Failure rate statistics

[ddrueding]

Yeah, I've had mixed results. Particularly when trying to run high-end RAID cards on low-end desktop boards. It sometimes assumed I had an external video and disabled the on-board. That was the last time I used an ASUS board

 

[Mercutio]

New Antec NSK-series cases are made with thinner steel than older revisions, and I've had a couple shipped with DOA power supplies. I'm not sure if that's lower quality PSUs or a problem that's come from shipping practices. The only memorable in-warranty power supply failure I've had in the last year or so was a 480W Truepower unit from a machine one of my students built.

I pull FSP Group branded power supplies out of machines fairly regularly. I don't track who made dead stuff I didn't put together myself, but if I had to guess I'd say I see one every six weeks or so. I don't think there's another single PSU brand I see that often in the random collection of stuff that people bring me to look at.

 

[time]

People have probably already seen this, and it's out of date (2009 stats), but interesting anyway:

Study: A Look At Hard Drive Reliability In Russia

 

[Mercutio]

I do recall seeing that before. I believe it was linked from Digg.com. It's interesting that those guys claim Hitachi as the most reliable and the hardware.fr guys claim it's the least. Regional variations? Or just a change with more recent models?

 

[CougTek]

It wouldn't be the only case of regional reliability variation. Samsung, for instance, seems to be a lot more reliable in the Soutern Hemisphere than it is around here. As WD is if you compare my experience with them and yours.

 

[Mercutio]

My personal belief is that any who says they've had a positive experience with Western Digital products as storage media is either lying or needs to up their chlorpromazine dosage.

 

[ddrueding]

http://www.ted.com/talks/lang/eng/kathryn_schulz_on_being_wrong.html

 

[Mercutio]

Take your antipsychotics, doofus.

 

[Stereodude]

It wouldn't be the only case of regional reliability variation. Samsung, for instance, seems to be a lot more reliable in the Soutern Hemisphere than it is around here. As WD is if you compare my experience with them and yours.

Coriolis effect? :scratch:

 

[ddrueding]

Take your antipsychotics, doofus.

?

 

[Mercutio]

?

Only a sociopath would suggest that WD is acceptable for anything other than construction material.

 

[Mercutio]

Sub-par construction material.

 

[Mercutio]

I'm talking first two Little Pigs bad, here.

 

[ddrueding]

The point of the talk was the observation that the default opinion one has on someone who disagrees with them follows a fairly sad pattern:

1. Ignorance: They obviously just don't have the information I do, allow me to enlighten them.
2. Stupidity: Now they have all the information, and are just too stupid to connect the dots as I have.
3. Malice: This person knows the truth and is intentionally denying it for their own evil purposes.

It had nothing to do with what is right and wrong, but instead focused on the unproductiveness and general silliness of taking this even to step 2 in most cases.

 

[Mercutio]

Well #3 would go along perfectly well with those who attribute positive outcomes to storing data on Western Digital products and/or using them to build some form of barrier.

 

[Tannin]

My personal belief is that any who says they've had a positive experience with Western Digital products as storage media is either lying or needs to up their chlorpromazine dosage.

Rubbish. Western Digital made the best MFM controller cards, and their Paradise video cards were extraordinarily good.

 

[Chewy509]

Sub-par construction material.

So you're going to be buying Seagate from now on?

Or alternatively ditch spinning disks altogether from now on?

 

[time]

It wouldn't be the only case of regional reliability variation. Samsung, for instance, seems to be a lot more reliable in the Soutern Hemisphere than it is around here. As WD is if you compare my experience with them and yours.

I don't know about the Southern Hemisphere, but there's definitely a reality distortion field around Ballarat.

Given the age of the data, the statistics didn't seem that unreasonable to me. I haven't used Western Digital for years because someone on Storage Forum had a few unkind things to say about them. And I did have one die without warning.

I haven't been happy with Samsung models until fairly recently, with first the F3 and then the superb F4 (I see that the European return stats back this up). I wasn't planning on using anything else ever again, but now that's all gone down the toilet.

If you look at the engineering culture of Hitachi, combined with what they inherited from IBM (and the sensitivity towards failures), I wouldn't be at all surprised if their drives made a few years ago outlasted other brands. No money in that though, so no future.

I'm afraid I still don't trust Seagate. Nothing in anything we've read here, or user feedback over the years, has allayed my concerns.

 

[Mercutio]

I've had luck with Samsung drives that varies between very good and truly excellent. In a bad year I might RMA 2% of the Samsung drives I've purchased for sale. I've likewise had good luck with Hitachi, though I buy those drives in much smaller quantity and I will admit I've had some DOA problems with the current model 1TB drives, which has been the cheapest 1TB full speed drive for at least the last year.

