It may be coincidence, but I have the second WD drive now (1TB), and had another one (2TB) approx 2 years ago which do clearly have a severe defect (large chunks of data sectors unreadable) whilst SMART claims that the drive is perfectly OK, and WD Data Live Diagnostics suspects a "cabeling fault" (3 different computers, 3 different OS, 3 different sets of SATA cables). If I use OEM tools (HDTune Pro) to scan the surface tons of defective sectors appear.

While it's not my primary job to do harddrive diagnostics it happens on a regular basis, and I have had dozens of harddrives on my lab bench during recent years, and no other but the two WD harddrives showed this annoying behaviour.

With all major OSses monitoring SMART to warn users of upcoming defects, I am suspicious whether WD programs SMART in a way that defects go unnoticed until the drive fails completely, probably to minimize warranty returns.

Please comment.

Arion.

I've nearly always had to return failed drives that were themselves warranty replacements. In those cases though it just seemed the manufacturers had reset or wiped the SMART info. Errors starting appearing more or less straight away in use.

Smart system also has the right to break.

I am not talking about manufacturers "refurbishing" disks and resetting SMART.

I suspect that with those WD drives SMART as well as the Digital Lifeguard Tools were programmed not to warn of the defects, only 3rd party tools found them correctly.

I therefore strongly disregard, opposite of what others say, to use a manufacturer's own tool for diagnostics.

Armin.

Btw, years ago I happened to talk to a guy who was employed by one of the major harddrive manufacturers.

He described the treatment of defective drives as follows:

Almost 1/3rd of a years full production was sold to one single large customer, who was, in those days the world's largest supplier of server storage systems, and probably still is. They got large quantities of returns ("by palette") every month from that customer. Some drives were indeed defective/DOA because of faults in the manufacturing process or bad shipping, a large quantity of drives was perfectly OK but nevertheless rejected by some technician for some unknown reason, some returns were fraudulent, and some had nasty, hard to find defects, e.g. heat-related, vibrations related, slack contacts and such, which slipped the factory tests which were rather perfunctory to save time and costs.

Since it was such a good customer, the returns were accepted with no exception without any questions asked. The drives then were re-looped into the production process right before the final testing equipment, and if they passed the tests, they were secretly marked/registered as "suspicious" and usually they were sold as new to "ordinary" customers, like you and me, at the regular price, or sometimes large quantities were sold to distributors with a discount, without any obligation that the distributor informs his customers or passes on the discount.

Only if such a drive was returned as defective *again*, it was finally scrapped.

Armin.

It's not really worth discussing (I think) and speaking of cheating without the RAW values. I mean I'd like to see if these files show reallocated or pending sectors. And also, it would be fair if those disks were allowed to run an extended self test. During every day use, bad sectors may not be discovered if they're not used yet. This isn't cheating, the drive simply never ran into them yet. Remember SMART is a reporting system. It can not report faults it hasn't discovered yet.

Other flaw with SMART is that the manufacturer decides when to cry wolf, set the thresholds. It may allow for significant amounts of bad sectors for example. So, without having to look at RAW values constantly, I rely on HD Sentinel to do that. HD Sentinel makes it's own estimation of a drive's condition. You could for example argue, it's not per se the number of bad sectors that are that important (of course it counts), but the rate at which they occur. But same goes for HD Sentinel, the hard disk first has to run into issues before that can be evaluated.

_________________
Joep - http://www.disktuna.com - video & photo repair & recovery service