RAM dying, why?

GoogalyMoogaly

GoogalyMoogaly

GoogalyMoogaly

Soldato
Joined
29 Aug 2010
Posts
8,122
Location
Cornwall
So I've just had to RMA some RAM and I don't think it was the first time.
Are there any common reason for a stick of RAM dying?

The Motherboard is Fatal1ty AB350 Gaming-ITX/ac and the RAM is Team Group Dark T-Force 16GB (2x8GB) DDR4 PC4-24000C16 3000MHz.
It always seems to be the RAM stick nearest the CPU socket (from memory, pun sort of intended).
It's not overclocked, it's just using the XMP profile.

Could it be a fault RAM slot on the motherboard? Could the CPU cooler be blowing warm air onto the RAM stick and causing it to overheat (seems unlikely)?

Overclockers have been really good, but I'd rather it not keep happening.

Getting a new mITX motherboard would require a new CPU too, which is getting quite expensive for what I use the PC for so I'm hoping there's something I can do.

Any ideas?
 
Perhaps there's some bug in the BIOS that's overvolting the ram? I've never had ram dying but I have from time to time on boards I've owned in the past seen default settings on there which I thought were a tad high. Try updating bios perhaps?

Seems odd if it's only one stick though.
 
Last edited:
3 things that can damage RAM, excess voltage, static discharge and heat.

The chance of having a dodgy RAM stick 2 or 3 times is pretty unlikely. It's either that slot is putting out to much voltage, or the heat from the CPU cooler is doing it.

I would either rotate your CPU cooler, or I would down clock the memory to 2666 and get the voltage down to 1.25, this will reduce heat. Also that memory has heatsinks on it, there designed to have cool air pass over them. If you pass hot air over them the opposite will happen, the heat sinks will help accumulate heat!

Now this has just re-freshed my memory. Back in 2008, I had some Corsair 1600 DDR3 that was cutting edge at the time and was designed to run at higher voltage to get that speed. Just like you I had a CPU fan pointing at the memory. After a period of time the computer would crash, I put this down to the CPU cooler overheating the memory (however the memory never failed). To solve the issue I down clocked the memory to 1333.
 
In my experience, heat. Put your finger on one or more of the sticks after something intensive like benching or gaming and feel them if they're hot to the touch they're too hot b-die seems particularly sensitive to heat degradation
 
Excess voltage is one of the main killers in my experience - some boards have way too high automatic/XMP settings especially if certain enhancements are enabled. Problems with heat really only happen if something creating a lot of heat is in very close proximity i.e. laptops where the RAM modules can be the opposite side of the mainboard to things like the CPU or GPU, etc. or your case cooling is very sub-optimal.

EDIT: Some boards also apply higher voltage than it says in the BIOS.
 
Last edited:
3090FE in my case the ram sticks sat directly above the upwards facing fan and didn't have a case that could fit a riser at the time, not sure who at Nvidia thought it was a good idea to exhaust hot air all over the motherboard components. Latest card has the regular 3 fans at the bottom thats much better.
 
Last edited:
It is in a SFF case (Silverstone ML08), so airflow isn't great. I suspect it is getting some air from the CPU cooler. Not sure if this is good (as it's airflow) or bad (because it's warm).
What gets me is that it's only 1 stick. If it was voltage I'd have thought it'd affect both.
One small thing, the motherboard (or CPU) won't run the RAM above 3000MHz, so I'm wondering if the voltage is too high for the speed?
 
You must log in or register to reply here.
Back
Top Bottom