Slashdot Mirror
Tom's Investigates Hard Drive Warranty Changes
Sherloqq writes "Tom's Hardware recently ran a story about major hard drive manufacturers drastically reducing their warranties on many of their products. Effective Oct 1, 2002, many IDE hard disks from Maxtor, Seagate and Western Digital will now come with just a 1-year warranty. This comes as a bit of a shock to me, as nobody seemed to have mentioned that previously (or I haven't been paying enough attention). Spokespeople for the big three cite disproportionate costs of in-warranty service vs. rate of failure, need to cut costs to remain competitive, advancements in technology used in manufacture of drives ("they're so reliable and cheap, you won't need a warranty anyway") as well as warranty period mismatch with OEM computer manufacturers (std. 1-year). Good news in all this: there are no plans for warranty period reductions for SCSI drives. For now... :)"
Seagate and Western Digital are both going to 1 yr warranties for the major potion of their product lines. WD will keep a 3yr on the Special Edition drives.
As of now....IBM is the only company to not announce a change in drive warranty...my guess is that will change once they introduce their new drives.
SIG:Slashdot: indymedia for nerds.
Story from news.com:
http://news.com.com/2100-1040-959831.html
As the article points out (along with several posters above), the warranties on drives in PCs and other devices (the vast majority of HD sales) were already that of the device in which they came, which is generally one year or less anyway.
Honestly, at today's prices I view hard drives as twinkies--they're cheap and they'll probably last 3 years anyway. There's plenty of worse things to get upset about than only getting a 1 year warranty with a $79 80GB 7200 RPM hard drive.
--Len
In the EU we have a minimum warranty of minimum 2 years on all products.
This is a new european law issued 2 years ago and effective since 2002, I think.
angel'o'sphere
Cost free eBook I read (by iBook/Kobo/Amazon/ObookO/Gutenberg etc.): "The Green Odyssey" by Philip Jose Farmer.
I've replaced two WD drives in the past 10 years, or so. I've also witnessed several friends and family do the same, along with a huge amount while I was a teach. In every case, WD took the drive back, even after 3 years, usually with little explanation. I haven't noticed a higher failure rate, but I would buy based entirely on their support.
In most instances, the older drives were replaced with a larger drive and they were sent for free, before I sent mine, so I could return it in the same box. A couple years ago they even had a return postage sticker for the return trip.
You have a two year dealer warranty in the EU, mandated by law.
Did you know you can fertilize your lawn with used motor oil?
Comment removed based on user account deletion
I would guess that temperature is probably a big factor - one of the manufacturers once showed me a graph of failures vs. case temp, and failures basically rose exponetially with case temp above 20C. But it isn't only overclockers who run hot - it is cheap PC builders who save a few dollars relative to the big boys by fitting small fans, or cheap fans which fail silently, leaving the disk to roast itself. Particularly the faster drives generate a lot of heat, and need help to get that out.
If you value your data, it is *much* more important to cool your disks than your CPU. If your CPU kills itself with overheat (and one thing you can say about the Pentiums is that they seem to slow themselves down nicely, unlike Athlons), it is a few tens of dollars, or the low hundreds if you went for the best, to replace. If you cook your drive, not only are you down roughly the same number of dollars to replace the drive, but you have the major hassle of recovering from backups - if you have backups.
I bet few people take image backups of a 40+ Gb drive every day or two: they only back up their crucial data regularly. So you are going to have to go back to your OS masters, clean install the OS. Then recover all the site-based configuration files which you backed up after you set up the system (you did, didn't you?). Then you are going to have to go to last night's backup of hot files and retrieve them. And I bet that, in between times, you installed something else which didn't get backed up, so you are going to have to dig out the install for that (if you remember where you put it). Thhe cost in hassle etc. and time is going to dwarf the cost of a new drive.
Damn - I am talking myself into a Raid very fast.
Consciousness is an illusion caused by an excess of self consciousness.
Nope It's SCSI that puts the I in RAID really.