Seagate is incredibly spotty for me. Seagate is frequently my only option for low-capacity drives and that is the sort of Seagate drive I generally buy. I've had years where I've returned in excess of 30% of the Seagate drives I attempted to sell, with the 80GB and 250GB Barracuda drives being the leading cause of that and the 5400rpm 2TB drives looking at that number in the rear-view mirror. On the other hand, I've had no problem with 160GB or 1.5TB 7200rpm Barracuda drives.

Every now and then I wind up with a WD drive. I haven't purposefully installed one in a computer in years, and I'm not going to start now. I don't trust them. I won't use them for anything remotely important. They're used as the protective layer that surrounds my other spare hard drives.

 

[Mercutio]

I've had luck with Samsung drives that varies between very good and truly excellent. In a bad year I might RMA 2% of the Samsung drives I've purchased for sale. I've likewise had good luck with Hitachi, though I buy those drives in much smaller quantity and I will admit I've had some DOA problems with the current model 1TB drives, which has been the cheapest 1TB full speed drive for at least the last year.

Seagate is incredibly spotty for me. Seagate is frequently my only option for low-capacity drives and that is the sort of Seagate drive I generally buy. I've had years where I've returned in excess of 30% of the Seagate drives I attempted to sell, with the 80GB and 250GB Barracuda drives being the leading cause of that and the 5400rpm 2TB drives looking at that number in the rear-view mirror. On the other hand, I've had no problem with 160GB or 1.5TB 7200rpm Barracuda drives.

Every now and then I wind up with a WD drive. I haven't purposefully installed one in a computer in years, and I'm not going to start now. I don't trust them. I won't use them for anything remotely important. They're used as the protective layer that surrounds my other spare hard drives.

 

[CougTek]

Only a sociopath would suggest that WD is acceptable for anything other than construction material.

That might be ok if I was the only one to have a positive experience with WD, but Buck isn't a sociopath and he often mentioned that he uses WD in his computers.

 

[BingBangBop]

I've really had much worse luck with Seagate than WD.

 

[jtr1962]

There actually are some technical reasons to back up Merc's disdain for WD:

"Western Digital's construction makes drives particularly vulnerable to shocks and pressure. Unlike other manufacturers, WD does not secure the hard drive axle with a separate screw to the drive cover. Because of this, pressure exerted on the housing or cover can shift the axle, resulting in it changing its angle, and then damaging the platters. The axle's attachment to the cover is another reason. If the cover is moved, the engine may be blocked. Except for this vulnerability, though, WD hard drives are mechanically and electronically reliable."

(From the article time linked to a few posts back)

I've stuck with Maxtor with good results for a long time. When they merged out of existence, I went with Samsung, again with good results. Not sure what I'm going to now that Samsung is merging away. I'm hoping by the time I need another drive, mechanical hard disks will be relegated to the dust bin of history. SSD, while it can have reliability issues, is inherently more reliable than a storing data in microscopic bits on a disk spinning 90 to 250 times per second. I've honestly never felt comfortable with that concept.

 

[ddrueding]

When I first went to 3TB drives, WD was the only option. I've since phased them out, and have 12 3TB Hitachi drives on the desk in front of me. I've dealt with about 30 drives in this capacity so far without a single failure. This included 6 WD internal 3 WD external, and 18 Hitachi. I can't remember what the rest were, and they are out in the field.

 

[Mercutio]

Buck isn't a sociopath and he often mentioned that he uses WD in his computers.

Buck has a reason for his particular bias. And there's certainly no evidence that he's NOT a sociopath.

 

[time]

Some more RMA stats, this time from an online store in New Zealand. The methodology and time frame aren't stated. Within each product class, brands are given a reliability rank between 0 and 100%, where 50% is supposed to be the average across the entire store inventory.

I thought it was also interesting to broadly compare different types of product, eg hard drives, motherboards and power supplies.

External HDD
100%: Hitachi, Samsung, Verbatim
90%: Iomega
70%: LaCie, Seagate, Transcend, WD
60%: Freecom

Notebook SATA
90%: Hitachi, Samsung
80%: Seagate
60%: WD

Desktop SATA
80%: Hitachi
70%: WD
60%: Samsung
50%: Seagate

SSD
100%: Crucial
90%: Intel, Transcend
60%: Kingston
50%: Mach Xtreme
40%: Corsair
30%: A-Data, OCZ

PCIe Graphics
90%: Leadtek
80%: Gigabyte, MSI
70%: Asus
60%: Sapphire
50%: HIS
40%: XFX

Intel s775 Motherboard
60%: Asrock, Gigabyte, Intel
40%: Asus, MSI

Intel s1366
100%: DFI
50%: Asrock, Asus, MSI
40%: Gigabyte
20%: Intel

Intel s1156
100%: DFI, MSI
50%: Asrock
40%: Gigabyte
30%: Asus
20%: Intel

Intel s1155
100%: Asrock, Gigabyte, Intel, MSI
50%: Asus

AMD AM3
50%: ECS
40%: Asrock, Asus, Gigabyte, MSI

Desktop RAM
100%: Crucial, Kingston
90%: A-Data, Transcend
70%: Super Talent
50%: Corsair
40%: G.Skill
30%: OCZ

Power Supply (not included with case)
100%: Antec, Enermax, InWin, Thermaltake (!)
80%: CoolerMaster, Highpower, Seasonic, Zalman
70%: Gigabyte, Silverstone
60%: Corsair
50%: Aerocool, Huntkey, Vantec

 

[Mercutio]

Is 100% good or bad? If 100% is good, those numbers for Intel motherboards look awfully fishy.

 

[ddrueding]

Is 100% good or bad? If 100% is good, those numbers for Intel motherboards look awfully fishy.

Agreed. DFI scored remarkably well if 100% is good, but if it is the other way around, the hard drives look wrong.

 

[Chewy509]

And why are the stats rounded to the nearest 10% ?

 

[time]

They're not stats, that's why I used the term "rank". They display them as 1 to 5 stars (with half stars), along with a factor of 10 that includes a '%' - which is confusing. It's really 0-10 and the '%' is probably just a formatting issue that I should have removed - a mod could clean it up.

The reliability rating is based on the actual returns of faulty products we have experienced ourselves. On a global scale the results could be quite different as this rating is based on and limited to our own sales. The rating does not relate to an individual item, but to all items by the same manufacturer in the same category. A rating of 2.5 stars represents the normal rate of return we would expect from all our sales across all categories. Products that are very unlikely to develop faults will tend to have high ratings (eg network cables), while more complicated or fragile products (eg notebooks, hard drives) will tend to have lower reliability ratings. It is the rating relative to other manufacturers in the category that is important. A rating of 4 stars would be excellent for a hard drive, but not so good for a cable.

 

[Mercutio]

... and then, two hours after I typed this, I had to swap out an Intel DH55PJ, only seven months old.

And why the hell are LGA1156 boards still so expensive, anyway?

 

[time]

Had an extended heart-to-heart conversation with the distributor for 'OceanSound' power supplies in this part of the world. The 430 and 500W S12II/M12II units have a return rate of less than 0.5%. Various sites were cited as examples of where these units were selected to satisfy a requirement for high availability.

Things were less rosy with higher power units. Personally, I suspect that is an industry-wide issue.

I may have been misinterpreted this part, but the expensive X series (semi-fanless) are surprisingly popular and are not causing any RMA heartburn.

HTH.

 

[tazwegion]

Only a sociopath would suggest that WD is acceptable for anything other than construction material.

There was an interesting "in house" study conducted by a Russian data recovery company, although the study couldn't be extrapolated to encompass the entire HDD industry it still serves as an interesting snap shot of drive reliability, failure rate & market share at the time of the study in 2009, of those drives sent in for recovery the WD's had a lower percentage of failure rate than Seagate and the Hitachi's only failed due to external forces (user mistreatment) having no recorded mechanical/firmware issues


Food for thought...


Perhaps as has been suggested varying climatic conditions combined with altered electricity quality (irregularities) has more of a dramatic effect on the serviceability of these drives?

 

[Mercutio]

I prefer the "born wrong" theory.

 

[tazwegion]

I prefer the "born wrong" theory.

So how does your theory explain the Seagate drives with a 31% National market share equated to 56% of all drives delivered for Data reclamation, this is more than twice the rate of the WD drives with a similar 30% market share?


One potential explanation could be that the people who were turning in these drives were business orientated thus making data recovery necessary and that this percentage of the population purchased Seagate solely on it's reputation, whereas WD owners generally weren't business orientated in nature and simply absorbed the data loss...

 

[Tannin]

Essentially, yeah. Collecting failute statistics from drives sent for data recovery is absolutely and completely pointless. Oh, it's quite interesting, but it doesn't tell you anything about between-brands reliability.

All it really tells you is something about the kind of clueless fool who is so stupid that they don't have meaningful backups of data so important that they are prepared to pay thousands of dollars to (maybe) get it back.

And, before anybody asks, I don't think you can even find valid generalisations about the brand-choices of the truly clueless from this data - in most cases they won't even have known what brand of HDD was in their machine, and we know nothing about what proportions of particular brands are sold to the clueless as opposed to the rest of the market.

 

[Mercutio]

So how does your theory explain the Seagate drives with a 31% National market share equated to 56% of all drives delivered for Data reclamation, this is more than twice the rate of the WD drives with a similar 30% market share?

My theory holds that Seagate drives are also born wrong. Which is why I've suddenly become very interested in MRAOPCR technology.

 

[LunarMist]

My theory holds that Seagate drives are also born wrong. Which is why I've suddenly become very interested in MRAOPCR technology.

What is that? Google only brings up SF for MRAOPCR.

 

[Tannin]

Well there you go, mate. Google is never wrong. Obviously, Mercutio is the World Authority!

 

[Mercutio]

What is that? Google only brings up SF for MRAOPCR.

Google knows full well what it is then!