Bad Random Access Memory (RAM) chips will cause data loss, and recovery attempts
require them to be replaced.
Faulty DIMMs will usually manifest themselves in spontaneous reboots, blue
screens or program failures. Things of all sorts can happen when data that is
read from memory differs from what was written to that memory. An implication is
that some wrong data from memory will eventually end up being written to the
A faulty memory page in a disk write cache will result in wrong data written
to the drive. On top of that, the operating system will usually be unaware of
data corruption. A "write miss" situation is also a possible scenario, albeit
rare. A RAM malfunction can cause correct data to be written to an incorrect
location on the disk, thus overwriting some other piece of data.
Soon after a memory chip goes bad, the logical structure of a hard drive
starts to degrade, ending up in an unbootable system or an unmountable volume.
The extent of damage varies, but most files are usually recoverable.
The key point here is: a bad memory chip must be replaced before
attempting any repair. This is because:
- Attempting to repair the disk in-place (using CHKDSK or
Norton Disk Doctor)
can cause additional damage.
Attempting a read-only data recovery with ZAR tools is unlikely to damage
the drive further, but may damage the target (good) drive during file
- Recovery can be safely carried out if you attach the
damaged drive to another computer, but you must replace the broken memory
before you bring the data back, otherwise you risk losing your data again.
Several memory testing tools are available, examples being
They can reliably detect memory failures and we suggest using them when in