;)
All mission critical RAID systems use some form of SCSI drive (UW, Fibrechannel etc).
This might be partially because of the higher performance with SCSI in general (yeah yeah I know ATA/133 is nice and fast but my 15000 rpm fibrechanel drives are faster
If I toddle down to our server room I find a huge number of RAIDed 18Gb (total 50-60 Tb) SCSI drives and no IDE ones.
Troc
Troc's dubious podcast and blog: http://www.trocnet.net
They don't. They almost always send you a refurb of the same product. As a matter of fact, I had an 8.4 GB Maxtor fail on me, and sure enough- they send the exact same drive, refurbished, and resealed in an ESD bag. And the drive was at least 4 years old. They aren't going to send you a newer product when they have boxes of refurbished drives that will still make you buy new products when they become too slow for modern software. Just try running a modern OS on an old ATA 33. It'll make you want to upgrade if you've ever had a taste of modern 7200 RPM IDE drives, or even faster SCSI drives.
IBM released a document a while ago showing a significant correleation with failure rate and operating temperature of some of their SCSI drives.
Personally I plan on actively cooling all my drives now. They are a rather important component, after all..
I normally never buy those types of extended warranties. It's been my experience that (with electronics), the part either works and lasts forever or it's broken out of the box.
One exception -- I received an iPod for my birthday. It's the 20GB model and my wife purchased it at CompUSA. The extended warranty they sell you for the iPod is the same warranty they sell you for a regular 20 GB hard drive. So, for $20 it's really not a bad deal for an extra year of coverage.
As below, so above and beyond, I imagine drawn beyond the lines of reason. Push the envelope. Watch it bend.
Same annoncement (3 year to 1 year warranty), on same product (ATA IDE harddrive), on the same day (1 oct 2002). Ring the bell to anyone?
:)
In the normal way of thing, one would have changed is warranty and the 2 others after a couple of months would have realized it was a good idea and would had done the same, not the 3 on the same day!
Secondly if only one of them would have changed is warranty customers and OEMs would have boycut the manufacturer just like IBM a couple of months ago. But now the 3 biggest harddrive manufacturer have changed their policies so you have not a lot of other reliable harddrive manufacturers to choose from (so far less boycut possible).
Come on, go fill a "File a Complain Online" on the http://www.ftc.gov/ Federal Trade Commision you are only 2 click away.
Take 5 minutes to denonce this collusion between the 3 manufacturers (Western Digital, Maxtor, Seagate), the problem is not that they drop the warranty but that they all agreed to reduce the choice and/or the price/quality ratio of the same products(ATA IDE hard-drives) thus reducing the customer in is fundamental right to choose the best product and to be in a fair trading country.
enough
It just occurred to me that people actually believe warranty costs are driven solely by failure rate and replacement drive costs. I guess I have to spell out other reasons that warranty costs could go up for a manufacturer:
1. Employee pay increases. Everyone from the technicians who test the drives to the janitors to the shipping clerks get paid. Sometimes job market conditions force employers to raise pay to attract and retain employees.
2. Employee benefit costs. If a company finds itself with unexpected increases in health insurance premiums, for example, their costs on warranty service rise.
3. Government regulations. OSHA and EPA rules and regulations (for example) might directly affect warranty costs.
4. Facilities costs. If the cost goes up for electricity, heat, water, building leases, fuel, etc., that affects warranty costs.
5. Shipping costs. When shipping costs increase, that directly affects warranty service costs.
Those are but a few of the things that can increase warranty costs even if failures stay constant.
As drives become cheaper and profit margins shrink, fixed warranty costs become disproportionate. It's no cheaper to ship an $89 drive than it is to ship a $300 drive of the same physical size -- and we've seen that kind of price drop. There was a time, not too long ago, when an inexpensive drive was $300. Drive manufacturers are now operating on razor-thin margins and downwards-spiralling prices. When you are making $1 profit on each drive, the shipping costs alone for a warranty replacement will eat up all of the profits for multiple drives.
