Mirrored ST2000DM001 both dead at same time. Recovery?!?!

[apjr1970]

Hi.
My first post

I've had a Netgear ReadyNAS NV+ for 10 years or so and it's been great. 2 disks in there, mirrored. A few years ago I "upgraded" and put into two Seagate Barracuda ST2000DM001 (PN: 9YN164-306) disks (cue pre-talkies villain music). A few days ago I discovered my ReadyNAS was powered down and wouldn't come back on. Fuse change... not that. Different socket... not that. New power lead... not that. OK, that's the limit of my electrical skiils, it's toast. After researching a replacement power supply (I suspected it's probably that) and discounting that option, based on price and no guarantee of a fix, I ordered a Netgear NV104. It arrived today and oh how excited I was!
Anyway, it is at THIS point that I decided to check the two Barracudas to ensure that all data was backed up before inserting to the new NAS (cue teary, heratbreak music). Stuck disk 1 in PC1 and disk not registered in Windows 10 or BIOS. Tried disk 2, same thing. Tried them both in computer 2, same thing. Bough a drive enclosure and tried that on computer 1, same thing. Computer 2, same thing. Tried on computer 3 (Linux), same thing. And this is what's breaking my heart man... I had two disks that were mirrored and both have gone kaput. At. The. Same. Time. I mean, come on!

OK, so I have backups of some of that data but not all of it. In my defence, I thought I was good cos my disk was mirrored!

Disks make no sound that I can detect when I switch on the (open) enclosure. So now, what are my options? Any help greatly appreciated.

Professional recovery is an option but a couple of hundred quid seems more than I want to pay. The stuff I have on there I can "reconstruct" by doing the work again. And I can do the cost calculation depending on how I value my own time. And it may yet come down to that...

And clearly, I would prefer to avoid the sost of pro recovery - if possible. I have two identical disks that failed at an identical moment, and that are displaying identical problems (on the face of it). I can possibly experiment on one to fix the other..... Also, I'm kinda suspicious that they both failed at the same time. I suspect maybe the fan failed and the whole thing overheated? I can't tell now as it won't switch on. Which makes me think that the problem is not mechanical but PCB related?

Man, I dunno. Any advice would be appreciated....

Andy.

[labtech]

The most important thing is that the diagnosis on the drives is correct.
High chance they may have an electronics issue (e.g. Damaged TVS - check with multimeter), but also maybe they were set to operate in standby mode (can check this with something like HDAT2). Both of these could prevent drives from spinning.

[fzabkar]

http://www.users.on.net/~fzabkar/HDD/ST ... 01_TVS.jpg

http://www.users.on.net/~fzabkar/HDD/TVS_diode_FAQ.html

If and when you get the drives to spin up, you will probably find a Linux file system. Try to mount the drive from a Ubuntu Live CD or Linux Reader for Windows or your Linux box.

http://www.diskinternals.com/linux-reader/

[fzabkar]

As already suggested, check whether Power Up In Standby (PUIS) has been enabled. AFAIK, your Linux box or Ubuntu Live should automatically spin up such drives and mount them, but ICBW.

[apjr1970]

Hi. Thanks labtech and fzabkar for those nuggets of advice and direction. I tried HDAT2 to investigate the PUIS issue but the drives are still not recognised. Will investigate the TVS situation once torx screwdriver has arrived and I can detach PCB. Will report back later. Thanks for everything so far.

Andy.

[apjr1970]

Hi Fzabkar,

First, my PCB looks a little different (but similar) to your pic.

At the 2k resistance setting on my multimeter I get a reading of 0.715 in one direction and 1 in the other over the 12V TVS. But I get 0.030 in BOTH direction over the 5V TVS.
Both the 12V zero ohm resitors and both the 5V zero ohm resistors seem to be fine, all registering zero ohms.

Does that suggest that the 5V TVS is no good?

Andy.

Attachments

PCB

[data-medics]

OK here's a picture of what to check:

#1: Check these which a multimeter. They should read 0 ohms (dead short). If they don't you can replace, or bypass with a piece of small wire.
#2: Your TVS Diodes. These will read very little resistance one direction but very high resistance the other direction when you switch your leads. If they are dead short both ways, simply remove.
#3: If you give up on fixing the PCB, simply move this chip over to a good PCB of the same PCB number (something like 100664987 REV A) which is printed into the board. It contains the adaptive ROM code and will make the new PCB compatible with the old one.

I'm giving this advice on the assumption that the drives don't spin at all. You said they "aren't making any noise", so that's what I'm assuming you mean. However I've had people describe that to me when they really just mean it's not making any bad noises. If it is spinning, then it's unlikely a PCB issue at all.

[fzabkar]

30 ohms is very low, so something is loading the 5V rail. Remove the 5V TVS diode and retest it out of circuit. Also test the resistance between the pads on the PCB after you have removed the diode.

Could we also see a photo of the NAS PCB? There should be a 5V regulator on this board. A damaged 5V TVS diode on your HDD would suggest that it was overvolted by the NAS PCB's 5V regulator circuit.

