Sign in with
Sign up | Sign in
Your question
Solved

Samsung hd103sj

Tags:
  • Hard Drives
  • Samsung
  • Storage
Last response: in Storage
Share
May 3, 2011 12:27:49 AM

Hello!

My samsung 1TB hdd has stopped spinning. I think its the pcb. I dont see/smell any burnt components. I have a second HDD of the same model and firmware but bought 6 months later. I checked the pcb and there is a 1 chip diff. One pcb is by Foxconn and the other FCI.

I have look everywhere and i can't find a pcb for it. Anyone who knows a site/store that sells hdd pcbs?

More about : samsung hd103sj

a c 336 G Storage
May 3, 2011 8:20:30 AM

Foxconn makes the SATA connector, not the PCB.

Could we see photos or scans of the component sides of each PCB?

Sometimes there is an easy DIY repair.
m
0
l
May 4, 2011 2:21:32 AM

Hi fzabkar,

Thanks for the response, here are the pics of my non-working samsung.


This is the PCB of the working samsung.


m
0
l
Related resources
May 4, 2011 3:21:31 AM

You may try to find out problem by check a resistor and 2 TVS nearest power connector but I bet, your HDD have some damage in service area and you should bring your HDD to the shop, where can help you fix firmware. Good luck!
m
0
l
May 4, 2011 3:55:03 AM

thanks hddlab, if your shop were anywhere near me i'd bring it to you. I'm just so stressed about losing access to my files.
m
0
l
May 4, 2011 4:04:52 AM

Hi cesar509, I'm sorry, my shop in Vietnam not too nearest you. www.hddlabvn.com
m
0
l
May 4, 2011 4:22:38 AM

hi hddlab, thanks for the reply. Right now i'll hope for fzabkar's DIY solution. Hopefully he can whip up some magic. From the posts I have read of him, he's quite the authority on failed hdds.

If all else fails, i might visit Ha Noi in the near future. I'm from Manila. :) 
m
0
l
May 4, 2011 4:42:10 AM

Hi cesar509, I hope @fzabkar' DIY solution can help you, too. Sometime, I heard that guy very kindly when he willing to give free his knowledge for helping people.
Welcome to Ha Noi-Viet Nam.
Goodluck!
m
0
l
May 4, 2011 6:10:30 AM

where did u get trained in data recovery and hd repair? yes fzabkar is quite helpful for us newbies.
m
0
l
a c 336 G Storage
May 4, 2011 6:44:28 AM

I'm not a data recovery guy, but I do have many years of experience in chip level electronics.

That said, if the drive is not spinning, then there is no reason to suspect the service area. A drive needs to spin up before it can read anything.

Hddlab has suggested that you check the zero-ohm resistor and two TVS diodes near the SATA power connector. A cheap digital multimeter should cost less than US$10. Select the 200 ohms range and measure the resistance of each component.

I hope this photo clip is clear enough:
http://www.users.on.net/~fzabkar/HDD/HD103SJ_TVS.jpg

This is a similar board:
http://www.users.on.net/~fzabkar/HDD/HD103UJ_TVS.jpg
m
0
l
May 4, 2011 12:57:25 PM

Hi fzabkar,

I bought a digital multimeter and set the dial to 200 ohms for the zero-ohm resistor. The reading was 00.7. Then I set the dial for the diode check function, the 5V TVS read 478 and the 12V TVS read 736. I just put the black(negative) terminal on the side with stripe and the red(positive) on the opposite side. I hope I got it right?

Thanks, i'm sorry i'm just a newbie about these stuff.
m
0
l
a c 336 G Storage
May 4, 2011 9:54:25 PM

You're doing well. All those components test OK. The problem is somewhere else. :-(

I'm assuming that the drive isn't making any noises at all, ie it is not behaving as if the motor is seized or the heads are stuck to the platters (stiction). You would normally be able to confirm this by swapping the boards, but I don't know if the difference in the MCU (Marvell IC, big "M") part number is significant. I suspect that it isn't, but it's your call.

