01010101100110101010110011010101011001101010101100110101010110011010101011001101010101100110
100110101010110011010101011001101010101100110101010110011010101011001101010101100110101010110011010101011011010101011
10101010110011010101011011010101001101010100110101101101101010100101
1011010101100110011010101011010101111010111
11010101101101010100111110
1101110110110011
0111011110
111101
011
11
1
PC-3000 EXPRESS / UDMA / PORTABLE
Seagate
ACE Lab
F3 architecture
Technical support: [email protected]
Phone: +7 863 201 50 06
ts.acelaboratory.com
19
Thus during translator restoration indirect methods have to be used to identify the type of the fork to hide. It can be
both a "left" or a "right" fork. To understand the types please keep in mind that hiding of post-processing defects occurs
in several passes. Different passes may reveal defects before or after the previously hidden ones. Besides, during each
recovery step you may discover "forks" added during the current test cycle, i.e. a "left fork" when the area hidden at the
factory is readable because it was written to these addresses but immediately followed by unreadable data, because the
LBA where it was supposed to be written will match the starting LBA of the "left" branch.
Therefore, a "left" fork
should be hidden in the direction of LBA number decrease UNTIL THE UNC OCCURRENCE LOCATION.
Testing may also reveal "forks" hidden during earlier scanning steps. These forks have been written before hiding with
larger LBA number than they occupy at the moment. In this situation reading of user data terminates abruptly with a
UNC error followed by the hidden area, then user data continue. Please keep in mind in this connection that post-
processing testing used to be based on recording of a plain pattern – the entire sector was filled with a single byte.
Initially drives were tested using sector filling with byte 0x77, then manufacturer started filling the sectors with zeroes.
While in the first case it is quite easy to recognize the 0x77 code pattern (the probability that user data will match it and
occur in the "fork" location is rather low), filling with zeroes causes ambiguity – before sale the entire drive surface is
filled with zeroes and there is no way to discern by sector content a sector written and hidden during surface tests from
a sector in the user space of the HDD. The difference can only be noticed if non-zero user data are present in the area in
question. Following from the above, you can examine the "fork" area and make assumptions about the position and size
of the hidden area based on the data placement. While you can read the data preceding the LBA with UNC error using
a regular ATA command, reading the LBA with UNC error and the following sectors requires the extended
functionality of the utility. To do that, use the sector editor from the Tools menu. To access the data in the UNC area,
you have to switch the reading mode to reading via the utility. The utility will display a prompt asking whether data
should be returned in case of a real reading error (UNC). To identify translation problems, you should respond 'No'
(responding 'Yes' will allow you to obtain the uncorrected data in case of a corrupted sector, but the mode is unusable
for the purpose of translator restoration).
Fig. 5.8