Do both HDD PCBs have the same problem?

[apjr1970]

The measurements I gave above are the same for both HDD PCBs.
I snipped the 5V TVS from ONE board only for now.
Measuring the 5V TVS out of circuit I get 674 ohms one way and open circuit the other.
Measuring the resistance acrosss the PCB pads I get 30 ohms.

Regrettably, I can't get the fascia off the NAS in order to remove the PCB. I've disassmbled most of it but the remainder is assembled with rivets. I assume the thing was rivetted before the PCB went in so I believe the remaining screws retaining the PCB are behind the fascia which must detach somehow. But I can't quite figure out how!

Does the above help with any further diagnosis?
Thanks for the assistance so far.

Andy.

[fzabkar]

The 5V TVS diode is OK. The problem is most likely due to one or more of the LX7157B regulators near the motor controller.

Measure the resistance between each of the V1, V2, and V3 test points and ground (eg a screw hole).

In the worst case you could replace the PCB and transfer the serial flash memory chip (W25Q40BWS05) for US$50:
http://store.yahoo.com/yhst-14437584971 ... 22209.html

If you choose to DIY, then be aware that you must not damage the chip, otherwise data recovery will become extremely expensive.


LX7157B, Microsemi (MSC), step-down PWM regulator IC with integrated high side P-Ch MOSFET and low side N-CH MOSFET, 2.2MHz, 3V to 5.5V:
http://www.microsemi.com/document-porta ... -lx7157bds

W25Q40BWS05, Winbond, 4Mbit serialflash memory, 1.8V, Dual/Quad SPI:
https://www.winbond.com/resource-files/ ... 101113.pdf


Edit: I would also measure the resistance between the 6855A coil and ground.

Attachments

SPI_flash.jpg (61.18 KiB) Viewed 19098 times

[apjr1970]

Hi Fzabkar,

The various points you mention measure as follows (moving from top to bottom and left to right in your pic):

V1.1 = 64 Ohms
V1.2 = 64 Ohms
V1.3 = 64 Ohms

V2.1 = 0 Ohms
V2.2 = 0 Ohms

V3.1 = 323 Ohms
V3.2 = 323 Ohms

Resistance between the 6855A coil and ground = 30 Ohms

I can have a go at removing the chip. (If necessary!). What are the secrets to doing it safely....? Enough heat to melt the solder but insufficient heat to damage the chip. Just the "right" amount.....?!?!


Thanks,
Andy.

[fzabkar]

There is a short circuit at the V2 test point. Hopefully it is caused by the uppermost LX7157B regulator chip.

download/file.php?id=8581&mode=view

This fellow (a technician) managed to repair his PCB for 2 Euros:

viewtopic.php?f=1&t=29204&start=20

It wasn't a neat job, but he recovered his data.

If you are keen on doing your own soldering, then IMO it would be safer for you to replace the uppermost LX7157B chip rather than attempting a ROM transfer. You could practice on some dead PCBs, eg old computer motherboards. In my younger days I have used a regular temperature controlled iron for QFP chips, but I think it would be preferable for you to purchase a hot air rework station. A technician at badcaps.net recently recommended a "Gaoyue 858 found on eBay for US$36.00". He says "you will need flux and heat protection tape. Electronic aluminum or heat tape".

There is a chance that there is damage on the load side of the R47 coil, in which case one or more of the other chips may be faulty. You can determine whether this is the case by desoldering pins 5 and 6 of the LX7157B and retesting the resistance at V2. If the short goes away, then you can be certain that the LX7157B is the culprit.

[apjr1970]

OK, thanks. A bunch of stuff for me to think about there. Need to establish how brave/adventurous/foolish I feel!
I'll get back to you.

Thanks,
Andy.

[data-medics]

Just for the sake of clarity, the drives don't spin at all right?

[fzabkar]

Just for the sake of clarity, the drives don't spin at all right?

Perhaps you missed the fact that one of the supply voltages is missing.

[apjr1970]

Yes, for the avoidance of doubt neither of the drives spins up. No noises - bad or otherwise.

Andy.

[fzabkar]

OK, thanks. A bunch of stuff for me to think about there. Need to establish how brave/adventurous/foolish I feel!

If you visit the badcaps.net forum, you should be able to find a technician based in the UK. Perhaps they could do the work for you at a reasonable price.

[apjr1970]

Hi Fzakbar,

I am struggling to find an LX7157B. I've ordered a replacement PCB. I think it may be best to send to an electronics shop and have them swap the ROM over. I looked for such an outfit today without any luck. I'll check out the forum you mention.

Air rework stations are available on ebay and worst case scenario I'll have a go myself... at least I'll learn something!


Thanks,
Andy.

[pcimage]

I'll swap the ROM chips over for a beer and postage:-)

Send the PCB's up clearly marked and I'll switch the chips and post them back.

PM me if interested.

Cheers

Sean

[apjr1970]

Hi Sean,

Hey, that sounds great! I'll message you.


Thanks,
Andy.