WD5000AAKS disconnects and reconnects constantly in Windows

[cmerrell]

Hey all, I just joined and I'm wondering how bad my situation is. I have a 500gig WD SATA drive with a MFG date of August 2007 that constantly disconnects and then reconnects in windows any time it is plugged in. This problem just started one day without any interaction from me. I came home to discover a whole bunch of messages in windows caused by the constant disconnecting and reconnecting. The first thing I tried was new data and power cables but that didn't fix the problem. I didn't really think it would either because the drive hasn't moved in a long time but I had to rule it out. I also tried a different SATA port on the MB with the same result.

Basically I'm wondering how hard it will be to safely get the data off the drive. This drive has basically all my family photos taken over the years and it is very important that they are not lost. The good news (I think) is I can see the data on the drive when it is connected and access it but it never stays connected long enough to do anything with it. While diagnosing yesterday it seemed like the drive was "spinning up" each time that it would reconnect, then it would disconnect and do it all over again. It seems to be some kind of component failure in the drive so I wanted to reach out to you all and find out what the likely cause of this problem is. It seems like professional data recovery/hard drive repair services are quite expensive and I will use one if necessary but I want to learn as much as I can first. Thanks.

Clint

[hhddrec]

Hi Clint

is internal or external drive?
if internal maybe a power unit fail. if not is this isue and data important, better contact with a pro

[cmerrell]

Yes sir this is an internal HDD.

[fzabkar]

See http://www.alexsoft.org/viewtopic.php?f=86&t=649&p=1789

[giannis267]

Hello everyone

I have exactly the same problem..
Disk was external(wd5000c032, My Book Essential)
I thought that usb connector has bad contact so i pull it out from USB enclosure
and test it in another enclosure to access data.
Drive (WD5000AAKS-00TYA0) behavior is the same, connecting for 5 sec, then disconnects and reconnects again for another 5 sec, a reboot loop. I saw all the data inside but i havent time(5 sec) to do anything with them..
I think PCB has a defect, looking for a overheating component but i dont see any..
Under a microscope, all seems ok, except a little oxidization at preamp and VCM contacts and on other side at JTAG contacts. I use eraser to clean them up and now are brand new..but nothing changed..

PCB etched number is 2060-701477-002 REV. A
PCB sticker number is 2061-701477-900 04P

I have another PCB from a WD3200AAJS with numbers:
PCB etched number is 2060-701477-002 REV. A (SAME)
PCB sticker number is 2061-701477-100 AD (NOT EXACTLY THE SAME)

Side by side is EXACTLY the same PCB, can i use it for a MCU transfer from original PCB to DONOR?
I have experience in micro soldering, hot air station and a WD Doctor from SD.

Any ideas?
Thanks

[DataPlanet]

mostly some component failure or firmware issue.. your first choice should be changing the PCB hdd electronic board.. hopefully your problem should be solve with this send me pictures of hdd so i can check my stock for donor drive.

skype: ontrack86

[giannis267]

Hello DataPlanet

Can the DRAM memory chip be the cause?

Drive is:
WD5000AAKS-00TMA0
25 JUL 2007
HHNCNV2AHB
Thailand

I'm thinking:
A)Replace the HY57V281620ETP on the patient PCB
B)Use as donor the second PCB from 3200AAJS, tranfer MCU and a new DRAM chip(this PCB have a HY57V641620ETP, a 64Mbit chip, not 128Mbit as the patient)

Is that right/safe?

[saltwater]

This could be a firmware module reading failure. It loads overlay modules (drive identified), and some module reading failure cause drive resets itself (dissappearing).

[giannis267]

Thanks for reply saltwater

What are the odds of 2 WD5000AAKS drives, my drive and cmerrell's drive, suffer the same problem?

1)PCB factory defect line(DRAM IC?)
2)Corruption of firmware(SA area? MCU?)

Any ideas of how can i begin with this problem?

[fzabkar]

@giannis267, your SD Doctor should be able to read the patient's ROM and SA modules. It should also be able to reprogram a donor's ROM.

Alternatively you could try SeDiv:
http://sediv2008.narod.ru/Easy3.9Passwo ... 567890.rar
http://sediv2008.narod.ru/Settings.rar

[giannis267]

Hello fzabkar

I did a FW Defect Scan for all modules (primary and secondary) and found no problems, all read succeed..so take a backup of them.
Then i did a SA Defect Scan for all tracks on 0 head and all are ok except track 0, but after a search i found that is normal for newest drives..so backup again..