What I would do is to remove both PCBs from the drives and perform comparative voltage measurements. There appear to be 3 onboard DC-DC converters, associated with 3 coils -- 2 x 4R7 and 1 x 1R2.

Measure the voltages at the boxed terminals:
http://www.users.on.net/~fzabkar/HDD/HD103SJ_regs.jpg

Connect the black COM probe of your meter to a SATA power ground pin:
http://pinouts.ru/Power/sata-power_pinout.shtml
m
0
l
May 4, 2011 11:21:41 PM

swapped the PCBs and nonworking drive started spinning using the working PCB, very smooth running, normal hdd noise and no stiction however the system does not recognize that there is another hdd.

i connected the nonworking PCB to the workinghdd and it did not spin at all. So I think the MCU has the problem.

My non working PCB:
PCB # BF41-00303A 00
MCU 88i9025-TFJ2 P1S7110.4 1003 B0P TW
SAMSUNG 010 K4H561638J-LCCC, H5616 WRB054GAN

My woring PCB:
PCB # BF41-00278A 02
MCU 88i8925E-TFJ2 NOV3410.05 0931 C1P TW
ESMT M13S2561616A- 5T AZC1P92OW 0946


m
0
l
a c 336 G Storage
May 5, 2011 12:08:08 AM

You don't yet have enough information to blame the MCU for your problems. You first need to check its Vio and Vcore supply voltages. What do they measure?

As for the SDRAMs, both are 16Mx16 chips, so that difference is immaterial.

K4H561638J-LCCC, Samsung, 16M x 16, DDR SDRAM:
http://origin2.samsung.com/global/system/business/semic...

ESMT, M13S2561616A-5T, DDR SDRAM, 256Mbit, 16Mb x 16, 2.5V:
http://www.esmt.com.tw/DB/manager/upload/M13S2561616A.p...

BTW, in Seagate, WD, and Hitachi drives, a straight board swap between identical models will usually fail. This is because each board stores unique, drive specific "adaptive" information. These data need to be transferred to your donor PCB. I don't know if this is always necessary for a Samsung drive, but in your case there is an 8-pin Atmel serial flash memory chip above the ESMT SDRAM. I wouldn't transfer this chip until you are sure that your donor PCB is compatible, and that such a transfer is necessary.
m
0
l
May 5, 2011 1:36:46 AM

fzabkar said:
You first need to check its Vio and Vcore supply voltages. What do they measure?


black terminal to ground and red terminal should be in the parts you pointed out, right? dial is 200ohms. Sorry this is a crash course for me. Thanks for your patience.

m
0
l
a c 336 G Storage
May 5, 2011 1:47:25 AM

The dial should be at 2VDC or 20VDC, ie volts for voltage and ohms for resistance.

The board should also be powered on.

You could practice you technique by measuring the voltage of a 1.5V alkaline battery or 3V lithium coin cell battery.
m
0
l
May 5, 2011 2:27:00 AM

dial at 20VDC



I really appreciate the help fzabkar!
m
0
l

Best solution

a c 336 G Storage
May 5, 2011 3:51:18 AM

All those voltages look OK to me.

The Vneg voltage should be -5V. It supplies the negative voltage for the preamplifier chip inside the HDA. The preamp also needs +5V power from the SATA power conector.

The SDRAM and Atmel flash memory both require +2.5V. This is supplied by the uppermost DC-DC converter. The same Vio supply is also used by the Marvell MCU.

The lowermost DC-DC converter appears to provide the Vcore supply for the MCU. I would have guessed +1.2V, but your result is close enough. You could confirm these measurements by testing your replacement board.

That said, I think your only reasonable chance of a DIY repair is to find a board that is more closely matched. I don't know whether you will need to transfer the flash chip -- some DR guys say no, others say yes. If yes, then your local TV/AV repair shop should be able to handle it.

