Catastrophic failures in Western Digital PCBs

Research and Development. This is the place to report experimental stuff related to data recovery.
fzabkar
Contributor
Contributor
Posts: 550
Joined: Tue Apr 16, 2013 9:28 am
Location: Australia

Catastrophic failures in Western Digital PCBs

Postby fzabkar » Wed Jan 28, 2015 7:10 am

Catastrophic failures in Western Digital PCBs

For a long time I used to think that WD's TVS diodes were configured as follows (because I was working from photographs and because I couldn't believe that anyone could be so stupid as to do it any other way):

Code: Select all

                         .---------.
+12V  o-----------+------|         |
                  |      |  motor  |----> spindle motor
               D4 -      | control |
                  ^      |         |
           ___    |      |         |----> VCM
Gnd   o---|___|---+------|         |
                         ----------'
           R64

Code: Select all

                                                       .------------.
                  +----------------------------------->| +5V        |
                  |      .---------.      .-------.    |     preamp |
+5V   o-----------+------|         |------| Vneg  |--->| -5V        |
                  |      |         |      '-------'    '------------'
               D3 -      |  motor  |      .-------.
                  ^      | control |------| Vcore |---> MCU
                  |      |         |      '-------'
           ___    |      |         |      .-------.
Gnd   o---|___|---+------|         |------|Vio reg|---> MCU, SDRAM, flash
                         '---------'      '-------'
           R67

In such a circuit, the fusible zero-ohm resistors (R64 and R67) would go open circuit whenever the TVS diodes were shorted by an overvoltage, thereby disconnecting the incoming supply from the remaining electronics. Other HDD manufacturers (Seagate and Samsung) also use series-connected protection devices (zero-ohm resistor, polyswitch).

However, when I eventually got my hands on one of these boards, I found that the actual arrangement was as follows:

Code: Select all

                .--------.
+V  o----+------|        |
         |      |        |
         -  D   |        |
         ^      |        |
         |      |        |
        .-.     |  HDD   |
        | |     |        |
        | | R   |        |
        '-'     |        |
         |      |        |
Gnd o----+------|        |
                '--------'

Now consider what happens when the drive experiences an overvoltage.

Firstly the diode goes short circuit, causing the "resistor" R to appear as a direct short across the supply.

Code: Select all

                .--------.
+V  o----+------|        |
         |      |        |
         )  D   |        |
         |      |        |
         |      |        |
        .-.     |  HDD   |
        | |     |        |
        | | R   |        |
        '-'     |        |
         |      |        |
Gnd o----+------|        |
                '--------'

What happens next depends on the current limit of the external PSU. If the PSU is a 19V adapter such as is supplied with various laptops, then the supply's short-circuit protection circuitry usually shuts it down before it can damage resistor R. However, if the drive is an internal one, then the computer PSU's much higher current limit usually (?) blows the resistor open. When this happens the TVS diode is rendered ineffective, and the HDD's unprotected electronics is then hit with the full force of the overvoltage.

Code: Select all

                .--------.
+V  o----+------|        |
         |      |        |
         )  D   |        |
         |      |        |
         |      |        |
         |      |  HDD   |
         o      |        |
            R   |        |
         o      |        |
         |      |        |
Gnd o----+------|        |
                '--------'

Here are the TVS diodes and zero-ohm resistors:

TVS_diodes.jpg

Here are two examples of SMOOTH failures resulting from an overvoltage on the +5V supply:

5V-Overvoltage.jpg

5V-Overvoltage_2.jpg

Note that the preamp would most likely be damaged, requiring an expensive headstack replacement.

Here are two examples of SMOOTH failures resulting from an overvoltage on the +12V supply:

12V-Overvoltage_2.jpg
12V-Overvoltage_2.jpg (84.67 KiB) Viewed 797 times

12V-Overvoltage.jpg
12V-Overvoltage.jpg (64.7 KiB) Viewed 797 times

In the latter cases a board swap would most likely get the drive going again, but this requires that the "ROM" be transferred from patient to donor. If there is a serial flash memory IC at location U12, then an inexpensive DIY repair (~US$50) is a real possibility. Otherwise, if the flash memory is embedded within the Marvell MCU, then a DIY repair would be much more difficult.

Ironically, WD's fusible zero-ohm resistors don't protect the user's data. Instead they are actually responsible for data loss. If these resistors had been replaced with a wire link or a blob of solder, then the TVS diode would have done its job and diverted the overvoltage, sacrificing itself in the process and shutting down the external supply. All that the user would then have needed to do to recover their data would have been to remove the shorted diode. Other HDD manufacturers such as Seagate and Samsung rarely, if ever, sustain such catastrophic damage because their protection diodes are connected directly across the supply, and any fusible protection devices are in series with the supply.

kaxi
Registered User
Registered User
Posts: 8
Joined: Wed Apr 16, 2014 11:37 am
Location: Macedonia, GREECE

Re: Catastrophic failures in Western Digital PCBs

Postby kaxi » Wed Jan 28, 2015 8:35 am

Exellent research fzabkar, this explains a lot.

User avatar
Spildit
Posts: 1554
Joined: Sat Apr 06, 2013 4:59 pm
Location: Portugal
Contact:

Re: Catastrophic failures in Western Digital PCBs

Postby Spildit » Wed Jan 28, 2015 10:48 pm

Thanks for sharing !
Nice discovery !
1Q9xrDTzTddUXeJAFRn37aqh1Yr6buDCdw - (Bitcoin Donations)


Return to “R&D - Experiments”

Who is online

Users browsing this forum: No registered users and 1 guest