HDD Discussion

[mojavewolfpup]

Mod comment : split from another thread to be HDD Discussion

[mojavewolfpup]

Re: DFI Lanparty 925X-T2 Motherboard

http://www.astro.ufl.edu/~ken/crash/

[Rainbow]

Re: DFI Lanparty 925X-T2 Motherboard

Wow, never seen that. I know that there is a firmware update for these drives. I've recently removed one 60GXP from a server that was running 24/7 and it was still working (replaced it because it was DeathStar). Then I updated the firmware. And then ran some tests - there were already some unreadable sectors! Zero fill fixed that but the drive is probably not good for any serious use anymore (the bearings are going bad too).

[willawake]

Re: DFI Lanparty 925X-T2 Motherboard

It appears that the head crash scrubbed the drive so clean that the magnetic substrate has been wiped clean off the platters.

is that for real? i am sorry i dont believe that. i am aware of the problems though. i do have one but it is not in any real use anymore, did not give me problems.

what did i have then, well at home i had a maxtor which gave me my first experience of extremely slow transfer and no other issue, so got retired. i have had numerous quantum fireball in the office suddenly ping ping and gone, particularly if you run msft scan disk it is bound to trash them. i have other fireballs still running today. actually my father has been using the same floppy for years and years, i just give up on that issue.

what concerns me, well it is the WD that concern me. i have seen too many recently including my personal disks to start whining early and then after a few months fail. i will buy only seagate now although they are quite noisy making a classical sound. i dont mind that though. i am watching samsung also, they got good reviews in a 2005 magazine annual poll, over 90% of users were satisfied

i have some zalman fans in my main pc which i commented in another thread that they have sleeve bearings, well at certain low speeds they have developed one by one a hdd fail type whine so they are getting swapped out. being an admin that sound just freaks me out too much.

After removing all of the platters from the unit, I noticed that there sure was a large amount of dust inside of the drive housing.

ah ok i believe it now, there should be no dust whatsoever inside

[gonzo0815]

Re: DFI Lanparty 925X-T2 Motherboard

I more convinced in Samsung HDD`S, i had been never a problem with those since my first 60gb 5400rpm drive. With maxtor i have made bad experience, on died with a exploded Chip, the other developed bad sektors after 2 week (both diamond max 9, hydro bearing). The replacement works fine so far. WD is similiar, i didn`t thrust them, but it is more a feeling nothing i can prove.

[arneson]

Re: DFI Lanparty 925X-T2 Motherboard

This mention of hard drive failure makes me cringe.
Any of the other bad components are easily and painlessly replaced but years of data going to oblivion is bad.

[Per Hansson]

Re: DFI Lanparty 925X-T2 Motherboard

willawake; the link provided by mojavewolfpup is genuine, it is many many moons old... I posted a link to that story years ago on Techspot...

What happened is that the heads crashed into the platters, as you might know all HDD nowdays are built using glass platters, it used to be aluminum but the gravity force on spinning aluminum was too big so it became unreliable at higher speeds/capacities

Then the glass is coated with a special magnetic alloy which the HDD head can read and write from, but when the HDD heads crashed, it could not find it's starting track, so it goes from the inner to outer track looking for it, again and again and again...

This HDD was probably in this condítion for many weeks for this to happen, but that is not suprising if it was some old forgotten machine...

So the dust in the harddrive is actually THE DATA, it was wiped clean off the surface of the platters by the heads!!!

Talk about doing a clean wipe!

[Rainbow]

Re: DFI Lanparty 925X-T2 Motherboard

I've bought 120GB Samsung SP1213N when my 40GB WD400BB failed. It works fine, the motor and bearings are quiet. We have also older 7.8GB Samsung in other machine and it works fine too.
That WD drive failed in a weird way. The bearings/motor started to make high-pitched sound. Later, when running defrag, it hanged with HDD light on. After reboot, some data was unreadable. So I bought the Samsung and copied everything (including Windows). Saved everything except one picture.
After that, I've done zero fill in WD utility - and no more bad sectors. Then I've done some tests and it seems that the drive works fine. But when it heats up, it does weird things and is not able to write, SMART goes bad. But when I turn it off to cool down, everything is OK again, it does not even remember that SMART was bad.
I have also WD300BB - it does not make the high-pitched sound and works fine.

