WD2500BEVT-22ZCT generates "random" pending sectors

[elmad]

Hi all.

I have a type of problem with this disk that I haven't ever seen before.

First time I found some pending sectors so I did a ATA SECURE ERASE to force the disk at clean them or relocate them. After the secure erase, Current Pending Sectors was 0 and Relocate sectors was 0 too. So I tought that it was only a soft problem because of some unfinished write.

So I started to clone an image on the disk and the cloning was fine. But when I later mounted the partitions I had some error on reading sectors. So I started to run some Smart test and the errors began to involve more sectors. I tried with hdparm to write some of these sectors to force relocate but the write was successful, so no relocating was performed.

I guessed that the problem was power supply so I tried with another power supply and I did another Secure Erase. Again pending sectors was 0 and relocated sectors was 0. Cloning and reading this time seemed ok and I started using this. But now (after some hour since secure erase) this disk has 192 pending sectors. So the problem is here again.

I suppose that the problem is an electronic problem and not a surface problem.

But I have little experience and litte knowledge, so because I am curious, I ask you if my assumption may be right and what can I try to "refurbish" the disk.

Or anyway, what do you think about these symptoms.


Thanks in advance, and sorry if my english isn't good.

[elmad]

I tried now with MHDD. First scan had many brown and red sectors. So I did a full erase with mhdd. As ATA Secure Erase did, pending sectors went to 0 and no reallocating was performed.

Now I'm scanning the disk and there are 2 brown sectors (scanned half disk). In the previous scan, at this point, there were many brown and read sectors. I guess that if I use this disk, slow sectors will grow as first time that I tried with secure erase.

What do you think about this behaviour? Had you ever seen it?

[arztt]

I have seen this too on a WD6400BEVT.

With a full erase all the pending sectors were gone, but "Reallocated Sectors" was still 0.
After some minutes, the drive got new bad sectors.

In my case it was a Head scratching on the platters, I guess.
Because after 5 times of full erasing, the "Reallocated Sector Count" started to grow.
And after some more erases, "Reallocated Sector Count" was full and so SMART bad.

Some (or all?) WD drives doesn't show the first 50 (?) reallocates in SMART. Just after some amount, the counting in SMART starts. But there are already some more entries in G-List then.


Somewhere I heard about this behaviour too, but the Reallocated Count was 0, even after 10 times of writing (with more than 200 pendings each time). So maybe firmware corruption.
Swapping the PCB (+ROM) didn't heal the drive.

[elmad]

I have seen this too on a WD6400BEVT.

With a full erase all the pending sectors were gone, but "Reallocated Sectors" was still 0.
After some minutes, the drive got new bad sectors.

In my case it was a Head scratching on the platters, I guess.
Because after 5 times of full erasing, the "Reallocated Sector Count" started to grow.
And after some more erases, "Reallocated Sector Count" was full and so SMART bad.

Some (or all?) WD drives doesn't show the first 50 (?) reallocates in SMART. Just after some amount, the counting in SMART starts. But there are already some more entries in G-List then.


Somewhere I heard about this behaviour too, but the Reallocated Count was 0, even after 10 times of writing (with more than 200 pendings each time). So maybe firmware corruption.
Swapping the PCB (+ROM) didn't heal the drive.

I had written and erased the disk many times for now, and reallocated sectors are ever 0 after an erase.

Now I'm thinking that can be a overheating problem. Contacts on pcb were all "brown" like metal overheated, so I cleaned them with a pencil rubber, but I dont think that this solves the problem.

Anyway, I stopped an erase at half sectors erased and checked the temperature. Smart reported 74°C! It seems very high, so the problem can be this?

Is this series of hard disks defective?

[moiz h/d]

i have a lot of this wd drive but i can't solve this problem this problem is physical problem maybe some had week or some platter scratchs

[fzabkar]

What temperature does SMART report immediately after a cold power-on?

[elmad]

What temperature does SMART report immediately after a cold power-on?

Now I'm trying it in a Acer notebook (it is his original disk) that haven't any hole for air where is the drive bay -.- .

