Barracuda XT, Barracuda, Barracuda Green, DiamondMax and other desktop SATA and IDE drives

[momo971]

Dear Seagate,

I bought a 7200.12 1TB 8 weeks ago.(CC34 Firmware)

This drive, in vertical position in my computer case, started to make some LOUD clicking noises, until the computer wouldn't POST anymore with the drive connected to the motherboard.

Not to mention the 700 GB data loss ... and the 12€ for RMA shipping.

I received the new drive yesterday (CC35 Firmware), and after 8 hours of use it started to make some LOUD clicks again.

The click sound goes like:

CLICK-CLICK (the drive spins down for 1 second and start again for 1 second) CLICK-CLICK (the drive spins down for 1 second and start again for 1 second) ... forever

My computer becomes unresponsive while the drive clicks, so I umounted the drive and put it in horizontal position outside the case to do some tests.

No more CLICKS.

I can reproduce this behavior just by mounting it in vertical position.

It's happenning in both Windows 7 32bits and Ubuntu 9.10 x86_64

Benchmarks with HDTune under Windows 7 reports very different access time and burst rate when comparing vertical mounting and horizontal mounting.

Vertical: 

Access Time: 25 ms

Burst Rate: 80 MB/sec

Horizontal:

Acces Time: 17.5 ms

Burst Rate:  115 MB/sec

Computer:

Case: Cooler Master Cosmos 1000

Motherboard: GIGABYTE P55-UD3 (F5 Bios)

CPU: Intel i5 750

RAM: G.Skill DDR3 10600

VGA: ATI HD4870

SSD: CORSAIR 32GB

HDD1 : SAMSUNG 250 GB

HDD2: Seagate 7200.12  1TB

All my drives are in vertical position, I don't understand why this drive is clicking so loudly and making my computer freeze while it's in vertical position.

[fzabkar]

That's very peculiar behaviour.

Can we see the performance graphs? SMART data?

Have you experimented with upside down mounting, or inverted vertical mounting?

Could a faulty shock sensor produce this kind of behaviour???

The following article may help you idnetify the components.

HDD from inside:

http://hddscan.com/doc/HDD_from_inside.html

[brucetyler]

Hi Momo
I have exactly the same issue with TWO external Seagate 1TB units I bought at the same time in mid 2009 here in Australia.
Both have the click of death while in the vertical position. Ironically, this is the way the external housing is made to have the drive sit.

One has had the problem from new, the other has only developed it recently.

Seatools doesn't detect any issue with the drive in the short or long tests.

As this is so obviously a hardware issue I'm going to email Seagate in Australia about it, and will report back with the response I receive.

Bruce

EDIT:

Have since tried the helpful suggestion above of vertically inverting the drive - sits just as well either way so why not.
Well, it's been running a backup for two hours perfectly! Will report back if any further issues.

Bruce

Message Edited by brucetyler on 01-10-2010 01:14 AM

[momo971]

This drive is now dead after 29 hours of use...

I'm starting a second RMA procedure

I Hope my next "Certified Repaired HDD", will not let me down that fast

[BennY-]

Just want to add that i have exactly the same problems, (vertical mounted 31000528as) and after reading your post i don't think my data is safe on this drive  

first drive dropped from raid after 3 weeks (>6000 reallocated sectors)

rma certified drive was dropped from raid after 5 days  (and now, it's clicking)

I've found some posts with the same problem!?

 http://forums.seagate.com/stx/board/message?board.id=ata_drives&thread.id=17717

 http://forums.seagate.com/stx/board/message?board.id=ata_drives&thread.id=17081

Message Edited by BennY- on 01-16-2010 04:32 PM

[momo971]

---

fzabkar wrote:
That's very peculiar behavior.

Can we see the performance graphs? SMART data?

Have you experimented with upside down mounting, or inverted vertical mounting?

Could a faulty shock sensor produce this kind of behavior???

The following article may help you identify the components.

HDD from inside:

http://hddscan.com/doc/HDD_from_inside.html

---

I didn't have time to make much graphs, but i promise the next drive will be tested with HDTune Pro on all possible positions.

7200.12 after hearing the first clicks, dismounted from case and in horizontal, Reallocated Sector Count (smart) ~76 if I remember correctly:

Looks normal right ?

A few hours later in vertical position, back in the case, Reallocated Sector Count 488 :

I got error code 98EDD8F9 from seatools

I expect the third 7200.12 drive within two weeks...

To be continued

Message Edited by momo971 on 01-20-2010 12:09 PM

Message Edited by momo971 on 01-20-2010 12:19 PM

[fzabkar]

Your results are very interesting. I was curious about alternative orientations because I seem to recall reading about preamp failures involving bad solder joints. The read preamp is attached to the head stack inside the HDA on a flexible PCB. I was wondering whether the different orientations were allowing the solder joints to open and close ???

The HDDScan article also talks about "microactuators - special piezoelectric or magnetic devices which can move or rotate slider, it helps tune up heads position under a track". I wonder if these can malfunction in different orientations ???

BTW, the SeaTools error code, 98EDD8F9, is unique.

About SeaTools Test Codes:

http://www.seagate.com/ww/v/index.jsp?locale=en-US&name=wvc&vgnextoid=573797f40cce3210VgnVCM1000001a48090aRCRD

"SeaTools Test Codes are based on the type of test that failed and the product serial number. No two codes are the same."

From what I've managed to deduce, the "F" in the test code designates a failure in the long DST test, probably due to bad sectors.

[solution]

keep backup data!

[momo971]

I received my third 7200.12 1TB drive and it's probably going to fail again.

After cc34 and cc35 i received a cc37 firmware.

I did a quick HdTune Benchmark in vertical positioning, with the drive not formatted, power-on time less than 5 minutes (Reallocated Sector Count = 0 at this time):

As you can see performance is terrible.

Six hours later, the drive started clicking, as usual with the 7200.12 Fail-accuda series.

After running Seatools auto long drive test HDTune S.M.A.R.T readings reports Reallocated Sector Count = 1

(According to Google Labs white paper on Hard Drives Annual Failing Rates, this exact situation give me a 60 days maximum lifetime for this drive.)

Let's run another benchmark in vertical positioning:

Access time is really bad.

Repositioning the drive horizontally, a few minutes later:

Access times are now normal. Just by mounting the drive horizontally ...

I can't explain this behavior but I've reproduced this situation 3 times now.

I'm sure this drive will fail within hours or days in vertical positioning, so i will let it die.

I have in the same PC a MAXTOR STM3250310AS 250 GB hard drive with 6769 power-on hours, and absolutely no sign of fail, my computer is 2 months old.

I'm considering buying a [ other brand ] drive. The ones who have longer warranty and reliable products

[fzabkar]

Amazing results!

However, I don't believe access time is the actual problem.

If you think about what the access time graph should look like, it should be a monotonically increasing band of data points about 8ms wide. This is because the access time consists of the time taken to seek to the target track, plus the rotational latency, ie the time required for the target sector to pass under the read head. The latter varies from 0ms to 8.33 ms (= 1 / [7200 RPM / 60]).

If the first read attempt is unsuccessful, then the next retry will occur 8.33ms later. If this second attempt fails, then the third attempt will occur 8.33ms later again. So you would expect to see a graph consisting of several bands of data points separated by 8ms. The actual separation looks more like 10ms.

Therefore I believe that the spread of data points correlates with single and multiple read retries. In fact even your "good" horizontal graph suggests that single retries are occuring in that orientation.

Message Edited by fzabkar on 03-02-2010 01:53 PM