modern removable media drives
On Sun, 15 Nov 1998 Chuck McManis wrote:
>My understanding of MO drives was the the servo tracks were optical (and
>thus very dense and accurate) but the data tracks were still magnetic with
>a conventional read/write/erase head.
No. The old "floptical" drives (made by Insite?) were I believe like that. Of
course this method doesn't have any greater data integrity than a floppy disk.
(What follows is really an oversimplification.)
Magneto-optical drives use, as the name suggests, both magnetic and optical
effects to read and write data. A layer within the media contains magnetically
sensitive elements. When this layer is heated to its "Curie point", about 200
degrees C, the polarity of the magnetic elements can be changed by an external
magnetic field from the drive head.
To write data, the laser heats up the target areas of the disk, and the
magnetic field is applied to record all 0s. On the next revolution of the
disk, the magnetic field is changed, and the laser heats up only those areas
which are to have 1s recorded.
This magnetic field does not have to be localised, so there is no magnetic head
in contact with the disk surface. Because the written polarity is "frozen into"
the disk, MO disks are not susceptible to magnetic fields as magnetic media
are.
To read data from the disk, the laser is used at a low power. Depending on the
magnetic polarity, the polarity of the laser light reflected from the disk is
rotated a few degrees either way. This rotation is called the Kerr effect. The
drive detects this, and thus determines whether a 0 or 1 was read.
"LIMDOW" media uses a slightly different method, which enables writes to occur
with no erase pass.
On Sun, 15 Nov 1998 Stephen Dauphin wrote:
>I'm not sure if the magnetic head in an MO bears any relation to the
>heads of other magnetic media drives. The reason for this thought is the
I don't think it does; the magnet is not used for reading, and its purpose is
to generate a magnetic field whose polarity only changes at a per-disk-
revolution scale, not a per-bit-written scale.
>each tiny heated spot is extremely small. It also doesn't have to be
>particualrly well targeted so I would think the head can be large, coarse
>and not particularly close to the platter.
That's right.
>I also gather it is not a fluctuating magnetism but simply a single
>direction attraction or repulsion. You need the two states, one to zero
>the section of disk and the other to write the bits that will be ones.
True. Even for LIMDOW media (which allows for improved write speed since no
erase pass is needed) the magnetic field polarity is constant over the whole
revolution.
>partition and initialize just like hard disks using hard disk formatting
>software. Or maybe they really don't do the format part. I notice, after
>pulling the shutter back, that the platter has small radial marks,
>seemingly corresponding to sectors. The disks also come with a fixed
>number of bytes per sector. Anybody know?
You can format them, both high level and low level providing you use the
appropriate software. If you have a Fujitsu drive, check their web site as
driver s/w and OEM manuals are available there. There is also free Mac MO
driver s/w at the Pinnacle Micro web site.
A low-level format allows you to tell the drive to rescan the disk surface and
map out any bad blocks using a sector slipping algorithm. This is normally
done at the factory. However, a very old disk may have more blocks mapped out
over time due to dust. After cleaning the disk, you could reformat it to be
able to access these blocks again. (Not that this really matters, e.g. a 640MB
disk has over 2000 spare blocks, so bad blocks don't actually reduce the
amount of data you can store -- but blocks found bad during use cannot be
remapped by sector slipping, so access to these is slower.)
MO disks have a fixed sector size; that's what the radial marks are for. For
3.5" disks this is either 512 bytes (for 128MB, 230MB and 540MB disks) or 2048
bytes (for 640MB disks). Unlike some hard disks, it is not possible to change
the sector size by doing a low-level format.
The data "density" of 540MB and 640MB disks is the same, the difference in
capacity being due to the different sector sizes. The same goes for 2.3GB vs
2.6GB 5.25" disks etc.
Another good thing about MO is that it is backwards-compatible, e.g. on my
640MB drive I can read and write 128MB, 230MB and 540MB disks.
>I guess they are cleanable. The disk seems to be made of a polcarbonate
>plastic. A manual cleaning doesn't look too easy as you would have to
Fujitsu and other manufacturers make cleaning kits for both drive and disks.
The disk cleaning kit holds the shutter back, and you turn a knob to revolve
the disc inside while wiping with a cloth.
I recently bought a dual-purpose MD/MO disk cleaning kit; cost was about the
equivalent of US$5. For the drive itself, you just put a special disk in the
drive and after about 10 seconds the drive spits it out.
>They are a little slow but faster than a diskette. Best read throughput I
>get is about 500Kb per second and when writing, the pace is about half
640MB drives are significantly faster than that, and modern units have a 2MB
cache which can speed up reads and writes significantly. Some 5.25" drives
have 4MB cache.
Data transfer rate for the current generation of 640MB 3.5" drives with
540/640MB disks is 2.8-4.7MB/s, random seek time 28ms. You'll achieve near
that for reading. Writing may be 1/3 of this unless you use LIMDOW media
and/or disable the automatic write-verify feature.
>that. One thing I do like is the lack of noise. Except for the fan, mine
>are dead silent and I have been contemplating whether I can eliminate
>that. Anybody know how hot would be too hot for the drive and disk?
Proper cooling is a good idea for MO drives, though you could always try
disabling the fan. The drive should shut itself down if the temperature gets
too high -- indeed you can interrogate the drive to determine what the problem
is; see the OEM manuals. Operating drive temp should be between 5 - 45 degrees
C.
-- Mark
Received on Tue Nov 17 1998 - 13:14:47 GMT
This archive was generated by hypermail 2.3.0
: Fri Oct 10 2014 - 23:31:19 BST