A longer warranty does not imply a better or more reliable product. Just look at cars. Hyundai and Kia come with 10 year powertrain warranties while Lexus, the most reliable car according to studies/surveys, comes with a 6 year powertrain warranty. So how does Kia/Hyundai offer such a long warranty? They cut costs elsewhere.
I'm willing to sell Maxtor hard drives with five-year warranties if you're willing to pay me $300 for each 40GB hard drive. I'll just go down to CompUSA, buy the drives there, buy some spares, and sit the spares on a shelf. That won't make the drive you get any more reliable, but it will have a longer warranty.
But dollar for dollar, I think over the years IBM has consistently made some of the most solid hard drives on the market. Warranty issues are the best in the industry. They fix and replace. And what did IBM do? They replaced all the bad ones. And still warrantied the new ones for three years. No change made. Hitachi will carry the ball, they have a good core of engineers.
I recently RMA'd two 9.1G SCSI SCA drives. They failed within three months of each other in the same array. I sent them back in a Seagate box with the preformed foam and egg crate. After IBM received the drives, they sent them back, broken still and said that I had voided the warrenty because one of the sides of the foam casing wasn't two inches thick. It was an inch and a half plus. That is rediculous. I know have two worthless IBM drives with no warrenties and don't work. I don't see how that = the best with warrenty issues.
MR
Search for "Sony TV tuner snowy picture", it is not quality you are paying for when buying Sony products.
People still use the tuners in the television? I thought with all the TIVO/DirecTV/digital cable folk we'd all have bypassed it and be using the composite, svideo or component video inputs in our televisions.
>"$115 for a motherboard? Don't you have anything cheaper, like, around the $75 range?"
But Tom's Hardware rated the K7S5A PC-Chips rebrand as great stuff, therefore PC-Chips must make quality products, and would never ship a special top-notch board out to a reviewer.
That's the sort of argument I would get from customers when I'd tell them to buy a nice, reliable, ASUS board for $50 more.
Fortunately, where I worked, warranties weren't always honoured without some serious arguing (no, I don't work there anymore -- it's no fun telling someone to sue us for $25 when the filing fees cost $75) so we didn't have to deal with the $399 CDN "compelete systems" blowing up in our faces.
If you could be told what you can see or read, then it follows that you could be told what to say or think - BoC
Drive cooling is wonderful. IF you like noise. I bought one of the half-height drive cover fan units - three small fans side by side that blow over the drive. Sounded really neat at the store. Sounded like a leaf blower when I installed it.
I leave my PC on full-time. I set the drives to sleep after a half hour or so. So they tend to cool down to ambient when I'm away from the machine. I didn't see any reason to cool a cool drive, and have to listen to it full time.
So I built a temperature sensitive circuit to try and limit the noise. I figured that as the drive heated up, I could spool up the fans accordingly, and keep it quiet longer. There were two problems, one fixable, the other not. The fixable problem was that I pulled the supply voltage from the drive power connectors, but my circuit was not voltage regulated - so as the processor load increased (yep, cpu cycle load), the fan speed changed. I could tell how busy the CPU was by listening to the fan pitch. A simple voltage regulator might fix that, although I'm surprised the voltage changes that much. The unfixable problem was that the drive heats up so rapidly (maybe one minute from idle/cool to spinning/hot) that essentially as soon as I sat down and got the modem dialed up it was howling away.
So I'm not exactly thrilled with the idea of air-cooled drives. Maybe a high-volume low speed fan for the entire case, vented near the drive, would work better. But the simplest solution is probably to stop mounting the drive at the TOP of the case, where the heat accumulates, and instead put it at the bottom.
--Brandon / Split Infinity Music
That statement is incorrect. Anybody that puts together a solid business model would have warrenty costs built into the product's original cost.
It may be true that they are only making $1 one each drive at this moment, but now they will make $2 on each drive because they will have to build in less warrenty cost into the product's price.