Anyway, when the disk is idle, temperature is about 51° and when is working reaches 65° and probably more. I should try a cold boot, but if it's interesting, I had removed the plastic cover of the disk bay on the notebook and placed it on a notebook cooler with a fan in the middle, pointing on the disk. In this way the temperature is about 37°C when the disk is working. I think that it's too much considering that other disks, without air, stay at 40/45°. But now it has not generating pending sectors so I think that is a problem of overheating.


Tomorrow I'll get the temperature with MHDD after a cold power-on.

Last edited by elmad on May 18th, 2013, 22:18, edited 1 time in total.

[BlackST]

Hi all.

I have a type of problem with this disk that I haven't ever seen before.
...

I suppose that the problem is an electronic problem and not a surface problem.

The problem is INTERNAL but 90-95% there's nothing you can do about it - simply the drive is worn out. Maybe a partial solution but I won't bet on it.
Also when reworking these drives in that state they end "killed" in the process most of the times. I see it in the dozens everyday

[elmad]

Hi all.

I have a type of problem with this disk that I haven't ever seen before.
...

I suppose that the problem is an electronic problem and not a surface problem.

The problem is INTERNAL but 90-95% there's nothing you can do about it - simply the drive is worn out. Maybe a partial solution but I won't bet on it.
Also when reworking these drives in that state they end "killed" in the process most of the times. I see it in the dozens everyday

The disk starts at 30° and using it temperature arises to 65° and more. Am I right that is a problem of overheat? Is there something internal that overheats? Just curious...

[fzabkar]

AIUI, the temperature sensor is on the headstack, probably implemented as a diode inside the preamp.

http://malthus.zapto.org/viewtopic.php?f=59&t=229&p=582

Some papers on the subject state that "the spindle motor is the primary heat source within the HDD". The voice coil would also be a heat source. However, ISTM that your biggest problem is poor ventilation.

You might like to experiment with your power management settings. There is also a WD utility named WDIDLE3 that can set the idle3 timer so that the drive powers down automatically after a period of inactivity. Many people don't like this feature, but you might consider the frequent autoparking as a trade-off against thermal problems.

http://malthus.zapto.org/viewtopic.php?f=36&t=125&p=175
http://malthus.zapto.org/viewtopic.php?f=36&t=125&p=153

[BlackST]

Measure THE RIGHT WAY the temperature of the chassis.
In any case it has nothing to do DIRECTLY with the 'error growth' .
P.s. The drive is very old and it is pointless to try to squeeze out some more life out of it as it is already degraded - albeit possible.

[positivebit]

This is a typical failure on that series, one head is failing slowly but constantly, i have seen many on that family of drives.

[elmad]

Thanks all for posting. I understand that the disk is failing ant it's just a learning thing.

I guessed that the problem is overheating because after that I wrote the disk while venting it, scan operation went fine. And I think that the error is generated at writing and not reading. I mean, it seems that when it's overheated it writes wrongs ECC in some sectors, so at the next read generates pending sector. Instead a scan operation when it's overheated (but writed at good temperature) generates only 1 or 2 brown sectors.

Maybe as you say it's a failing head but is it possible that a failing head has a different behaviour at reading and writing? And, apart that, if a disk has 2 or more platters, how are "partitioned" the platters? For example, a 500GB disk, 2 platters: first platter is 0-250 and the second 250-500? Or the first platter has even sectors and the second one has odd sectors? Or...?

[elmad]

Well, if you want to spend some money you can try to get another PCB and copy ROM from your PCB to the new one to make it compatible.
I still think that it might be some problem regarding RAM chip on PCB or any other PCB problem that apears when temperature got higher. If you want to try something i would start by swaping the PCB, at the risk of loosing the money of another PCB and the problem could be with heads ...

Interesting experiment but for now I will not spend money to study this disk, because it could be also a mechanical problem. For example, I thought that something inside the disk could be barely inside the tolerance due to use, and overheating send it out-of-tolerance.