Code:

Re: read head 0 track 0 failed , this fine ?
Yes this is normal. Track 0 will not back up on WD drives. There is no check sums in the track to compare so it always shows as an error. In older drives it can do it but the newer ones does not work. I have tons of them here the same. I can get older ones to read track 0 on head 0 and 1 but most of the time one or the other fail on it.

Also for record, found 4 defects in G-List and 2462 defects in P-List

BTW i found that if i connect the drive in USB enclosure and power it on, i saw the drive on "my computer" without disconnect-reconnect. Can stay there for ever.
When i try to access it, i see all the folders and after 1 sec, disconnects by itself.

What is wrong with this drive??

EDIT: I read the ROM and take a backup of it, but i dont know how i can test if ok or not? WD said only READ COMPLETED.

[fzabkar]

The ROM should be OK. If it were not, then the drive wouldn't spin up. FWIW, I have written a tool to perform an integrity check of a WD ROYL ROM ...

As for the disconnection problem, it doesn't appear to be temperature or APM related. Perhaps the drive is going offline due to difficulties with bad sectors. Have you tried the "slow responding" fix? You could clean the J1 contacts, just for good measure.

Otherwise, see this "WD Tornado" thread:
http://www.alexsoft.org/viewtopic.php?f=3&t=255&p=750

Apparently the Marvell 88i8745-TFJ1 MCU chip commonly fails with "head mimic" faults.

[lcoughey]

This is very likely just a flaky PCB, as was very common with this series of drive. So, yes, it is possible that both drives have a failing PCB and the more you fight with it, they eventually become drives with bad sectors and finally drives with damaged heads.

A common symptom for the drive is to detect very quickly and even read a single sector, but then go stupid if you try to read any further sectors.

Just find a known good PCB, transfer the ROM, clone the drive (likely has bad sectors, at the very least) and then you should be good to go.

[giannis267]

I refresh with new solder the J1 pads, with a bit flux and make it flat again with soldering wick!
Now are shiny like new! Test it again, now drive disconnects after 20-30 seconds.
Then i add a better foam under the pcb,for better heat absorb and drive stays connected for a minute or two.
Temp issue??

SMART attributes are perfect BTW, deep green all of them.

So if i use as donor the pcb from 3200AAJS, has exactly the same etched numbers, will be ok?
Should i transfer DRAM ic as well, patient has 128Mbit, donor 64Mbit, both from Hynix.

Thanks again for support..

[lcoughey]

Trying to fix is a waste of time...too many variables. Match the PCB number and move the U12 chip, if it exists on board. If no U12, you will have to read the ROM from the source and reprogram the donor.

[giannis267]

5000AAKS(patient) PCB and 3200AAJS(donor?) PCB:

U12 not present, ROM is inside Marvell.

Transfer MCU or reprogram it?
Do you know if DRAM must be changed too?

Attachments

[lcoughey]

Should be a suitable donor. Just read the ROM from the patient and write it to the donor. Remember to backup the donor ROM first.

[giannis267]

Thanks for quick response!

So steps are:
1)Read patient ROM(backup)
2)Read donor ROM(backup)
3)Write patience ROM to DONOR
4)Transfer pcb from donor to patient
5)power up patient

Firmware modules? ONLY ROM must be written?
Patient has 16MB CACHE, donor 8MB, leave DRAM ic as is?

[data-medics]

Only ROM resides on the PCB. Modules are in the service area on the platters.

Just out of curiosity, have you ruled out the really easy stupid things like a bad power connector or SATA cable? I've had drives act that way from something as simple as that. Could even be underpowering it, I've had a few docks that weren't providing enough power to spin up certain models.

[giannis267]

Thanks data-medics for reply

First things i checked, measure +5V and +12V, found them 5.14, 12.41 at 2 TVS diodes.
Also check +1.2, +2.5 and -5. All ok.
I also attach the drive internally at 2 desktops...same problem.

My mind is stuck at DRAM IC, i have see many memory ic's going bad at memory modules for desktop pc's and cellphones, why not at HD's?
Like a computer, when memory gone bad, computer restarts/crashes when trying to access that area, can be this the faulty part of this pcb?
And this is one of my concerns, can the drive working well with 8MB cache(from donor PCB) or need to transfer/find a 16MB ic?