I was wondering if your brand new sticks of RAM DD3 could go faulty after a hard shut down of a PC? No power surge and did not turn it back on immediately

UPDATE: Computer turns on but does not POST. There are indicator lights on the MB that state the RAM is causing the issue. When I remove the ram it does give me beep codes but other wise it does not. It was just working yesterday..


I have 4 sticks of RAM, how can they ALL go bad? I have tried them in every combonation possible (single, double, triple,etc) I have tried every slot with each stick and same result.

  • By "hard shutdown" do you mean you held in the power off button or the mains was lost? – Tog Dec 21 '11 at 20:45
  • I have never seen any of my ram go bad, even after decades of use. we use UPS (uninteruptable power) proper cooling, and do not raise the voltage of the ram too high, even if overclocking. The first thing that comes up in Wizlogs link "There are other possibilities" is true more often than not. With DDR3 and direct cpu connection , even the connection points of the CPU socket could effect it. Togs comment about how the the power went down (or up and down) a good surge can damage anything. There are many reports of a motherboard (now) tagging the ram , without the ram being bad at all – Psycogeek Dec 21 '11 at 21:44
  • Decades of use? Time to upgrade :) These computers now a days will amaze you. And yes, RAM can go bad. That's why every motherboard known to man will have a RAM test function built into it. – Safado Dec 21 '11 at 23:30
  • Everything in the computer was manufactured no more than 2 years ago – nGX Dec 21 '11 at 23:38
  • 1
    RAM can go bad. Since you tried all 4 sticks, it sounds like it might be a different issue. If you have another computer available, I'd try to run a memtest on the RAM in a different computer. If it comes out clean, there's another problem. – nhinkle Dec 22 '11 at 0:48

There are a lot of reasons RAM can go bad. Too many to name.

You should run memtest86+ to verify that the RAM is bad. You can also find a copy on the Ubuntu install CD and other Linux boot CDs.

  • 17
    Cosmic rays...definitely cosmic rays... – EBGreen Dec 21 '11 at 20:47
  • 2
    I am so glad we are not experiencing coronal mass ejections right now. Solar flares directed at Earth make me cry. – Jeff Strunk Dec 21 '11 at 21:04
  • Damn random bit-flipping! – Garrett Dec 21 '11 at 21:12
  • 1
    I have 4 sticks, I have tried them individually as well but same result. I do not have any other sticks of ram – nGX Dec 21 '11 at 22:53
  • 1
    Sounds like the mother board is at fault then. Unless it requires matched pairs, in which case try two at a time. – Jeff Strunk Dec 21 '11 at 23:03

As with any electronics, RAM can suddenly go bad.

Because your ram is relatively new, I'd try returning or exchanging it for a replacement.

  • 1
    I once told a mentor at work... "But... it was just working. It couldn't JUST have gone bad!" He ever so casually replied, "When things do go bad - they frequently worked and one point then not at another... having literally "just" gone bad." There wasn't a lot to say back to that :) – OG Chuck Low Dec 21 '11 at 23:56
  • 1
    @OGChuckLow you mind up-voting my answer then (as it meets all the requirements set out in the faqs – wizlog Dec 22 '11 at 0:37
  • Sure it can go bad, but that doesn't mean it did go bad. Don't you think it would make more sense to do a memcheck before going to all the effort of returning it? – nhinkle Dec 22 '11 at 0:46
  • @nhinkle To quote from the question "I was wondering if your brand new sticks of RAM DD3 could go faulty...". The answer to that is that it could happen. Now that I think about it, I'm not even sure if my recommendation belongs, as he doesn't ask for it, or for any solution... – wizlog Dec 22 '11 at 0:59

It's unusual for any chips inside a PC to fail without cause. The likely failure modes are damage (electrical or mechanical) while handling, overheating, damage (cracks) from thermal cycling, electrical surge, or corrosion on socket connections. Only very rarely does a chip "just fail", without some external cause.

Memory chips are particularly sensitive to static damage and are apt to be electrically damaged if not carefully handled during installation, especially in winter when static is a bigger problem. (Though newer technologies are an order of magnitude less sensitive to this than the stuff of 20 years back.)

Another common failure from the past for memory chips is corroded socket connections. While rarer than in the past, it's no doubt still possible for corrosion to develop at the connection, causing a failure. Traditionally this was dealt with by applying grease to the contacts, to exclude the air that causes corrosion.

  • 1
    I'll add that there are two significant forms of "grown defects" -- "Out of the blue" failures for no apparent cause: First is "whiskers" -- metal fingers that crystallize and grow between wires in the IC. This is especially a problem with new RoHS standards, as a dash of lead in the mix prevents this. The second is a much older and more debated cause -- the "migration" of wiring at corners. Similar to a river changing course, the wiring will supposedly move slowly over time as current moves individual metal atoms to the outside of corners. But both of these are rare in standard parts. – Daniel R Hicks Dec 28 '11 at 19:54

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.