[kc8adu]

Re: DFI Lanparty 925X-T2 Motherboard

i have a ibm deathstar that ground itself to dust like that.customer brought it in to see if i could repair its electronics and recover the data.the repetitive clicking was due to the inability to load its firmware from the platter.almost nothing on the scope on head amp output so i figured the head stack would need to be swapped.customer provided an identical drive and stayed to watch the operation.
i lifted the cover and stated you are so fucked!
all platters were completely clear.
this thing had been running for a month offline in a remote location and there was no backup.
where is my data?
i dumped the dust into a ziploc baggie and handed it to the customer.its here but you need to reasemble the 1's and 0's in the proper order!

[willawake]

Re: DFI Lanparty 925X-T2 Motherboard

damn that stuff is crazy, most interesting issue i heard for a while.

here is the 2005 reliability survey from UK Pc Pro magazine

Samsung (6 points)
Maxtor (5)
Hitachi (4)
Excelstor (3)
Seagate (3)
WD (2)

Excelstor is the company which makes the ibm drives, ibm storage division was sold to hitachi. Excelstor was permitted to make their own branded drives using the same technology i think.

The article says almost 90% of respondents who owned or currently owned a samsung said it never failed or needed repair, 8% said it failed and 3% needed repair. Then 87% of maxtor owners didnt have fails either.

One thing to consider i suppose is that Samsung hasnt been as long in the market as the others so that might slew the results a bit. I guess that the maxtor owners do not include dell owners with the infamous maxtor slim drives inside.

[Rainbow]

Re: DFI Lanparty 925X-T2 Motherboard

Samsung HDDs are not that new - they were just not widely known - see http://www.samsung.com/Products/Hard...cyProducts.htm

[Chris1992]

Re: DFI Lanparty 925X-T2 Motherboard

>Bad caps on Socket 7 boards are rare.
no, I have a Soyo with bulged Tayeh awaiting recapping.

About the HDD stuff: I had 2 WD fail on me recently, 120GB and 20GB parallel ATA. The 120GB, just before the fail, was doing wonky stuff like reporting the temperature as 0°C and spinning down randomly. The 20GB gave no indication of the fail. I'll use these fails as an excuse to buy myself a nice Samsung or Seagate SATA for quietness.

[arneson]

Re: DFI Lanparty 925X-T2 Motherboard

Ebay has lots of hard drives, as is.
I waisted a little money on some just to see how bad they were. They were.

[willawake]

Re: DFI Lanparty 925X-T2 Motherboard

well i got another wd120 whining now and keeps the hdd light on constantly when attached. i am backing it up to dvdr. i wil get 2 samsung 200gb cos that will enable me to deal with another supplier and i can order some other interesting stuff i want that i could not get before. i am finished with wd now.

whatever wd i dont want i will benchmark until they die and send them back to wd. any ideas how to work a hdd to death?

and yeah Jim, secondhand hdds no thanks.

[willawake]

Re: DFI Lanparty 925X-T2 Motherboard

Samsung HD300LD, SP2514N, SP2504C Hard Disk Drives of the SpinPoint T133 and P120 Series
http://www.digit-life.com/articles2/...g-hd300ld.html

[linuxguru]

Re: DFI Lanparty 925X-T2 Motherboard

The IBM 60GXP is OK (at least the lower capacity 20 to 60 GB ones - all of mine from 2000/2001 are still running, but I don't run them 24/7). The 75GXP is supposed to be terrible, but I don't have a single one.

Try updating the firmware on the 60GXPs - IBM provided an update that wiggled the heads slightly during idle rather than let them stay on a single track. This apparently fixed the problem of the platter lubricant thinning out at that track, leading to failure. My 20 GB 60GXPs accepted the update, and have been happily working without issues. The 60 GB already had a higher rev. and would not accept the update, but it's OK too.

The other fix is to remove the drive PCB, locate the solder bumps that make contact with motor and head electronics, and reflow them with fresh 60:40. This apparently fixes another common failure mode on the 60GXP.

I also believe that these drives were more sensitive to regulation on the +5v & +12v rails, so were more vulnerable to bad caps in the PSU; but the problems were blamed on the drive, not the PSU. I've observed that 60GXPs rarely failed in IBM systems with Delta PSUs, and most reports of failures came from retail 60GXP drives which may have gone into assembled systems with flaky PSUs.

I have a bunch of slim Maxtor Diamondmax Plus 8 units (20-40GB), as well as a U-series 60 GB Seagate ST360020A, all without issues. I have not extensively used a pair of WD 400BB (IBM sticker), so no opinions so far. The only dead multi-GB drive that I have is an old 27 GB Maxtor 92720U8 dating back to 1999 - it could have been exposed to bad PSUs as well as shocks during transportation, so no useful info here.

The general observation that I'll make is that cheap, entry-level, single-platter, 5400 RPM drives are far more reliable in the field than expensive, top-of-the-line, multi-platter, 10k RPM drives with exotic state-of-the-art magnetics. Never buy a drive with bleeding-edge materials tech in it - allow the tech to mature across 3-4 generations and stabilize, and then evaluate its reliability.

[willawake]

Re: DFI Lanparty 925X-T2 Motherboard

I also believe that these drives were more sensitive to regulation on the +5v & +12v rails, so were more vulnerable to bad caps in the PSU

what is common failure mode of hdds due to poor power regulation?

[linuxguru]

Re: DFI Lanparty 925X-T2 Motherboard

> what is common failure mode of hdds due to poor power regulation?

Just guesses here:

1) overvoltage, killing the controller/interface ICs
2) noise on Vcc causing interference on the analog circuits, including read-head pickup lines
3) improper operation of Servo and voice coil drive circuits caused by bad regulation
4) possible overwriting of factory-formatted servo tracks caused by mis-positioned write-head.

This stuff is arguably the most complex and sensitive instance of electro/mechanical tech, with the tightest tolerances on almost everything measurable, that we're ever likely to encounter in day-to-day living, but we tend to take it for granted. With a little care, it can be made to last for much longer than its designed lifetime.

[willawake]

Re: DFI Lanparty 925X-T2 Motherboard

Here is a question that my local shop assistant answered incorrectly. Are SATA2 HDDs backward compatible with SATA1 controllers? The answer is YES. However you can experience issues like system not recognising the new drive or if you connect one SATA1 drive and one SATA2 you can get system lockups. This is because your controller does not support SATA autonegotiation speed.

So you have to add a jumper to force the drive to sata1 1.5 gigabits.
http://www.seagate.com/support/kb/di...sata_lock.html

[willawake]

Re: HDD Discussion

well the jumper works. i have a 7200.9 sata2 running on a sata1 controller happily at the moment testing. Actually the 7200.9 series are the most available seagate series here currently. I was surprised to get a 120gb pata today and see it was also 7200.9.

Unfortunately there is something concerning the series that might bother some people. I had bought a 250gb seagate recently and have been using that on my main pc. I noticed that it was making a high frequency sound, i cannot recall if it was making it from the beginning but it became obvious to me. The sound is similar to a WD hdd whine which develops after some time and which i usually consider to mean the hdd will have an early failure.

So i thought it might fail early and so would replace it with a new one and retire it to less mission critical duty. I bought some seagate 7200.9 sata2 hdds thinking that they would be the newest technology. then i heard the same high frequency sound on both new disks i was testing.

It appears this sound is normal. the funny thing is that it comes and goes in volume periodically. it does not occur in idle and actually the 250 sata1 i was worried about is a 7200.9 also!

then i found SPCR discussing it in their review
http://www.silentpcreview.com/article283-page1.html

i guarantee that some people wont like this sound, especially the quiet crowd. personally if its normal then i will get used to it. i dont want to buy more WD at the moment and Samsung were a problem to get.

other impressions

seagata 7200.9 PATA 120gb (Singapore) - the drive had some vibration, it was not excessive.
seagate 7200.9 SATA2 200gb (Thailand) - there is almost no vibration on both drives

the sound of the disks reading data is quite loud also. i think louder than WD but it is a nice sound IMHO. being an admin i like the sound of hdds reading data and FANS also. but sounds which i normally associate with failure are not something i would prefer to have around.