Atom C2000 AVR54 issue compensation for customers out of warranty

[PKoc]

Hello.

My motherboard with Atom C2550 processor died this January after three years since purchase and one month after warranty expiration. Considering the chip is affected with a build-in defect of clock signal malfunction (as mentioned in AVR54 note in Specification update http://www.intel.com/content/dam/www/public/us/en/documents/specification-updates/atom-c2000-family-spec-update.pdf http://www.intel.com/content/dam/www/public/us/en/documents/specification-updates/atom-c2000-family-spec-update.pdf) that was likely the cause of the death of the board how will Intel compensate customers affected who can not RMA their boards by manufacturers?

It is a very bad publicity Intel wasn't able to take care of quality control on its server component and I hope Intel will at least take care of those affected by paying all the costs associated with this issue.

Best regards.

[n_scott_pearson]

This is something you will need to discuss with Intel Customer Support directly; not here in a user-to-user support community. Here is information regarding how to contact them (by geography):

http://www.intel.com/content/www/us/en/support/contact-support.html# @11 Intel Customer Support Contact Information for US and Canada

http://www.intel.com/content/www/us/en/support/contact-support/emea-contact.html Intel Customer Support Contact Information for Europe, Middle East and Africa

http://www.intel.com/content/www/us/en/support/contact-support/apac-contact.html Intel Customer Support Contact Information for Asia-Pacific

http://www.intel.la/content/www/xl/es/support/contact-support/lar-contact.html Intel Customer Support Contact Information for Latin America

...S

[PKoc]

I think Intel should publicly say how he's going to fix the issue and compensate the damages caused by lack of quality control. There could be millions of units affected as Atom C2000 is on the market since fall 2013 and many of the units are now out of warranty.

It is surprising this happened to server product where the most thorough quality control should take place.

[idata]

Hello:

Thank you very much to N.Scott.Pearson for the information posted above. To petr.koc: Thank you very much for providing that link. First of all, we just wanted to apologize for any type of inconvenience you might have experienced with the processor. In regard to the information on that link about the build-in defect of clock signal malfunction, I just wanted to let you know that the issue is related to the processor itself, and as it says there the system may experience inability to Boot or may cease operation, but it will not damage the board, the problem happens just on the processor, it will not damage the rest of the components. So, in that scenario in order to fix the issue, what we do is to replace the processor if it still is under the 3 years of warranty period. Any further questions, please let me know. Regards Alberto

[idata]

Hello petr.koc:

I just wanted to check if the information posted previously was useful for you and if you need further assistance on this matter? Any questions, please let me know. Regards Alberto

[PKoc]

Hello.

Thank you for the answer, As said my motherboard is out of warranty as I bought it in 2013 already, therefore the 3 year warranty period expired in 2016. The motherboard died in January 2017.

Sorry to say but next time I will seriously consider consumer silicone based products (Xeons / Pentiums for LGA1151) as those are tested by millions of customers and the error will likely be discovered sooner or I will consider AMD (I know they had some issues too but none as serious as complete silicone failure on server class product).

In the age of Pentium IIIs and Pentium IVs Intel created newer revisions of the same processors that fixed major issues. What has happened with this policy? Recently there were so many issues with Intel products (P67 chipset SATA dying, Haswell USB chipset bug, TSX not working on Haswell, Skylake freezing on Prime95). Does Intel even have some quality control if such a bug as processor freezing in Prime95 is not discovered before product launch? Prime95 is a standard reliability test. Strange.

[idata]

Hello petr.koc:

Thank you very much for providing that information. Just to let you know Intel, really appreciates and respects the comments, suggestions and feedbacks provided from any of our peers in our community. So, we are sorry to hear that you had a bad experience with the Intel processor, and I am going to send your feedback to the proper department so they can be aware of your suggestions and comments. Once again, we apologize for any inconvenience and is by any chance you need our assistance in the future, do not hesitate in contact us again. Regards Alberto

[n_scott_pearson]

This is the silliest discussion I have heard in a long while. There is absolutely zero proof that this errata had anything to do with this board failing. There are hundreds (if not thousands) of possible causes for this type of failure that have nothing to do with this processor errata.

[idata]

Hello:

Thank you very much to N.Scott.Pearson for the comments posted previously, I hope that information was useful for all the peers viewing this thread. Regards Alberto

[PKoc]

Mr. Pearson, what could it be when the motherboard worked and then suddenly died? There are lot of issues mentioned on the web about motherboards with Atom C2000 dying suddenly such as:

Asrock - https://www.reddit.com/r/freenas/comments/4x1kh1/asrock_c2550d4i_sudden_death/ https://www.reddit.com/r/freenas/comments/4x1kh1/asrock_c2550d4i_sudden_death/

Synology - https://forum.synology.com/enu/viewtopic.php?f=7&t=119727 DS1815+ died (no lights) - how to move disks to another DS1815+ without having configuration backup - Synology Forum

Cisco is doing recall of all their products containing this chip - http://www.cisco.com/c/en/us/support/web/clock-signal.html# ~faqs Clock Signal Component Issue

According to unnamed sources the issue has been there for some months as failure rate was much bigger than usual - https://www.theregister.co.uk/2017/02/07/intel_atom_failures_go_back_18_months/ Intel Atom chips have been dying for at least 18 months – only now is truth coming to light • The Register

What we see here is Intel designed a (server) chip that is definitely going to die due to hardware design issue. It is a question of when rather than if and the life expectancy is much shorter than what an average customer would expect. The way I see it Intel should take full responsibility and bear all the costs related to this issue (including software licencing costs) to save their image. It will be lot of money but still much less than what a bad reputation will cause then in long term. And as mentioned, there are customers who already are out of warranty.

[n_scott_pearson]

Again, there are hundreds (if not thousands) of possible reasons for a system to die. Just because an issue like this exists does not mean that this is the cause of your system's failure. Could it be? Certainly. Is it likely? That can be debated. Regardless, in your case, it really doesn't matter. You received a warranty when you purchased this processor. Your processor worked properly over the entire duration of the warranty period. Once this period has expired, you are not entitled to any compensation if the processor then fails - regardless of the cause.

...S

[AlHill]

petr.koc, how far beyond warranty do you believe a product, any product, should be supported?

Just curious.

Doc

[PKoc]

2 Al Hill: I think the manufacturer should take full responsibility for design defects for the whole expected lifetime (say on computers for 10 years).

One thing is the normal wear. If you use it heavily, it should last the warranty period. That is as a customer you have a guarantee that if you stress it, it will not wear out before warranty period (or warranty specifications in case of products such as SSDs).

The other thing is the design defect. I would definitelly not buy this Atom C2000 if I knew it has a design defect that would cause it to die so soon. I believe this is standard as home appliances are often recalled after many years in service if design defects are found, the same with cars (see Volkswagen Dieselgate for example, Takata airbags), exploding laptop batteries etc.

[AlHill]

You are mixing Apples and Oranges. Safety recalls are a completely different subject. 10 years on computers? Let's be real here.

Dieselgate? Totally irrelevant.

You made it beyond the warranty period. Time to move on.

For me, when something goes bad, even when in the warranty period (with few exceptions), I either upgrade or replace the item, and learn from the experience rather than have my hand out.

I am trying to find a better term than "silliest discussion", but cannot.

Doc (no longer following this thread)

[n_scott_pearson]

Many of the components used on motherboards have an MTBF of 7 years. To think that a computer will last longer than that is thus completely unreasonable (bordering on idiotic). Am I saying that the expected lifetime for a processor should also be 7 years? Absolutely not. The expected lifetime cannot be anything other that the warranted lifetime. Warranties cost money. If you want longer warranties, you must pay more - often significantly more. If you don't like the warranty, you don't have to purchase the product. Of course, since AMD's warranties are very similar to Intel's, this means you won't be purchasing a product (period).

I too am done with this thread.

...S

[PKoc]

Sure, it could be something else. But since there are massive number of failures of this CPU it is the most probable cause. Usually when electronics is about to die it starts behaving strange - here it stopped working suddenly what is the exact description of this issue.

The life expectancy DOES matter a lot. Nobody buys something expecting it will last just the warranty period. In Europe standard warranty is 2 years yet people do not buy houses on 30 years mortgage expecting them to fail after 2 years. They expect them to last 100 years as that is a standard. Similar with computers - usually one expects the computer to last about 5 to 7 years if properly maintained and not stressed heavily. If Intel ever mentioned their CPUs are designed to last warranty period only it would be the end of their business. IT professionals trusted Intel because, while somewhat expensive, it was a guarantee of quality. Yet recently this image is eroding because of this issue and the other chipsets and CPUs issues mentioned.

Anyway I will wait to see AMD Ryzen performance and then will consider some low power LGA1151 board. No chance I will go again with Atom C2000 or try untested Atom C3000.

[PKoc]

According to both of you when I buy a house I should extend warranty from 2 years standard to 100 years because if the house collapses after 2 years it will be my fault as I should have expected it to collapse since the warranty was only 2 years. That is what I would call idiotic.

[n_scott_pearson]

No, that's what I call a silly apples to oranges argument.

[ATech4]

Has there been an platform level workaround identified for this problem? If an OEM or major distributor with certified refurbishing relationships wishes to participate in the RMA process and refurbish these processors SoC in OEM products, how do we gain information to determine if it's feasible for a high-volume service line can fix these cost-effectively; on behalf of the OEM and/or Intel.

[Alberto_Sykes]

Hello atalba:

In regard to your question, the Intel® Atom Processor C2000 units are sold to the OEM (Original Equipment Manufacturer) and to Intel approved Distributors. So in order to get the information you are looking for the best thing to do will be to get in contact with either one of them to get further assistance about this matter.

Regards

Alberto

[CPerc1]

@PKoc​ 

I came here looking for a copy of AVR54, for Jan 2017. My Synology NAS has shown signs of imminent failure (improper shutdowns). I submitted a support ticket and received notification of RMA process being started within 3 hours. No messing about, system hasn't failed yet and is 3 years 1 month since purchased - so kudos to synology support, i'm in awe of how hassle free it was. Sorry to hear OP how it not the same for everyone with defective CPU (continued...)

@n.scott.pearson​ 

"This is the silliest discussion I have heard in a long while. There is absolutely zero proof that this errata had anything to do with this board failing. There are hundreds (if not thousands) of possible causes for this type of failure that have nothing to do with this processor errata."

Shame on you. A retired software engineer of 35 years experience 21 OF THOSE WITH INTEL. Dismissing someones frustration at their failed hardware not least because you are a former intel SOFTWARE engineer but these cpu's have shown failure rates above normal and the AVR54 erratum even suggests the chip could fail without warning if it was manufactured pre feb 2017 and even more so if in continuous use. I don't dispute there are MANY (hundred / thousands? only in a literal sense if you count every component on the motherboard, so really?) other reasons a motherboard could fail, but in this instance I wouldn't wan't to hedge my bets against it not being the CPU, would you?

This whole situation isn't the norm as far as reliability goes and so was the case 2 years ago too. Who knows how bad it might or might not get? What about after 4 years or 5 years, warranties are warranties but if people knew that they had a say 25% chance of only getting 3 years and 50% chance of getting 4 years out of any piece of IT hardware would they buy it in the first place? Of course not, so OP's hardware dies after 3 years, comes here to vent a bit just to be belittled by current and former employees of intel.

If i'm being unclear, the next post after you calling OP silly was from intel employee:

"Thank you very much to N.Scott.Pearson for the comments posted previously, I hope that information was useful for all the peers viewing this thread." THANKING you for it.

Most consumers are customers not experts, everything the OP uses in his posts are pure google results, so assuming savvy but not technical person. Search Intel c2000 cpu and guaranteed it will all be about whats wrong with them or links to intel specs etc. I'm already 100% AMD and if this berating of peoples problems is the norm on intel forums, it's going to stay that way.