Can RAM go bad?

[Aesalon]

hey guys, just got a quick question for you. can RAM just randomly go bad? all of the sudden in the last 2 days, my computer has been getting blue screens with the 0x00000008, 0x00000050 errors. after some quick research, looks like the RAM I'm using, the OCZ 2GB Platinum (5-2-3-2). my current setup is as follows:

DFI NF3 Lan Party
Athlon64 3200+
2GB OCZ Platinum
2x250GB Maxtor RAID Array
Nvidia 6800GT
Antec NeoPower

I was running the FSB at 240mhz with a 180% RAM divider at 2.9V with 7-3-3-2.5 timings and things were just fine. then randomly, I come home to my computer blue screened. figured it was a fluke and rebooted. then it started happening more and more. I ran MemTest and came up with tons of errors on test 5.

I checked all my PSU voltages and they seem to be fairly strong (5v - 4.93, 3.3v - 3.24, 12v - 11.58). all temperatures and such are fine as well. I dropped everything to default settings and I'm still getting errors but it seems to be blue screening less.

I'm doubting that it's anything else other than the RAM. the thing is, I've done literally nothing to this machine in the last couple of days with the exception of downloading some music.

my next step is to contact OCZ but I just wanted to try here first. can RAM just go bad after not even using it for a month?

 

[Footer4321]

i have had a stick of mushkin go out on me... called them and they replaced it no questions asked....

 

[Anime_Fan]

Ram is pretty sensitive stuff and yes it can go bad. Whats the stock voltage for that RAM? 2.9 seems a little high to me. Whats the speed rating?

 

[XeRo_Dark]

7-3-3-2 timings?
whoa dood thats high

 

[Talonz]

What's a 180% ram divider?

They are normally given in ratios. 5:9?

Basically, is your RAM being run out of spec, and if so, does it give you errors at stock?

 

[ZOMGWTFBBQ]

7-3-3-2 timings?
whoa dood thats high

Dude hes talking about 2.5-3-3-7, CAS 2.5, not CAS 7 lol.

CAS at the end is the 'proper' way to write it, just no one gives a poo really.

 

[Ingonuts13]

I think 2.9 is outside the rated spec of the RAM so it is likely that it could be failing due to being too hot for too long.

How long did you run the ram at that Voltage and how is your case cooling? Also, have you tried running the RAM at stock as a previous poster suggested?

Also, does that board not hit a voltage wall at 2.8V or 2.85V?

 

[Aesalon]

Ram is pretty sensitive stuff and yes it can go bad. Whats the stock voltage for that RAM? 2.9 seems a little high to me. Whats the speed rating?

the RAM is pc3200. oddly enough, never thought that 2.9v would be an issue as it never was with my corsair 1gb pc3200 kit which ran at that voltage for over a year.

 

[Aesalon]

What's a 180% ram divider?

They are normally given in ratios. 5:9?

Basically, is your RAM being run out of spec, and if so, does it give you errors at stock?

the ratio is 9/10 and as for testing, yes, test #5 goes nuts at everything default.

 

[Aesalon]

I think 2.9 is outside the rated spec of the RAM so it is likely that it could be failing due to being too hot for too long.

How long did you run the ram at that Voltage and how is your case cooling? Also, have you tried running the RAM at stock as a previous poster suggested?

Also, does that board not hit a voltage wall at 2.8V or 2.85V?

to be honest, the RAM never felt too hot so I never thought it would be a problem. plus I ran a 1gb pc3200 kit at that same voltage for over over a year.

I ran the ram for less than a month and my case cooling has a 120mm in the front and a 120mm in the rear. the only problem is there is a little heat build up in the middle of the case because the rear fan is blowing fresh air in through the radiator (water cooling) and the front fan is blowing air in to provide adequate cooling on my raid array. I would flip the front fan around but I worry the hard drives would get too hot and I didn't want to push hot air out through the radiator. heh, this might warrant some changes in airflow.

now the board's ram voltage can go higher, I just didn't want to go THAT high. the whole intention was to boot up with the overclock. I couldn't match the FSB so I had to use a divider, which was a major bummer. figured if I loosened up the timings, I'd hit my original overclock with smaller ram but I couldn't. I know that more ram means more trouble hitting good timings or overclocking but I've read plenty of success stories. all that's left is to hope that OCZ can replace them. if not, I think I'm still within warranty/replacement time from monarch computers.

 

[Ice Czar]

absolutely

especially overclocked RAM with a voltage bump
ESD & Electromigration Rant

heat and voltage being the two components of Black's Equation accelerating electromigration displacement


http://www.memtest.org/

note fully testing RAM with software requires some dancing

http://www.lostcircuits.com/memory/ultrax/

The advantage of having a non-DOS or Windows based test program is quite simple. If an operating system is running, the area of memory containing the OS or reserved for the OS cannot be tested correctly. Usually, the workaround in a situation like this is to alternate DIMMs between two memory slots since the operating system will be loaded onto the higher (Win9x) or the lower number DIMM (all other operating systems). Thus, by switching DIMMs back and forth between slots, one DIMM can always be kept free of the OS which, then allows complete testing of the module. The natural drawback is that every DIMM has to be tested twice which can be rather time-consuming.