@RAMGUY: does Corsair 'advanced RMA' ?

Hello Ramguy,

my case number is #2597090

does corsair offer 'advanced rma' ? (send customer replacement parts with expectation of return parts recieved within X weeks, or credit card is charged for full cost of shipped parts).

i recently purchase (2x) 2 x 2gb kits of corsair RAM, with matching lot #'s.

One of the sticks died, so i was advised to RMA one 'pair.' There was no way to identify which of the 3 good sticks came with the dead one, so i chose one at random.

I sent the RMA away, recieved a replacement pair, with a completely different lot #.

I tested both of the replacement sticks for 1hr each with memtest86+ (2 complete passes) and they both had zero errors.

I installed the remaining 3 sticks of ram (1 had just finished testing).

Machine booted fine, ran windows updates fine, etc etc. Wife went to load a game, and it BSOD'd and wouldnt boot.

Some basic troubleshooting led me to running memtest86+ with individual RAM modules. 'lo and behold, one of the two that i recieved as replacement from corsair is now dead, throwing thousands of errors within seconds of running memtest (only the one is dead, the other of this pair is fine).

Would you advise RMA? is this just coincidence? (i mean, i know i've got a dead stick of RAM, that's clear). ... Wondering if something more sinisted might be at play? Faulty motherboard (intel dg33tl) or PSU.

and last but not least, can using the same part number but different lot # sticks of ram kill one or more sticks? this seems highly unlikely to me... figure i should at least ask though.

the motherboard is set to run ram @ SPD, and there is no ram voltage control afaik. No overclocking involved (no option to with this mobo). No other issues with the system either (comonents dying).

Thanks for bearing with me on this. If you need more info, let me know and i'll post it here.

Cheers,

Andrew

does corsair offer 'advanced rma' ? (send customer replacement parts with expectation of return parts recieved within X weeks, or credit card is charged for full cost of shipped parts).

Yes: http://www.corsair.com/support/warranty/

Consumers may request advance replacement but must provide a valid credit card to secure the shipment

mixing ram is not supported for this reason.

you can try 1 stick, reset the bios to defaults and add 1 stick between reboots and see what happens.

but to answer your question (one of them :): ) yes advanced RMA is possible but you must CALL CS explain its a second RMA they will email a prepaid mailing label and get your credit card info for the advanced RMA.

you would have been better with a single 4/8G kit to avoid this as it may happen again as mixing kits with even the same lot#s are still not officially supported, suggested or recommended.

it really sounds like you need to call CS and explain whats going on and see what they say. you can even ask to speak to Ram Guy. if you get his voicemail leave your number, he calls back!

good luck.

Thanks for replies guys.

So... just to clarify Synthohol: using sticks from different kits and or lot #'s can actually result in a dead stick/sticks?...

that's what i would have thought.

I guess i'll call corsair tomorrow.

I'm wondering if they will have a 4 x 2gb matched kit (or 2 x 4gb) matched kit that they could send me, and have me send back both kits i've got. ... 'officially supported' would be nice.

What about the whole having tested the now dead module for 1hr in memtest / 2 complete passes, and then it die when installed alongside the other 3 modules (that are still 100% good) thing?

is this just a fluke, and i should expect niether module in the next (third) kit to not throw memtest errors(a.k.a die) after having been installed alongside the other 3 modules? Seems like really really low odds of this happening.

I understand that 'mixing' kits, even with matching lot #'s, may result in instability, bsod's, or whatever. Not RAM stick death though ;)

What about the whole having tested the now dead module for 1hr in memtest / 2 complete passes, and then it die when installed alongside the other 3 modules (that are still 100% good) thing?

 

is this just a fluke, and i should expect niether module in the next (third) kit to not throw memtest errors(a.k.a die) after having been installed alongside the other 3 modules? Seems like really really low odds of this happening.

 

I understand that 'mixing' kits, even with matching lot #'s, may result in instability, bsod's, or whatever. Not RAM stick death though ;)

 

he motherboard is set to run ram @ SPD, and there is no ram voltage control afaik. No overclocking involved (no option to with this mobo). No other issues with the system either (comonents dying).

It's not a dead stick if it passes memtest by itself. Since you returned the wrong stick originally you say , now you have a possibly a third kit in the mix..

