Are Ram failures common? and "just happen"

[surrealchereal]

Or is it the cheap stuff and that's why everyone goes for the fancy brand names.
I must apologize if is, I kind of thought folx went for the fancy expensive stuff like people go for designer name clothes. Kind of like Mushkin is a nerds Calvin.

[Steve R Jones]

Yes ram can just fail/go bad. Probably doesn't matter is its cheap or expensive. It just hurts more when its the good stuff.



[This message has been edited by Steve R Jones (edited 07-28-2001).]

[Roy]

I buy RAM by price and specification. I've had cheap "generic" and just received some cheap Mushkin(!) RDRAM for a new P4 system I'm building.

I've experienced two failures, one name-brand, one generic. What I've learned is to consider the possibility of RAM failure when strange symptoms show up. It can seem like a virus attack. In one case, the faulty stick caused totally different problems in three different systems.

[Axel]

If the ram goes in and functions - it's rare that it "just fails" - there is typically some sort of electrical "event" that causes it to go bad.

Things like lightning, power spikes, powering the machine off in the middle of the hard drive spinning up - several different things can impact RAM and other system components -

I've even seen miss-matched RAM keep the floppy drive from being recognized by an ASUS motherboard - I'm 100% sure it wasn't the ribbon cables - tried 3 known good - and it wasn't the drive - tried 4 known good - - but when I took out some of the RAM, some of which didn't exactly match the other three SIMM sticks - the drive was again recognized - the box otherwise worked perfectly - just couldn't read a floppy disk.

I've seen a brand of power console ( MASTER ) that had an electromagnetic signiture that caused sticks of NEC RDRAM to show up with errors. - took out the power switch pod and the systems worked fine....

Dirty power with spikes and half-sine waves can cause RAM errors -

I've seen someone pack as many hard drives as they could in a mini-tower - putting one low under the 2 5.25 bays - right over the ram and the drive motor impacted the RAM -

Heat can do it as well - lose a cooling fan and you might fry your RAM -

too much humidity - I.E. moisture - can corrode the contacts and cause your system to fail - typically the RAM or processor goes first in this case - heat aggrivates this situation. Not enough humidity can cause the plastic components to dry out and crack -

Typically - the connections on the RAM are gold or gold plated so they don't corrode - but the motherboard SIMM / DIMM slots aren't always gold plated, but can be copper or aluminum - different metals coming in contact cause what is known as a galvanic reaction that produces a charge - over time - this will cause you problems - but memory and processor slots aren't made that way any more because of those reactions. This reaction is also aggrivated by heat.....

Physical shock can also cause failures - I.E. - don't drop the computer.

Over time, the expansion and contraction of the material making the RAM will eventually cause hair-line fractures in the material - as these become numerous enough to change the electrical resistance of the RAM components, it will fail.

But for RAM to "suddenly" fail on a relatively new box that you don't frequently abuse and have on a good quality power strip w/ a surge protector is rare.

There is a term in computing called "burning in" - for end users, this term means leaving a computer on for a day or so after you first get it - Any truely faulty components will die in a day or so.

Some smaller low volume companies with high dollar clients will build the users a system and leave it running on the bench playing a music CD for 2 or 3 days - then, if nothing fails, they box it up and ship it. Under those conditions, they can be very sure they are shipping out a good system to their user as far as the hardware is concerned.

The big PC companies cannot afford to do this - too much over-head for the quality control department cuts into profits. They take a sample out of every major production run and calculate average time to failure.... They set standards based on most PC's only being around 2 years or so - figure out how much use John Q. Public might put on the machine in that time and build with components that will make it that long. Nothing is truely built to last these days....

[surrealchereal]

Wow thanks for the info,,
one thing that always bothered me, (ok a couple) Getting an AGP video card in and the ram. (not good, very bad design)
The fan from the heatsink for the chip pelting the ram wit dust from the fan, causeing a really bad build up. But until I had a stick go out I never treated it with respect. (toss it in my purse have it in my mouth ... http://www.sysopt.com/forum/redface.gif)
Thanks for the added info Axel! http://www.sysopt.com/forum/wink.gif

[wedor]

I've had RAM go bad from both generic and quality manufacturers, the big difference has been in how the problem is resolved. I have started buying Mushkin RAM almost exclusively because when I have any problems all I have to do is call, a new module is on the way overnight and shows up at my house, or anywhere else I need it, the next morning usually by 9:30. In contrast with generic RAM I have to argue with the support people on the other end, then send in the suspect bad module, and then wait for weeks for them to replace it. So if you are into gambling it is sometimes better to go with the cheap stuff and take your chances that nothing will go wrong, but if it does.....well that's why it's gambling.

[Richard_Cranium72]

I killed a stick of 128 on a mobo last month.

The mobo had SIMMS and DIMMS, being stupid like I am, , I stick a nice new kingston 128 in the DIMM side.

It won't see the memory in POST,, hmmm

Try as I might it don't work.

Later,, I realize the mobo furnished 5.0 volts to the 3.3 volt stick..

Time for a RAM "Fruneral"(spelling on purpose) http://www.sysopt.com/forum/smile.gif

[Dudster]

There's definately a lot more to buying brand name RAM than fashion http://www.sysopt.com/forum/smile.gif I used to use 128MB of generic and only discovered that it was causing instability (especially in Quake 3 Team Arena) when I replaced it with 256MB of Micron RAM. I'd never buy generic again and I'm not even into overclocking.

[Cadfael]

You can also kill simms and dimms with static electricity (and any other component for that matter), and there is an easy way around it. You don't need a wrist-strap to discharge static... just turn of the power point at the wall, plug in the power lead, and keep touching the case before you handle any component... er... at least in the UK you can, we have three pin plugs with an earth lead.

[Scorpio69]

Wow Axel, that was a lot of info, but you didn't mention ESD (Electro-Static Discharge) from improper handling, though. It sounds like some memory modules have been abused here. I am definitely guilty of this, myself.

With the prices these days, it's not going to break you to buy a new stick, even the good stuff. What is really a pain is having symptoms that you just cannot figure out. I wish RAM would either work properly or just DIE. I hate the gray areas in between.

As someone mentioned, learn to identify and consider the specs of the RAM. I have run into modules that had me pulling my hair out, because they work fine in one machine, but will not run on another. If you have generic RAM, often it is difficult to determine the specs.

[Richard_Cranium72]

Cadfael is correct

That is "written" procedure, I forget where I saw it.

The main point is to keep the power plugged IN, that is the only way for the ground wire to drain the static.

DrVette

[Cadfael]

But... do remember to turn the power off at the socket, if you power point don't have a switch, use an extension lead that does.