As for which chip is actually faulty, that is a hard call. AIUI, in a normal power on sequence, the motor controller verifies that the supply voltages are up to spec. It then releases the POR signal (Power On Reset). The MCU then wakes up and runs some self tests, including a test of the SDRAM and serial flash memory. If the POST passes, then the MCU commands the motor controller to spin up the drive. In your case there could be a faulty MCU, SDRAM, flash, or motor controller, ie just about anything. Hopefully the flash is intact.
Share
May 5, 2011 4:53:07 AM

Best answer selected by cesar509.
m
0
l
May 5, 2011 4:56:00 AM

Vneg is -5.15 yes sorry for the typo. Thanks fzabkar. I measured the replacement board and has almost the same results. I'll have the rest if the board checked tomorrow morning. I want to thank you for all our help. You have been most helpful. I'll inform you of further developments. Its so exhausting and frustrating but also a refreshing learning experience. :) 
m
0
l
December 11, 2011 4:47:51 PM

Hi Cesar509,

There is something Funny.

I've the same problem as you.

I have 2 samsung F3 HD103SJ.

An old one which doesn't work anymore. In this one, I want the retrieve data

And I've another Samsung wich work fine.

I tried to exchange the pcb each other. But tis doesn't work. The two drives are not compatible.

The thing funny is that for me the pcb which does not work has this serial:

BF41-00278A 00

And the PCB which works well has this serial:

BF41-00303A 00

Are you interrested to make an exchange of pcb ?


m
0
l
December 15, 2011 6:35:29 PM

I've the same problem as you with my samsung F3 HD103SJ
I think I cause this problem by connecting the supply plug reversed

So I removes the two large TVS chips (near the Sata connector) by using a cutter
also removed the 0 ohm resistor using soldering iron (near the Sata connector 5V leg) and instead of the zero ohm resistor I solder a small short wire.

and that it I fixed my disk now it works OK

Thanks to all in this forum

AK from Israel :) 
m
0
l
January 6, 2012 4:12:33 AM

hai my samsung hdd hd502hi is dead locate the board

What I would do is to remove both PCBs from the drives and perform comparative voltage measurements. There appear to be 3 onboard DC-DC converters, associated with 3 coils -- 2 x 4R7 and 1 x 1R2.

on of the 4r7 coil is damaged

what will happen ?

what will do next?

thanks



m
0
l
February 25, 2012 8:24:50 PM

fzabkar said:
All those voltages look OK to me.

The Vneg voltage should be -5V. It supplies the negative voltage for the preamplifier chip inside the HDA. The preamp also needs +5V power from the SATA power conector.

The SDRAM and Atmel flash memory both require +2.5V. This is supplied by the uppermost DC-DC converter. The same Vio supply is also used by the Marvell MCU.

The lowermost DC-DC converter appears to provide the Vcore supply for the MCU. I would have guessed +1.2V, but your result is close enough. You could confirm these measurements by testing your replacement board.

That said, I think your only reasonable chance of a DIY repair is to find a board that is more closely matched. I don't know whether you will need to transfer the flash chip -- some DR guys say no, others say yes. If yes, then your local TV/AV repair shop should be able to handle it.

As for which chip is actually faulty, that is a hard call. AIUI, in a normal power on sequence, the motor controller verifies that the supply voltages are up to spec. It then releases the POR signal (Power On Reset). The MCU then wakes up and runs some self tests, including a test of the SDRAM and serial flash memory. If the POST passes, then the MCU commands the motor controller to spin up the drive. In your case there could be a faulty MCU, SDRAM, flash, or motor controller, ie just about anything. Hopefully the flash is intact.

m
0
l
May 4, 2012 2:12:03 AM

Hi...I'm having the very same issue with my Samsung drive, it's the same hard dive model and the same PCB. I as hoping someone would have some additional advice. My drive was doing the same as cesar509's, was not spinning up or anything. I followed everything listed in this forum, I removed both the TVS's by the power and removed the 0 ohm resistor and soldered the gap where the 0 ohm resistor was.

