Huge list of failure rates on PC components (French, but I translated nearly everything)

[Fred Castellum]

OCZ had borrowed money they had to pay off at the same time too.... It really is a shame heir Vertex 4 line was highly competitive with Samsungs SSDs.

[KamiKatze]

No - we sell about 50/50 WD's vs Seagate at work yet we still get more WD's coming back to us. I've also seen a school that had about a 12% defect rate on non-user error issues with WD 2.5" 320GB Blues.

 

That's interesting. A while back someone posted a graph that showed the reliability of WD, Hitachi and Seagate. All 3 started off very good but Seagate's curve went down after 2 years of use.

 

I personally only had WD. My external 2.5" HDD died within 2 years, which is a cheek, let's see how long my WD red NAS survives.

[LogicalDrm]

Suck on that you Corsair fanboys!

 

Seriously though. Those numbers are so much depending on sales. Here's one list from finnish eretailer (released 26.11.2013) http://www.verkkokauppa.com/fi/lehdisto/2013-11-26/ala-osta-mitaan-paiva-29-11-2013-ala-ainakaan-osta-naita

Top of that list on repaired products are:

1. Salora 24” Full HD 100Hz TV /w DVD player (33%)

2. Blackstorm Gamma G400 prebuilt gaming PC (30%)
3. CEL POWER8 Lithium toolset (25%)
4. Asus PadFone smartphone & tablet (23%)
5. Audio Pro Addon Six pair of active speakers (22%)

 

Top for returns:

1. Skyartec Wasp Nano CP 3D nano chopper (31%)
2. FXD 5 channel chopper (22%)
3. ProCaster PDVD-007 portable DVD player (20%)
4. TeleWell TW-LTE/4G antenna (20%)
5. Green Cyclone STEAM 2000W steam washer (20%)

 

The shop sell also PC parts but since their sale rates are lower than these other items, they don't hit TOP10.

[mikeangeles 75]

Well damn, I'm using a Sapphire 7870 right now. Don't die on me you video card!

[OddlyShapedTree]

I had to RMA a 7870 before

very good post.

[brownninja97]

and people ask me why i love gigabyte so much

[NoVaGuy]

Sapphires' RMAs are high. :\

No problems with my 7870 so far though.

 

And OCZ was worse than I thought. I sense some incompetence with their management.

[Henry]

Seasonic??? Last place???

[nokturrduk84]

I only heard good about FSP Aurum PSUs. So, there's also a prove now, but sadly people here only know & buy in 90% cases Corsair.

[Strider]

It also pays to consider that some of those numbers are a result of the fact they simply sell more than some of the other brands. If you produce more, sell more, you're chance of RMA/failures also goes up, and it does not always even out.

 

Still, a very interesting read none the less. Personally, I will stick with my own persoanl experiance based on the fact I have been a professional system builder for about 20 years, and trusted online reviews when researching the hardware I intend on buying. Between the two I figure it's my best bet for making the best choice I can at the time.

 

=]

[LAwLz]

Behardware have been posting numbers from the same etailer for a few years in a row now. They seem to be pretty accurate judging by other sources I have looked at, as well as general reputation.

Just keep this in mind though: these are the RMA numbers to the store. That means that companies like Corsair "cheats" because they want customers to send defects products to them instead, thus avoiding these stats. Some of the "defect" products are probably caused by user errors as well.

 

Thanks a lot for the translation OP.

 

 

It also pays to consider that some of those numbers are a result of the fact they simply sell more than some of the other brands. If you produce more, sell more, you're chance of RMA/failures also goes up, and it does not always even out.

God dammit why do people keep saying this? That's NOT how percentage works... Yes if you produce a higher quantity of a product then you will get more defects, but you will also get more functional products which will stabilize the % of RMA'd products.

1000 sold and 10 defects will have the same percentage as 100 sold and 1 defect.

 

 

I only heard good about FSP Aurum PSUs. So, there's also a prove now, but sadly people here only know & buy in 90% cases Corsair.

There is more to PSUs that just not dying though. I agree that Corsair are very over rated (they just rebrand mostly Seasonic PSUs for crying out loud) but they are still very good.

 

 

My issue with all of this data is that it isn't clear what the percentages are weighted against. Are they weighted against eachother or number of units sold?

 

For instance, how many people/OEMs are using Corsair PSUs compared to Cooler Master units? More units = higher % failiure rate.

I think it's very obvious. It's the number of products sold vs the number of products returned. If they sell 100 Gigabyte motherboards and 10 are RMA'd, then it will have a 10% return rate. It's as simple as that. That's why the whole "more units sold = higher failure rate" doesn't make any sense whatsoever in this context.

I don't really understand how you even think that this would be weighted against each other. We would have completely different numbers if that was the case. For example we would have a ~30% failure rate on ASRock motherboards, or something along those lines.

[Slinkey]

Yeah I also think it is mostly based on sales, if you we could see how many sales each company had along with the failure rates, it would be more "accurate".

[Buttsubushi]

It also pays to consider that some of those numbers are a result of the fact they simply sell more than some of the other brands. If you produce more, sell more, you're chance of RMA/failures also goes up, and it does not always even out.

God dammit why do people keep saying this? That's NOT how percentage works... Yes if you produce a higher quantity of a product then you will get more defects, but you will also get more functional products which will stabilize the % of RMA'd products.

1000 sold and 10 defects will have the same percentage as 100 sold and 1 defect.

Not necessarily. If you have a product that sells well, and you have to manufacture more to keep up with the demand, skimping on quality parts or error checking may be required to get more of the product on the shelf, which will in turn increase failure rates for that product.

[JAW96]

Yeah I also think it is mostly based on sales, if you we could see how many sales each company had along with the failure rates, it would be more "accurate".

not how percentage works, if less are sold but a lot fail for example 100 are sold 20 fail that is 20% but if let say 1000 sold and 200 fail look still 20% see math simple math at that

[Slinkey]

Quite true my mind derped

[JAW96]

Quite true my mind derped

derpage happens to the best of us

[mgsstar]

Surprised that Seasonic had a pretty high failure rate of around 5% and great that I got the 6+8 pin version of the 7950 since the MSI R7950 Twin Frozr 3GD5/OC BE had a failure rate of 4,06%.

[hawaiims]

Well damn, I'm using a Sapphire 7870 right now. Don't die on me you video card!

These rates of failures are all based on returns which were filed between 0 and 12 months since purchase. If you've had the card for longer, it shouldn't be a problem. Besides, as the article stated, the failure rate is in a U shape, with most failures however being right after purchase. If your card has lasted for a couple months already, you shouldn't have issues.

 

 

Not necessarily. If you have a product that sells well, and you have to manufacture more to keep up with the demand, skimping on quality parts or error checking may be required to get more of the product on the shelf, which will in turn increase failure rates for that product.

I'm pretty sure I can say with confidence that every single part listed in the whole article was manufactured in the tens of thousands at the very least. After all, keep in mind that these pc components are sold all over the world and that this is just data from a French retailer (and France doesn't have a very large custom pc building market like the US or Japan). 

[mikeangeles 75]

These rates of failures are all based on returns which were filed between 0 and 12 months since purchase. If you've had the card for longer, it shouldn't be a problem. Besides, as the article stated, the failure rate is in a U shape, with most failures however being right after purchase. If your card has lasted for a couple months already, you shouldn't have issues.

 

 

I'm pretty sure I can say with confidence that every single part listed in the whole article was manufactured in the tens of thousands at the very least. After all, keep in mind that these pc components are sold all over the world and that this is just data from a French retailer (and France doesn't have a very large custom pc building market like the US or Japan). 

Yeah, thanks for clearing that out, the card's still running like a champ btw.

[NoVaGuy]

Yeah, thanks for clearing that out, the card's still running like a champ btw.

 

Mine too. Although it doesn't OC super well.

[hawaiims]

Mine too. Although it doesn't OC super well.

wow really? That's quite surprising to me because crypto miners always rave on sapphire cards and are prepared to pay the highest premium for sapphire cards in particular because they can usually squeeze out the highest hashrates. The AMD GPU brand most hated by cryptominers is also XFX, which for some reason always gets abnormally low hashrates on their cards. 

 

Then again perhaps you can't really compare the overclocking prowess of cards for cryptomining and gaming because artifacting and such things are not of any importance for mining. The only thing that matters is for that is to get the highest crash rate without the card crashing completely or throttling due to temps that are too high. Mining for coins is basically the same thing as running furmark 24/7 because it pushes your gpu to the absolute max, while with gaming that is not the case. 

[LAwLz]

Not necessarily. If you have a product that sells well, and you have to manufacture more to keep up with the demand, skimping on quality parts or error checking may be required to get more of the product on the shelf, which will in turn increase failure rates for that product.

Ehh... Yes? I don't really see how that is a counter argument to what I just said. Selling more does not increase the risk of failure. Cutting down on the quality will increase risk of failure.

Again, the number of units sold does not have any effect on the numbers on this list. Low quality parts will however have an effect on this list. If someone has a high failure rate on this list then it's not because they sell more, it's because they have worse reliability.

[Leonard]

Most RMAs are the result of stupidity.

Your source please.

[aEi1aesainee2]

Just so that you know, it isn't anonymous ^^

 

 

Hardware.fr belongs to LDLC.com
(which is the leading internet electronics store in France)

[jagjordi]

This is not a good spot for CORSAIR. 

I'm glad i bought a Samsung SSD.

Excellent post. Saved in Favorite favorite list.