Also consider this. They have reduced the warrenty period from 3 years to 1 year
Now, they will make it so the drives only last 2 years (1 year less than their previous warrenty required and one more year than the current one requires). Any one that can't see the basic business behind this decision needs to make a stop to the eye doctor (and please stay off the roads!)
I wish I was the V.P. that got credit for this lesser warrenty idea
HallmarkOrnaments.Com
I have a large installation, and of the recent large
disk installs (~80GB IDE, at least 20 disks each) I have replaced between 15-30%. This disks are from seagate, wd, and ibm and have been bought in the last 3 years, and all had some failures fairly quickly, without there being a real
noticeable quality difference between the three.
After the 1st year, we still are getting a few
failures, not as many, but still a amazingly high
number, mostly with blocks going bad.
The quality is no where near the scsi reliability, either in the long term or
the short term.
Those of you who defend IBM do so wrongly. During the 'bad run' if IDE GXP based drives, which I believe from what I have read invloved GXP 60 and 75 ranges, IBM behaved irresponsibly, and should have come clean and advised customers as soon as it knew it had problems.
After a considerable amount of time, they issued new doctrines, such as 'drives must not be used for more than 11 hours per day'.
In addition, they did not handle warranty coverage correctly, nor customer advice correctly.
Users who shipped bad drives got either the same drive back after IBM had used their utility on the drive, or a replacement.
The joke was if you were a customer they advised you to wipe, using their downloadable tool, and by common account within 3 weeks the grinding noises and data loss returned.
Many people got returned drives, and then lost data again when the new drives failed.
Sending out bad batches is one thing. Sending out bad families of drives is a new scale altogether. Add to that the warranty handling, the multiple returns, the failure to make public the actual issues. The failure to withdraw a faulty product they knew damn well was loosing customers data. Resupplying the customer with said same drives with pretty clear knowledge the drive was a likely failure. Lastly the issue of new guidelines making the problem the customers (ie, daily no more use than 11 hours).
I had 5 of these drives. 3 were replaced. Out of a total of 8 drives 7, that is 7, died, made grinding noises, lost data, etc etc.
The bottom line is now this. I do NOT know if I can trust IBM disk again. I am neutral when it comes to brand. But given that IBM have not publicly accepted the problems, or given the true reasons for failure, OR SAID , on our new family of drives we cured the problem by X,Y,Z, that means until I know for an absolute FACT that IBM make IDE harddisk that are utterly bombproof, I doubt I or anyone I advise will buy an IBM IDE 3.5" harddisk in the future.
I just do not dare to put my data on their drive, and that is the bottom line.
It is a shame as they, looking at www.scan.co.uk come at a good price, good (speed) performance, and one huge gigantic stone round their neck care of the GXP issue in the past.
AdmV
We`re all equal
Nope. Very, very different mechanically. Your typical IDE drive will spin at 5400rpm or 7200rpm for the nicer ones. That difference alone between the IDE drives requires slightly different sets of mechanics (different spindle motors, for instance).
Some numbers to back up my reasoning? Let's compare some basic stats on the various classes of drives. These are sample numbers that I got from various manufacturer's websites. They're more or less typical (within a few tenths of a millisecond) for each class.
Seek Time: 12.0ms
Difference: 6.4ms
Seek Time: 9.0ms
Difference: 4.8ms
Seek Time: 4.7ms
Difference: 1.7ms
Seek Time: 3.6ms
Difference: 1.6ms
Latency, as used above, is defined as the time it takes for the disk to spin through half a turn. It's the average amount that the disk has to wait for the data to be in line with the head. A disk's average seek time cannot be faster than its latency.
So, if you look at the difference between the latency and the average seek time for each platform, you can see how hard the actuator is being driven. If a 5400rpm drive has a difference of 6.4ms, while a typical 10k drive has a difference of less than 2ms, it's pretty obvious that the mechanics are going to be radically different. For SCSI drives, the magnets get beefed up, the spindle motors get better bearings, etc.
Additionally, the major OEM customers that buy the SCSI drives have much more stringent standards for them than the ATA drives they also buy. If you're selling some schmoe a $1000 computer with a standard 1 year warranty, you really don't care if the drive lasts longer than one year. Meanwhile, if you're selling a company a $20k server, complete with a service agreement and such, you want those components to be as reliable as possible. You're going to give the manufacturer a much tougher set of standards (temperature, shock, etc) to meet before you OK their drives in your product.
To wrap this up, you asked if the main cost difference came from just the different electronics interface. No. It does contribute significantly to the cost of a drive, but if you grab a new ATA drive and a new SCSI drive from the same manufacturer and void your warranties by opening them up, you'll see very, very different sets of mechanics inside. The SCSI drive will have components inside that just don't exist in the ATA drive, let alone the ones that are made of higher quality materials.
"A company has to put $XX in the bank and not touch it to cover their warranty costs. If they reduce the warranty, they can use the $$ on other things, such as R&D to make more reliable drives."
Not true. A company does NOT have to accumulate & set aside cash to cover future warranty claims. A company must accrue the expected future warranty expenses, record that amount as liability, and book the increase in the total accrued warranty liability as an expense (or to income, if total warranty liability decreased).
Future warranty claims have no effect on curent cash, just on current income. Big difference.
For WD, IBM, Seagate and the like, this is an easy wasy to increase their current net income & EPS. It will have an effect on future cash to the extent that the company has to use otherwise-revenue producing drives to service warranty claims. It has no effect on current cash. It DOES have an effect on current and future net income.
Um....welll...Where should I be storing data...? I'm a small shop, and I don't have some kind of mini computer sitting in the back.
Why on a fileserver in the back, one that does have at least IDE software RAID1 and a CDRW or tape backup that you take offsite.
That way you can use dirt cheap POS systems and workstations (well to the extent that work can still be done on them) and don't worry about flaky hardware. At least that's how I've set it up at a number of small (
Yeah, but Windows works after it's installed. When I installed Redhat I had to compile a custom kernel before it would even think about seeing half my hardware/disk partitions, change about 9 configuration files (including having to tell XWindows that resolutions my monitor can run at and what GFX card I have - even though they're both PnP) recompile Netscape and ask about 4 people for help. *OR* I could install Windows and everything will be working in about 20 minutes.
You can look at this 2 ways...
1: I'm a fucking idiot and I need idiotproofness.
2. I have work to do and/or a life - so I don't want to spend ages getting an operating system working.
That's just me - some people might find a broken (new) installation a really fun exciting challenge that they can fix. I don't.
I haven't tried a large Linux distribution recently, so I'll give it another go before complaining again, but I'm pretty sure I still won't want to use it as my main OS (yet).
Ever noticed how only Linux users complain about Windows' instability? I've never seen a BSOD in my life (apart from machines with fscking ATI cards/drivers on them) - and I've been using Windows nearly every day since Windows 95 RC1. Win2K seems pretty stable out of the box - and my XP machine hasn't crashed yet *at all* (4 months of daily usage).
I keep hearing how the cheap Maxtors that I buy from CompUSA, are unreliable crap. I believed the rumors, so I expected 'em to fail, so I bought several and RAID5ed 'em. With several, I should really have some horror stories to tell, right? Alas, no, the failure just doesn't happen. Maybe because they're in a well-ventilated aluminum case that cost me something like $250? Hm...
The last disk I had fail was a Micropolis SCSI (which was a warranty replacement for another failed Micropolis SCSI). But it was in an poorly-ventilated Amiga 3000 case. Hmmm... (I love my A3000, but when it comes to moving air across disks, I gotta admit she's not so great.) Then I replaced it (no warranty, Micropolis went outta biz) with a Quantum SCSI, but I also moved it to an external enclosure that has fans. And it didn't fail either.
Cool your drives, folks. Buy a CoolerMaster case for that next RAIDed home server. They don't just look great, they work great too.
As copyright owner of this comment, I authorize everyone to defeat any technological measure which limits access to it.