thanks for your reply as I was actually hoping to get one from you since I already found your posts / advice on this forum very helpful in the past.
Unfortunately I'm guilty of ommiting some details, so here is a more detailed anamnesis of the patient:
The HDD started to amass reading/writing errors under the Linux OS which in turn required reparation of the ext3 FS. For this reason I figured the drive would near it's end of life and I'd better take an image of it. Therefore I connected it to this external USB adapter and started cloning the disk with ddrescue
. The transfer rate was very low at 1MB/s, it ran a whole night and only had imaged 20GB with several errors (which were expected after all). I thought maybe it would run faster if I connected the drive to the internal SATA bus of the PC - nothing to lose right, since I can always interrupt and continue the rescue operation afterwards? Indeed a big mistake in hindsight! Anyway I left the power adapter connected just swapped the SATA data cables (PC was powered down for this purpose) and the drive failed to come up again ever since.
The Marvell MCU stores unique, drive specific data in its internal flash memory (the vacant location at U12 is reserved for an external serial flash). This means that a board swap is not a DIY option.
Well I am aware of the fact that adaptive data is stored in the MCU. However I failed to find any information regarding to what effect it will have on the drive if the PCB is replaced without
transfering the adaptive data. I would presume that the drive won't go "crazy" and overwrite its own data in the service area if it receives read only requests from the OS / rescue tools? Surely there must be some kind of safety measures in the microcode to prevent such incidences? Of course I'm referring to pure data recovery means, it's clear that it won't be used as a storage medium thereafter. After all I read some success stories, for example it seemed to work out for this guy (his WD drive model seems to be quite similar): http://forum.hddguru.com/transplant-platter-wd3200bevt-t18339-20.html
At the very least another PCB could be used to drive the motor and thus check if it is still working? Since I have to get a new drive anyway I might as well try to get an identical model.
The first thing to check is whether the USB adapter is delivering +5V to pins 7,8,9 of the SATA power plug in your cable.http://pinouts.ru/Power/sata-power_pinout.shtml
If the +12V supply was applied to the +5V pins, then the motor controller is most liekly dead, in which case the MCU has most probably been overvolted by faulty DC-DC converters internal to the motor controller.
As mentioned above I'm quite certain the power adapter is fine, it stills works with other drives. But just to be sure I measured the pins and they deliver 5.22V and 11.71V. The 3.3V rail is not connected.
With power removed from the board, set you meter to the 200 oms range and measure the resistance between SATA power ground and each of L1 and L2. I expect that these inductors will be part of the Vcore and Vio supplies for the MCU, or maybe the supply for the preamp.
Gnd to L1: 10.8 Ohm
Gnd to L2: 44.6 Ohm
Also measure the resistance between ground and each of the pins of transistor Q2. I suspect this may be a linear regulator.
Yes, the SOT223 casing suggests it might be one? BTW the imprint on it is AEQH. The measurements from top view of the SMD:
Top: 643 Ohm
Bottom (from left to right): 1017, 643, 10.8 Ohm
BTW, the fact that both chips are getting hot suggests that the board has no protection devices.
Why on earth does a manufacturer save $1 on circuit protection devices if it could make life so much easier in many DR cases? Frankly I don't get the point.