still 
cringe when I think about it...
 
> The one I always found impressive was the CDC 300MB removable platter drive 
> that was OEM'd to a lot of companies. The spindle motor in that would be 
> large enough to power a small car. From a standing start it was something 
> like 15 seconds to online when you pushed the start button. 

Yes I remember those - the heads were part of the pack, then known as a "Head/Disk Assembly" (HDA), and it looked rather like a flying saucer with 
a handle on top.  I don't remember its IBM model number, as I'd moved from operations into programming by then.  The one chronologically in between 
the 2319s and that one, which had 100MB per pack but the heads still stayed in the drive, was the 3330.  It's amazing the amount of junk that stays in 
my memory... rather like my house, actually!  :-)

Cheers,


Howard Winter
St.Albans, England


-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist