Sign in with
Sign up | Sign in
Your question

Hard Disk Failure / New Hard Disk selection

Last response: in Storage
Share
July 4, 2008 10:07:45 PM

I have had very bad experiences with seagate 7200 Barracuda drives of 80GB and 1000 GB capacity recently.

4 drive failures in 12 months. Clicking disk failure...i.e...the internal mechanism went bad.

It seems WD drives are better than Seagate.

Want to buy a new Western Digital hard disk of low capacity, preferably a 5400 RPM drive.....

Which WD 5400 RPM drive do you suggest? A low capacity drive that will not die on me....

Please reply soon as i need to make some urgent decisions.
July 5, 2008 12:42:18 AM

Just curious why 5400 rpm and not the more standard 7200 rpm? I haven't had many bad experiences with hard drive failures, i've recently bought a couple seagate HD's and I hope they hold up better --- but whomever makes a suggestion probably needs to know if it is IDE or SATA connections.
Anonymous
a b G Storage
July 5, 2008 1:00:51 AM

well I think all mobos now (unless its really really really really really really new) support IDE... so... but yeah I don't understand why you want a 5400 rpm drive... you can get massive 7200 rpm Seagate drives that are the new 7200.11

its an updated firmware that supposedly is very good
Related resources
July 5, 2008 7:09:52 AM

supremelaw, you didnt comment on 5400 rpm drives versus 7200 drives.

My experience with 7200 Seagates (barracudas - both IDE and SATA) has been a nightmare in the last 12 months. 4 failures (2 X 80 GB ide, 160GB sata, 1TB SATA).

So much so that I dont ever want to hear the word 'seagate' again!

What about WD in terms of reliability? What about 5400 rpm WD drives? wont they be better and more reliable because of less wear and tear /operating temperature?

Anyone else who wants to comment on my response, most welcome. I am in a bad shape. Seagate makes me want to throw up.


Quote:
If you can find them ...

WD Caviar® Black™
1 TB & 750GB, 3 Gb/s, 32 MB Cache, 7200 RPM

http://www.wdc.com/en/products/products.asp?driveid=488

See Specifications, e.g.:

Transfer Rates
Buffer To Host (Serial ATA) 3 Gb/s (Max)
Transfer Rate (Buffer To Disk) 145 Mbytes/s (Max)


2 of these in a RAID 0 should really scream,
PROVIDED that the RAID controller is on the PCI-Express bus
(not the old PCI bus).


Sincerely yours,
/s/ Paul Andrew Mitchell
Webmaster, Supreme Law Library
http://www.supremelaw.org/


July 5, 2008 8:00:11 AM

Your understanding that a small 5400RPM drive is necessarily less prone to failure is inaccurate, although you will grow old while it accesses your data. Sorry you got burned by Seagate. I have seen that the 7200.11 drives, which are touted as being very fast, are very prone to failure. Also there are other things that could cause HD failure, the first suspect is a garbage PSU. If you are he!! bent on having a smaller drive then partition it.

Get the Western Digital Caviar SE16 WD6400AAKS 640GB 7200 RPM 16MB Cache SATA 3.0Gb/s Hard Drive
It's OEM so make sure you have a SATA cable.

Low Price and Great Performance--Western Digital WD6400AAKS Review
a b G Storage
July 5, 2008 8:03:12 AM

Sounds to me like you may have a motherboard controller problem or some other issue. It's really hard to believe you would have that many hard drive failures in such a short period. Seagate drives have 5 year warranties I would just send them back and try them on another system.

This might not be a 5400rpm drive but should do the trick if you want a new drive.
http://www.newegg.com/Product/Product.aspx?Item=N82E168...
July 5, 2008 8:42:53 AM

fauxpas1900 said:
I have had very bad experiences with seagate 7200 Barracuda drives of 80GB and 1000 GB capacity recently.

4 drive failures in 12 months. Clicking disk failure...i.e...the internal mechanism went bad.

It seems WD drives are better than Seagate.

Want to buy a new Western Digital hard disk of low capacity, preferably a 5400 RPM drive.....

Which WD 5400 RPM drive do you suggest? A low capacity drive that will not die on me....

Please reply soon as i need to make some urgent decisions.


I find it hard to believe that you had a drive failure on the average of every 3 months. No matter what the MFG of the drives. Drives fail for other reasons, such as bad RAM causing dirty writes, bad PSU's causing premature motor failure, bad wall voltages or ground causing PSU voltage specs, etc. Note that the RAM will not actually cause a drive failure, but it may be perceived as such. And dirty noise on your 12V side of your PSU may not be fatal as well, but is more likely to be.

So how did you determine that these drives failed? Seatools would be my first choice, especially as I'd be looking to RMA them

Note that if you have dirty wall power or a flaky PSU or bad RAM, changing drive MFG will not be an improvement.

Google did a paper, "Failure Trends in a Large Disk Drive Population" that they presented to the usenix conference on drives and storage. While they would not give MFG statistics, the end result was that while their reported trends might shift a bit with one MFG VS another, over all there was not much difference.

I know that I have not helped you in your search for a 5400 RPM WD drive search, but hope that I have given you pause to look closer at 'why' you have lost so many drives in such a short while.
July 5, 2008 12:32:08 PM

Hi croc, thanks for the response. I was thinking along the same lines. Faulty power supply.

I checked the power supply voltages using a multimeter. The power supply is OK.

There is no grounding problems in the cabinet either.

What do I do? I dont understand what you mean by 'dirty noise'. What is to be done to determine and fix the 'dirty noise'.

I have some RAM testing tools on something known as the ultimate boot CD.

So how to proceed? How to know whether the RAM is causing physical disk failure?

Some of the things you wrote that I did not understand:

"Note that the RAM will not actually cause a drive failure, but it may be perceived as such. And dirty noise on your 12V side of your PSU may not be fatal as well, but is more likely to be."

I checked the drives I mentioned with SEATOOLS - they all fail the DST (drive self test).

If the PSU power supply is correct (using multimeter) and there is no grounding problem (none that i have detected so far) what is the conclusion?

All these drives that I mentioned make funny noises (motor struggling with something). And then they hang. One of them refuses to work totally.

The 1 TB Seagate went bad within one month of purchase. The others lasted longer.

July 5, 2008 1:33:41 PM

Hi Zorg, yes, I heard the same thing - 7200 Seagate drives are a bad product , prone to failure.

I am looking for an 80GB drive. The one that is available in WD is a 7200 rpm drive named BD800. What do you think of it?

Hi PsyKhiq - How could a motherboard controller cause physical failure in a drive? [clicking disk failure, drive failing the Drive Short Tests etc.]

_____

I checked the power supply of the PSU - it gives a stable 11.7 volts......could that be a problem? 11.7 volts instead of 12?

_____
July 5, 2008 10:10:31 PM

What model PSU are you using? By saying BD800 I assume you mean WD800. That is not specific enough, can you provide the full model name and link if possible? are you looking tor SATA or PATA?

Also what's the power like where you live?
July 5, 2008 11:44:41 PM

When people talk about "dirty power" they usually mean that the line voltage and frequency fluctuates. It may not be readily apparent so it's a tough call. I happen to use a UPS that has "line smoothing" to maintain voltage and frequency within specs. Zorg is hinting that your recurrent probem may be due to a decrease in line voltage that is also causing your power supply to give fluctuating votage to the hard drives.

The problem with getting an absolute answer is that you need to take measurements under a load condition. That may be almost impossible for the average person to do without some sophisticated equipment. Capacitors and other electrical/electronic components in your power supply age and degrade and affect power supply stability. The best way to check that is to take the power supply out and have it checked by a reputable shop.

For line voltage, an indication of instability is lights dimming when many appliances are plugged in or turned on or if the outlet you use for the computer also has a motor or high load component on the circuit (refrigerator, fan, toaster, etc).

The fact that you've received so many failures indicats that there is a causal factor outside of the brand of hard drive you're using although I can certainly understand your reluctance to go back to Seagate after you've been bitten so many times - their fault or not.
a c 167 G Storage
July 5, 2008 11:53:52 PM

1) Why in the world would you want a 5400rpm drive when faster drives perform so much better and do not cost any more?
2) Something is amiss in your system. In good conditions, the MTBF for a drive is supposed to be a million hours. That's over 100 years. To get 4 failures tells me something is wrong.
3) Excessive heat can shorten the length of drives. Is your drive cooling good?
4) Identify your PSU and whatever else is in your system. A low quality PSU can deliver flaky power under conditions of heat and load. A voltmeter test is only the most basic of tests.
5) Are you overclocking? Bad writes can sometimes occur.

Try a complete format of the drives to see if they are recoverable. Otherwise, RMA them. Sell the RMA units if you are uncomfortable with Seagate.
a c 353 G Storage
July 6, 2008 12:42:34 AM

Additional insight on “Dirty power”

Line fluctuations and spikes can cause Noise on the output as pointed out in previous post. The cheaper the PSU the more noise on the output. Another culprit is the switching Power supply itself. This is the biggest drawback to switching Power supplies (switching noise riding the output) and why audio “freaks” pay big bucks for “old” analog series regulated PSs. A multimeter will not show this noise in DC mode. Sometimes you can see this by using the AC mode, should read zero; but alas this will not show spikes – for that you need an o’scope.

11.7 V at idle is a little low in my books as the loaded value could be dropping even lower. 12 – 3% = 11.64 V. Down load cupid as it shows min – max in addition to present value. Then run a program to load the 12 V. I use ATI tools and run their “View 3 D mode. While monitoring the min value
July 6, 2008 4:03:59 AM

So, do you have the PSU model, HD model that you want, and info on your local power?
July 6, 2008 6:00:52 AM

See if you can borrow a recording voltmeter and monitor your PSU while it is loaded heavily. A Fluke 87 or equivalent would do the job nicely. The Fluke would also be able to check the RMS of the DCV output. May be more cost effective to just egg it and spring for a new QUALITY psu especially if your is several years old and you intend to keep it running for sometime in the future.....your call.
July 6, 2008 6:37:33 AM

You can get a psu tester to see if everything is okay.
July 6, 2008 9:20:04 AM

Hi all, thanks for your wonderful inputs.

I think I am going to trash this cabinet/PSU and get my hardware shifted to a new cabinet with a new PSU.

My old PSU is 2 years old. I am from India - power conditions from the supply are quite ridiculous.

But I have an APC (www.apc.com) UPS that has surge protection in it. And provides a stable 230-233 volts AC to the computer.

There are frequency fluctuations in the outside power from time to time but I have no way of measuring that.

_________________

Any suggestions on how to get to know whether the new PSU is a good one? Is using a multimeter enough? I have absolutely no idea which PSUs are availbale in the market, my side of the world - but please do suggest some reliable brands that I will try and search for. OR maybe I can buy a new PSU from some hardware shop like newegg. But i dont know which PSU is reliable.

It would be nice if it was possible to test the new PSU before unleashing it on my system I guess....! Is that possible using some software?

My current PSU model/make is something known as 'NAV TECH ATX450T'.
________________________________________


Geofelt - You asked "5) Are you overclocking? Bad writes can sometimes occur.". How do I know whether I am overclocking? I havent done anything intentionally. Can overclocking destroy a hard disk?

Zorg - here's the WD800BD drive's link: http://www.wdc.com/en/products/products.asp?driveid=161&language=en.
This is the only WD model that is available at the shops in the 80GB category - i am looking for an 80GB - to keep my winXP on it and nothing else.

My current PSU model/make is something known as 'NAV TECH ATX450T'.

RetiredChief - You wrote "Down load cupid as it shows min – max in addition to present value. Then run a program to load the 12 V. I use ATI tools and run their “View 3 D mode. While monitoring the min value".

Where do I download 'Cupid'? And can you re-explain what you have written in the above sentence?
July 6, 2008 12:48:14 PM

The WD800BD should be a good drive. If it fails then you definitely have another problem. Given the large number of failures you probably do anyway, as was said several times already.

The UPS should clean up bad power from the wall, I suspect it is a must in India.

I can't find anything about your PSU, I wouldn't trust it. Get a new PSU before you install the new HD.

I don't know what is available over there but I think you should be able to get a Corsair, they are good PSUs.

I don't know what size because I don't know your configuration.

The VX550 should cover you nicely and give you some headroom since you are using a 450, but that's just a guess.

a c 353 G Storage
July 6, 2008 3:11:52 PM

Here is the link to CPUID's HWMonitor. http://www.cpuid.com/hwmonitor.php

On my system the 12 V is 12.1 V This is at idle, not much running.
If I run ATI tools ( free download, use google search) my 12 volts drops to a steady 11.84 V (0.26 volt drop). Graphics cards are the single biggest load on 12V Rail. The 2nd is the Processor.
You can also open up a DOS window and do a "dir x: /s" where x is your drive with the largest Number of files. My 12 V droves Momentarily to 11.97 then back to 12.1.

Zorg gave you good info on HDD's
a c 167 G Storage
July 6, 2008 7:26:45 PM

Overclocking is the process by which the cpu speed is increased beyond the original specifications. It is done by selecting new parameters in the BIOS on certain motherboards. It is most unlikely that you did it unintentionally, and even if you did, it would be unusual to harm the hard drive.

Here is a link to a list of PSU's tiered by quality, at least according to some opinions. If your PSU came with your case, it is probably not very good. If you like the case, just replace the PSU. If you want a new case, buy the PSU separately. The only exception might be Antec, which puts decent PSU's in cases. http://www.xtremesystems.org/forums/showthread.php?t=10...
July 6, 2008 9:15:50 PM

Hi Zorg, RetiredChief, geofelt, and everyone else who has responded.

Thank you all so much for your valuable suggestions. I am very grateful.

I bought the WD BD800 since the last of my seagates is failing and i'd be without an OS without the WD; WD seems to be running fine.

Zorg - On what basis do i decide what PSU I need? You mentioned VX550. My PC stays on about 15-16 hours a day.

Geofelt - Thank you for the list of PSUs. Will show the list to my supplier.

RetiredChief - I downloaded CPUID but its not showing the 12 V at all. Does this mean my PC does not have the sensor for 12 V? or is it something else?

Here's a screenshot of what CPUID shows on my PC. Does it look ok? I cant figure out what is what! (One of the failing Seagates is attached currently, as the image shows.)





My supplier mentioned a brand named 'EYEBALL'....in PSUs. I dont know where it stands on quality, cannot find it on the net.
July 6, 2008 10:41:22 PM

This screenshot is of my system. Measured by 'hardware sensors monitor'.

I think we may have found the culprit.....

The voltage in the following chart is 11.31 V and it should be 12 V.

Is this the problem??? The 11.31 V?? Can this destroy a hard disk??

I will get the PSU changed immediately, but might take 24 hours. Is it ok to run the new WD BD800 for 24 hours on this voltage?

a c 353 G Storage
July 6, 2008 11:28:37 PM

11.3 V under minimal load is low considering it will drop further under load probably down to 11 V or LESS..

Now as you can see, ie CPUID readings, software readings are not always correct. I always verify by using a multimeter. Once I know I can trust the software I feel confortable using it.

And yes low voltage can wipe out your HDD. As Voltage decreases Current increases to keep power constant.

Myself - I would dicontinue usage until you have replaced the PSU.

PS
Multimeters are relatively cheap and easy to use AND are handy for many things. You can get an inexpensive one for about $20 -$25. Put on DC function, insert black lead into the Black pin on a molex connector and with the red lead measure the other two pins. Ones the +5 the other the +12.
July 8, 2008 1:21:18 PM

In order to determine the minimum required PSU size you need to post all of the components in your system.

I've never heard of the eyeball PSU and therefore wouldn't trust it.
July 8, 2008 1:56:38 PM

I got myself a new PSU - named 'CoolerMaster', taiwanese PSU. Am running the machine on it right now.

http://www.coolermaster.com/index.html

It cost 5 times the cost of an average PSU in the market and is rated 390 watts. (Its so expensive, so I guess its a good one. The PSU list given by geofelt- none is available).

Zorg - Corsair PSU was not available either.

Is 390 watts enough? I have a DVD writer/AMD athelon 64 3200+, MSI motherboard, 2GB RAM, NVIDIA 7200 GS graphics CARD 256 MB VRAM, 2 HARD DISKS.

RetiredChielf - The voltage that this new COOLERMASTER PSU is giving me is exactly the same as the old NAVTECH PSU. i.e. 11.31 Volts. The other voltages are also the same.

Screenshot of Hardware sensor monitor with COOLERMASTER (new PSU)




So perhaps now its confirmed that it is SEAGATE 7200 RPM disks that are of bad quality, and not my PC hardware/PSU?
July 8, 2008 2:36:12 PM

Cooler Master is known for making junk PSUs as well. Apparently some of their larger ones are OK, but I wouldn't use them. I don't understand why people ask for confirmation after they buy the part. A list of available PSUs before the purchase would have been much wiser.

The PSU is large enough just not a good brand.
July 8, 2008 3:51:07 PM

Zorg, I couldnt wait because I cannot afford to be without a PC, and I have just one PC. :-(

I can exchange cooler master for something else. What do you suggest?

Here's the other brands that are available here.

1) Microtek
2) Quantum
3) Intex
4) QHMPL (not too sure).

I can also try newegg for a certain brand, but i need to check whether newegg delivers to india, which is doubtful. But in case they do, I can send the Coolermaster back to the supplier and take some other accessories in return.

But Zorg - Both the old PSU and this new Coolermaster are giving the same voltage, as is shown by the screenshot i put on. What does this suggest to you? Is this enough to conclude that most probably the seagates i got were bad.
July 8, 2008 4:58:04 PM

It suggests that the numbers from the mobo are wrong. The only way to verify that is to get a cheap meter and test the voltage at one of the connectors.

Unfortunately, I'm not familiar with any of those brands. You might try Google to see if you can find any reviews that say more than how pretty the box is.
July 8, 2008 6:40:03 PM

Yesterday I asked someone to get me a multimeter from which I will check voltages manually. Waiting for the multimeter to come to me.

Just 25 minutes back I saw a particular column in the BIOS (after pressing DEL on the Boot Screen) that says H/W monitor....

Here are voltages that are shown by the BIOS.


V (12V) : 11.985 V

V (5V) : 5.171 V

V (3.3 V) : 3.342 V

So if the BIOS reports these voltages, perhaps the PSU is giving correct voltages?

I replaced coolmaster with the old PSU just to check voltages from the BIOS, and it reports the same.

Are values straight from the BIOS reliable?
July 8, 2008 6:50:58 PM

They look better. Unfortunately you can't necessarily rely on the BIOS to give you the proper voltages. The only way to be sure is to use a meter, as I think RetiredChief mentioned earlier.
a c 167 G Storage
July 8, 2008 8:40:39 PM

The real test is when the PSU is under maximum load and everything is heating up. Try running prime95 and the voltage monitor to see if those voltages hold up under load. 350w may be ok of there is sufficient amperage on the 12v rails. On the side of the psu is a label with the specs for each of your psu's Can you post a pic of them?
a c 353 G Storage
July 8, 2008 10:35:15 PM

As Zorg and I have pointed out, You really should verify with multimeter.
I would tend to think that the Bios reading is probably correct.

HOWEVER, as geofelt stated, measurement when in Bios is at a reduced loading. I would run ATI Tools, 3d view as this would place a heaver load on the +12V than testing the CPU using prime 95.

1. open up your program for viewing voltages. Connect your multimeter to molex connector Gnd/Rtn -> +12V. Compare Multimeter with software value. This will give you the Delta @ idle. Leaving the multimeter connected, start ATI Tools, 3d view, Monitor your multimeter and software values. You know have a delta for under load. If the multimeter reading drops below 11.5, then I would suspect your PSU, Then either 10 your PSU is poor quality, or 2) your PSU is not capable of suppling sufficent power on the +12 V Rail.

AS ZORG POINTED out research the PSU before you buy!.
July 8, 2008 10:40:35 PM

Am arranging for a multimeter.

Will check with ATI tools 3D tools and post the results here.
May 8, 2009 5:39:46 AM

Bro, i've been using seagate barracuda (7200 RPM) 80 GB for the last 8 years with not even a single problem, must be something wrong with your system or supply plz check.
!