Your issue sounds mire MB related than anything. A bad slot, or something. It could even be a bent pin in the CPU socket. It could also be the inability to adjust voltages for the increased load on the memory controller. That actually makes more sense than anything....

It's not a dead stick if it passes memtest by itself. Since you returned the wrong stick originally you say , now you have a possibly a third kit in the mix..

 

Your issue sounds mire MB related than anything. A bad slot, or something. It could even be a bent pin in the CPU socket. It could also be the inability to adjust voltages for the increased load on the memory controller. That actually makes more sense than anything....

I did not return the wrong stick.

The "dead" stick, have you tried it in different slots?

I have to admit wired, i hadnt tried the dead stick in other slots.

So right now i went and tried both the sticks from the set that i set aside as having the isolated stick that was throwing memtest errors.

I tried both sticks in both slots #1 and #2 (individual stick in slot 1 alone, then slot 2 alone), and neither are throwing memtest errors immediately as one of them was before. (it was immediate, within 5 seconds tops).

...

....

Might this be a simple case of a contaminant that was physically in slot #1?

Could be as simple as that. I did have one pair of sticks running (including one in slot #1) and it was fine.

EDIT:

i've put both sticks from the kit / pair that had the 'dead' stick in now, and am running memtest86+.

I've absolutely no idea if it's going to throw errors at this point.

Never had a situation quite like this, i seated & unseated the 'dead' stick three times i think, in slot #1 and it threw errors immediately each time. other sticks in slot #1 did not throw errors. :confused::confused::confused:

(no, i haven't cleaned the contacts on the RAM. They didn't look like they needed it)

i'm going to leave the machine running memtest86+ for a few hours, and will report back.

For reference, the two 'good' sticks are NOT installed at the moment.

Only the kit that had a 'bad' stick is installed, in slots 1 & 3.

I've a feeling that this might just boil down to a contaminant in slot #1, or on that one stick that seemed bad. 4 minutes of wall-time right now, and it took less than 5 seconds last time around with one of the sticks in there now to throw thousands of errors.

:o::o::o: ... yup. Embarrassing to say the least.

It's not a dead stick if it passes memtest by itself. Since you returned the wrong stick originally you say , now you have a possibly a third kit in the mix..

 

Your issue sounds mire MB related than anything. A bad slot, or something. It could even be a bent pin in the CPU socket. It could also be the inability to adjust voltages for the increased load on the memory controller. That actually makes more sense than anything....

peanutz, i just want to clarify this.

the stick in question passed memtest for an hour on it's own in slot #1. So did it's sibling. I then installed all 4 sticks and one of them died shortly thereafter. I isolated that stick and set it aside, along with it's sibling. These sticks were the ones that corsair sent me as replacement for the originally isolated dead stick. Once again, that stick was only tested in slot #1. For some insane reason, i was under the impression that you had to have ram in slot 1 to boot. :o:

These same sticks are now the ones that are running memtest at home.

20,xxx,xxx errors after 2.5hrs with two sticks from the suspect set in slot 1 and 3.

I checked the bios, indeed i can adjust ram timings, the auto detected speed is correct however for cm2x20248-6400c5c (800mhz 5-5-5-18)

My wife being home, testing will come to and end.

Should i bother trying to isolate which of these two sticks is bad, or should i try instead these same two sticks in slots 2 & 4 of the motherboard?

ok, well, i had the time... so i found the suspect stick of RAM.

fwiw, I tried turning off legacy usb support, all that did was make memtest not respond to the keyboard. Turnd that back on.

I have tested the faulty stick of ram in all 4 slots, running the ram @ 667mhz 5-5-5-18 ... just in case ya'll said to try at a lower speed. There is no voltage control that i can see in the bios for my motherboard. Apparently some folks can see it in there, but i dont think the latest versions of the bios from intel support that.

Every slot has the same results with this stick of RAM, thousands of errors almost immediately (within a few seconds) of test #1.

I'm guessing that the earlier individual testing was with cold ram. After having been hit by memtest for 2.5hrs, the ram has heated up thoroughly, and now throws errors immediately.

one complete pass without errors @ 667-5-5-5-18 on the trouble free kit of RAM, slots 2 & 4.

I think I've got another kit of bad ram, not a faulty motherboard.

Sound right?