Viper 60s Theory Of Operation And Maintenance Manual page 153

VIPER 5C51S0, 125
&
150 THEORY OF OPERATION AND MAINTENANCE
The read gap is used for read-after-write checking of new data as well as normal reading of
data. Since the inter-record gap is only
0.013
inches, the controller begins writing block n +
1
before
it
finishes read-checking block N. By the time a CRC error can be found in Block N,
block N
+
1
is already half written. Because of this, the following protocol is followed:
1.
Viper drive begins writing block
N.
Refer to Figure
6-16
while reading this
protocol.
2.
When block
N
reaches the read head
(3000
flux reversals later), read checking
begins.
3. The write function finishes writing block N by appending a block address and
CRe.
4. A short resynchronization gap is generated and written.
5. The write channel begins writing block N + 1.
6. The read channel finds a CRC error in Block N.
7. Because a CRC error was found, block N must be rewritten. Block N + 1 is already
half written.
8. The drive finishes writing block N + 1, then begins writing block N a second time.
9. If no errors are found in the second iteration of N during the read check, then the
drive continues writing the second iteration of block N + 1 and proceed with its
normal sequence.
10. If
the second writing of block
N
still has a CRC error, then a third copy of
N
is
written, and block N +1 is also written. The drive repeats writing block N for a
total of
16
attempts.
If a CRC error is still evident, the drive stops tape motion, sets the Exception Flag
and rewinds the tape to BOT. The drive then informs the host of an unrecoverable
or hard error.
This error recovery protocol is transparent to the host if no hard errors are discovered. To
keep the host informed of the media quality, the drive can report the number of blocks
rewritten in response to a Read Status Command. The total number of soft errors is cumula-
tive.
6-22