14 Dec
2002
14 Dec
'02
9:09 a.m.
Johnny Bilquist wrote:
Huh? No. Actually, the data track itself is used as
the servo
signal. The head centers in on the track by centering on where the
amplitude is strongest.
Tony Duell wrote:
> Do you have any evidence for that?
Johnny Billquist wrote:
Yup. I quote:
"1.3.3.2 Sector Format - As shown in Figure 1-3, each sector contains:
* Servo information for head positioning
* Head (address) information
* Data - 128 words of 16 bits each, or
256 bytes of 8 bits each, or
170 words of 12 bits each
Only the data portion can be written by the user. The servo and header
information is protected by the drive logic and controller to ensure
disk integrity and cannot be written in the field.
Each sector starts with a sector pulse that is produced by a sector
transducer mounted on the drive unit. It senses the sector notches that
are machined into the hub of the disk cartridge.
During the time that sector notch passes by the sector transducer, the
heads detect two servo pulse bursts (S1 and S2) that are prerecorded on
the platter. These servo bursts are used by the drive logic for head
positioning."
Note that it says that S1 and S2 are prerecorded on the platter, but it
does NOT say that they are on-track.
Assuming that the RL drives work the same as other embedded servo drives,
the issue here is that the servo bursts S1 and S2 aren't really "on-
track". Rather, they're each recorded off-track slightly, in opposite
directions, or halfway between tracks. During the portion of the
sector time in which the servo bursts are active, nothing is recorded
on-track, so that the servo bursts can be read. The S1 and S2 bursts
are typically at different frequencies.
When the head is on-track, it reads the servo burst at a reduced
amplitude due to the burst being off-track. The amplitude is
proportional to the distance. Assuming that the S1 and S2 bursts were
recorded at equal amplitude, and have positions that are off track in
opposite directions by an equal distance, the track following can be
accomplished by positioning the heads such that the amplitude of S1 and
S2 are the same.
If the S1 amplitude is higher than S2, they move the head in one
direction, and if S2 is higher, the other direction.
Note that if the servo bursts are in fact halfway between tracks, then
for alternate tracks the relative directions of S1 and S2 from the track
will be opposite. So for proper positioning, the drive needs to know
whether it is on an odd or even track.
It is not clear to me how having an embedded servo burst that is actually
on-track would do anything very useful. The drive wouldn't be able to
tell where the peak was without seeking in both directions.
A servo writer normally needs to be able to position the head to within
very tight tolerances to a fraction of a normal track position. As has
been previously mentioned, this is usually accomplished using an optical
interferometer, which is not present on standard drives (due to cost).
However, I don't actually *know* anything about the RL internals, so it's
certainly possible that it works differently.