Author Topic: Diagnosing dead 3.5" hard drive--won't spin  (Read 658 times)

0 Members and 1 Guest are viewing this topic.

Offline iXod

  • Frequent Contributor
  • **
  • Posts: 311
  • Country: us
Diagnosing dead 3.5" hard drive--won't spin
« on: June 22, 2021, 07:36:09 pm »
SATA 1TB Samsung HD103SI/A. Completely dead.

No shorts between 5v and ground, 12v and ground. Checked TSP diodes across 5v and 12v, no shorts. There is one large 8-pin FET-looking device (maybe the motor switch?), also no shorts.

Other than swapping out the board (and swapping in the original flash memory chip with drive-specific parameters), is there anything worth checking for?

And please, no "It's not worth it" replies. I fix stuff. It's an interesting project.

Thanks.
 

Offline fzabkar

  • Frequent Contributor
  • **
  • Posts: 990
  • Country: au
Re: Diagnosing dead 3.5" hard drive--won't spin
« Reply #1 on: June 22, 2021, 07:46:08 pm »
The 4R7 coils are the Vcore and Vio supplies. Check the voltages between them and ground.

The 1R2 coil and the associated MOSFET is the inverting boost converter which produces the -5V supply for the head preamp. Measure the voltage at the anode of the diode.

If those supplies are OK, then you can look for diagnostic output at the drive's terminal (UART) port.

How to connect a terminal adaptor on a Samsung drive ?
http://www.hddoracle.com/viewtopic.php?f=116&t=189

BTW, the "ROM" (flash memory) is embedded in the Marvell MCU.
« Last Edit: June 22, 2021, 07:48:59 pm by fzabkar »
 

Offline iXod

  • Frequent Contributor
  • **
  • Posts: 311
  • Country: us
Re: Diagnosing dead 3.5" hard drive--won't spin
« Reply #2 on: June 23, 2021, 05:09:35 pm »
All measurements taken with pcb not mounted on, nor connected to the HDA. Power provided by the SATA cable and power supply.

The 4R7 coils measure 1.25v and 2.5v

The 1R2 coil measures -25mv (it is negative) measured at the anode.

I discovered that the TI SH6125B chip (motor controller) is very hot, almost (but not quite) too hot to touch. This, with the board not connected to the HDA.

Thanks for your interest. Much appreciated.
 

Offline Mario87

  • Regular Contributor
  • *
  • Posts: 125
  • Country: gb
Re: Diagnosing dead 3.5" hard drive--won't spin
« Reply #3 on: June 23, 2021, 05:58:23 pm »
All measurements taken with pcb not mounted on, nor connected to the HDA. Power provided by the SATA cable and power supply.

The 4R7 coils measure 1.25v and 2.5v

The 1R2 coil measures -25mv (it is negative) measured at the anode.

I discovered that the TI SH6125B chip (motor controller) is very hot, almost (but not quite) too hot to touch. This, with the board not connected to the HDA.

Thanks for your interest. Much appreciated.

Hot component and voltage rails being pulled low, sounds like a partial short within the MCU :-BROKE :--
 

Offline fzabkar

  • Frequent Contributor
  • **
  • Posts: 990
  • Country: au
Re: Diagnosing dead 3.5" hard drive--won't spin
« Reply #4 on: June 23, 2021, 08:30:53 pm »
All measurements taken with pcb not mounted on, nor connected to the HDA. Power provided by the SATA cable and power supply.

The 4R7 coils measure 1.25v and 2.5v

The 1R2 coil measures -25mv (it is negative) measured at the anode.
Vcore = 1.25V
Vio = 2.5V

The -5V supply should be measured at the anode of the diode, not the coil.

See the example in Part 2 of my tutorial:
http://www.hddoracle.com/viewtopic.php?f=59&t=231

I have attached the datasheet for an early SMOOTH combo spindle motor/VCM and PSU controller (L7250).

Edit:

I've never tried this, but you could disconnect the 12V supply from the PCB (by placing tape over the pins?) and see whether the IC still heats up. I suspect that it will shut down all the output voltages, even though they are derived from the 5V supply, but I could be wrong.

« Last Edit: June 24, 2021, 11:49:16 pm by fzabkar »
 

Offline RayRay

  • Regular Contributor
  • *
  • Posts: 228
Re: Diagnosing dead 3.5" hard drive--won't spin
« Reply #5 on: June 25, 2021, 11:08:21 pm »
Other than swapping out the board (and swapping in the original flash memory chip with drive-specific parameters), is there anything worth checking for?
IMO, nope. But for the record, you're mistaken in regards to the chip thing. For this particualr PCB, the BIOS is integrated into the largest (Marvel) controller chip, and that's what you'd need to swap (needless to say, this requires a hot air station, good flux, and experience!) if you don't feel comfortable doing this yourself, then you'd be best finding a phone repair shop that'd agree to do it for you. Look up the PCB number on hddzone.com or harddriveparts.com (it's printed on the board, underneath the white sticker). Good luck.
« Last Edit: June 25, 2021, 11:10:19 pm by RayRay »
 

Offline iXod

  • Frequent Contributor
  • **
  • Posts: 311
  • Country: us
Re: Diagnosing dead 3.5" hard drive--won't spin
« Reply #6 on: June 30, 2021, 01:21:26 am »
I want to diagnose the missing -5v supply (with the 1R2 coil).

What generates the switching of the FET for the -5v supply? Presently the FET has steady 5v on the gate and source terminals, a few mV on the drain terminals.
 

Offline fzabkar

  • Frequent Contributor
  • **
  • Posts: 990
  • Country: au
Re: Diagnosing dead 3.5" hard drive--won't spin
« Reply #7 on: June 30, 2021, 01:55:29 am »
The -5V supply and the Vcore and Vio supplies are all generated by the motor controller from the +5V supply. The controller generates the PWM pulses, which drive the gate of the MOSFET, but it does so on command from the MCU. Some MCUs look for a preamp (on the headstack) and will switch off the -5V supply if they can't detect it. In short, the -5V supply may be missing because it is not being switched on.

FYI, I have a database of datasheets here:

http://www.users.on.net/~fzabkar/Datasheets/DATAURLS.HTM

I should add that the MOSFETs and diodes for the Vcore and Vio buck converters are inside the motor controller.
« Last Edit: June 30, 2021, 01:58:57 am by fzabkar »
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf