4 Sticks Dead At Once?

[Daves255]

I realize this would be best for the Dell forum (I posted it there too), but I'm pressed for time and need a response soon.

Dell has given me hell. I'll make the backstory pretty short to explain why I need your help. A little over a year ago, I ordered a XPS 410 with an 8800 GTX graphics card. The power supply proved to be inadequate for the card (Nvidia's own specifications exceeded the power supply's specified rating), and the computer would constantly lock up in games. Dell graciously replaced the 410 with a 720 with upgraded parts.

The system has been very unstable since getting it. During games, the graphics card would lock up during games, sometimes the screen would go black, other times it would give graphics card memory errors or driver blue screens. Basically, it would blue screen/lock up 3+ times a day. This, for me, wasn't a big problem. However, my brother's reference system, a similar system with an 8800 GTX and the same OS, amount of memory, etc. didn't experience nearly the volume of lockups that I had on this system (even playing the same game).

The other day, I attempted to turn on the computer, but I got a beep code error. The beep code is 1-3-2, and the front LEDs indicate a memory problem. It is saying that the memory is corrupted. I understand that memory becomes corrupted, and that replacing a memory module isn't a big deal; it happens. But what happened next is truly troubling and puzzling.

After calling Dell, they had me remove all memory modules but one and test each one. 3 out of the four modules gave the same beep code and LED lights on the front (we didn't bother testing the fourth). We also tried multiple combinations to the same effects. Having no memory gave a memory not found beep code, which led them to believe that the problem was the memory and not the motherboard.

Now that it's becoming clear that something is going to have to be replaced, Dell is refusing to honor the warranty of the second computer that was shipped, saying that the warranty starts at the purchase date rather than the day we received the second computer that actually worked. I find it highly unlikely that their diagnoses that all four memory modules failed at once is correct. What could this indicate?

I've considered taking out the graphics card to see if it was a video memory error rather than a RAM error, but it's heavily bolted down.

What would be the best course of action, and what steps would you recommend to fix and prevent this from happening again?


Thanks.

 

[mustacheman8]

Try the RAM in another computer system. If it works there, the RAM is not dead, and most likely your motherboard is fried If it doesn't work...my bet would be that the voltage was screwed up somehow...dont ask me about that im no expert....but a power issue could have fried all of your RAM.

 

[techinsider]

I used to work for Dell's Home and Small Biz tech support division, as both a tech and later a manager(left in late 04'), and I'm sad to report that warranties from every OEM (HP, Dell...even Falcon and Alienware) are tied to the purchase dates, even on those rare occasions when they choose to replace the entire unit. The warranty can be enforced however, if you 'reported' the memory issue on this system before the warranty expired. In this last case, a manager will have to issue you a short-term extension, before you can receive warranty support.

When phone techs test memory, its usually easiest to have the customer do exactly what your tech asked you to do (heck, I think I wrote that script for them!). To test the physical sticks, we have you remove all but one, and try the stick in each seperate slot. If you get errors from each stick in each slot, then they ask you to set that stick aside, and repeat the (grueling) process with each stick, until we can reach a logical conclusion.

If all 4 sticks fail in all four slots, that's a bad mobo. If all the sticks fail in the same individual slot, that's a bad mobo. If one stick fails and others work, that's a bad stick. And the ultimate proof of a bad motherboard slot is take the memory to a different, compatible computer and test it there.

Memtest86+ (downloadable for free at http://www.memtest.org ) can also provide definiative proof about what's going on w/ your memory, slots, and/or memory controller.

As for your graphics card memory, I'm thinking no. When GDDR3 goes bad, you'll first see lots of tearing and jagged edges, and the issue will gradually get worse and worse and worse...it's very obvious when you see it.

As for what killed your slots (or sticks), its usually related to a simple power surge, or overclocking, or over-voltages beyond 1.8-2.0 volts. It's also possible that your bios settings changed to something-other-than-stable. Setting them back to the defaults should eliminate that particular cause, if applicable.

-Lee