Now, the drive will start to spin up and run for about 10-15 seconds, then shuts down.

Any ideas on what I could try next? Just trying to save the data from the drive.

Thanks!
Dale
m
0
l
May 8, 2012 12:49:12 AM

drcleven said:
Hi...I'm having the very same issue with my Samsung drive, it's the same hard dive model and the same PCB. I as hoping someone would have some additional advice. My drive was doing the same as cesar509's, was not spinning up or anything. I followed everything listed in this forum, I removed both the TVS's by the power and removed the 0 ohm resistor and soldered the gap where the 0 ohm resistor was.

Now, the drive will start to spin up and run for about 10-15 seconds, then shuts down.


Any ideas on what I could try next? Just trying to save the data from the drive.

Thanks!
Dale

Your HDD have some error in service area and you must pay at least $5000 for proper tool. Don't try this at home. Go for DR in your local.
m
0
l
June 13, 2012 9:44:59 PM

paulolesgaz said:
Hi Cesar509,

There is something Funny.

I've the same problem as you.

I have 2 samsung F3 HD103SJ.

An old one which doesn't work anymore. In this one, I want the retrieve data

And I've another Samsung wich work fine.

I tried to exchange the pcb each other. But tis doesn't work. The two drives are not compatible.

The thing funny is that for me the pcb which does not work has this serial:

BF41-00278A 00

And the PCB which works well has this serial:

BF41-00303A 00

Are you interrested to make an exchange of pcb ?

hello i need a pcb BF41-00303A 00 can you sell it to me ?
thank's
m
0
l
September 28, 2012 5:16:38 PM

Sorry florpanama, but since I've no solution for my HDD BF41-00278A 00, I keep my hdd BF41-00303A 00.

I've just acquire a pcb BF41-00278A 00 in second hand. But whith this new pcb, the motor of my hard drive does not start. Here is the mesure I have made:



Is there a solution to make this pcb work ?
m
0
l
a c 336 G Storage
September 29, 2012 7:57:46 AM

@paulolesgaz, your second hand PCB appears faulty. Without a Vcore voltage the MCU will be brain dead. I would measure the resistance of the coil, and I would also measure its resistance to ground.

The absence of the -5V supply is OK. In some designs the MCU needs to enable this supply by sending an appropriate command to the motor controller, so if the MCU is down, then the -5V will be missing.

m
0
l
September 29, 2012 8:23:02 AM

Thanks fzabkar for your help.

Sorry, I don't understand well what I would measure. I did not represente which component is the coil on the pcb.

If this pcb is really faulty, there is no way to repair it ?
m
0
l
a c 336 G Storage
September 29, 2012 8:49:45 AM

The coil is the component marked "4R7" in the bottom right corner of the following photo:
http://i54.tinypic.com/6zy1hk.jpg

If the resistance between the Vcore test point and ground is zero ohms, then this could be because the MCU is shorted, in which case a DIY repair would be impractical.

Otherwise, if the coil is open circuit, then you might be able to replace it.
m
0
l
October 23, 2014 12:50:29 PM

Hi,

I know this is an old thread, however ...

I have the same drive. It clicks around 10 times with about 2 clicks/second. Then it spins down.

I measured the voltage across the two diodes just next to the power connector. They showed +-12V and +-5V depending on the polarity.

I guess the diodes are dead then?

Can I replace them with any diode?

Thanks in advance :-)

Martin
m
0
l
a c 336 G Storage
October 23, 2014 1:33:06 PM

@astromarlar, there is nothing wrong with the TVS diodes. The problem sounds like an internal fault.
m
0
l
October 23, 2014 1:57:19 PM

thanks, fzabkar. Do you have any advice?

I have read about the freezer trick, is this something I should consider?

Also: If I gently shake the drive while connected, I can make the clicking noise go away for some time and it takes longer before it spins down. Does that tell something?
m
0
l
!