Formatting media issues (was Re: Amlyn Minipac diskette changer)

[Tony Duell]

> That's one issue I recall with the transition from RK05 to RL01 as
> "inexpensive" cartridge media in the DEC world.  The RK05 has a
> manually-aligned external positioner, and the RL01 (and RL02) uses
> embedded servo data.  One advantage to the RK05 technique is that you
> can format packs that have been bulk-erased (since the platters only
> contain data, not positioning info)  One advantage to the RL01 is that
> you don't have to be quite as precise with head alignment (you have to
> be close enough for the embedded servo data to be legible, then, IIRC,
> the positioner circuit can lock on the rest of the way).

Actually, it will try to lock no matter how far out the heads are (think 
of what hapapnes when you load a pack, it moves the heads in until it 
detects the guard band and then servo data on the currently selected head 
(nromally head 0). But if the heads are misaligned between the 2 sides, 
it may lock to the wrong cylinder when you switch heads.

But the RL's sure are easy to align, and you don't need an alignment pack 
which a plus...

> 
> There was another method that was acceptable for multi-surface packs -
> a servo surface.  It's not worth dedicating an entire surface to
> positioner data if you only have two surfaces - you get a poor

But that's what the CDC Phoenix drive (9448 IIRC) did. It ws an 80 meg 
fixed/16 meg removable drive. 3 fixed platters (5 data surfaces of 16M 
each and a servo surface) and a front-loading pack with one platter (one 
data surface, one servo surface). Apprently they didn't feel the radial 
alignment between the removeable pack and the spindle/fixed pack to rely 
on the fixed servo surface when accessing the removeable platter.

> space-to-cost ratio.  If you have 4-8 surfaces, it's more practical to
> give up a surface (vs a loss of some space for embedded servo info).

E.g the RK06/RK07 (3 data surfaces, one servo surface).

> An advantage of a servo-surface pack is that you can easily format the
> data surfaces as long as you leave the servo surface untouched.
> 
> I've heard of a field service device for writing servo data on a wiped
> RL pack, but I've never seen one in person.

I am sceptical. You would need to have a special positioner with 
mechancial feedback (so as to be able to position the heads without using 
the servo inforamtion), and I don't believe it could be made rugged 
enough to be able to be used in the field.

-tony