Is 1.65V Memory bad for Sandy Bridge Processors?

[ensign_lee]

I just now read that Sandy Bridge processors are supposed to work with 1.5V memory instead of 1.65V memory. The tolerances are supposed to be +-5%, which 1.65V memory definitely doesn't fall into.

What exactly could happen by running my memory at 1.65V instead of 1.5V?

 

[kittysox]

I've been running mine at 1.65v since they launched on jan 9 with no issues whatsoever and the guys at the mushkin forum assured me that the memory specs for Sandy Bridge are exactly the same as for any other i3/i5/i7.

 

[Zap]

What exactly could happen by running my memory at 1.65V instead of 1.5V?

1.65v appears to be the limit that the CPU's IMC can handle in all the recent Intel CPUs. What can happen is the CPU can die. Actually, it is more of a voltage differential between the IMC and something else. In any case, don't go over 1.65v and you should be fine.

Personally I prefer lower voltage memory because any DDR3 rated at 1.65v is factory overclocked and basically wasn't originally designed to run at those speeds.

 

[ensign_lee]

Zap: ORLY? I didn't know that.

I did notice that my corsair RAM that was specced at 1.65V and DDR3-1600 can run just fine at 1.5V and DDR3-1333. I confirmed with Corsair that I shouldn't experience any problems doing that either.

So should I just go ahead and run my RAM at 1.5V DDR3-1333 then?

 

[ensign_lee]

*bump* - Is it common knowledge that any ram rated at 1.65V is just 1.5V memory that's been overclocked by the manufacturer?

 

[Zap]

So should I just go ahead and run my RAM at 1.5V DDR3-1333 then?

Sure.

*bump* - Is it common knowledge that any ram rated at 1.65V is just 1.5V memory that's been overclocked by the manufacturer?

Well, apparently not common enough. :\

There's a group called JEDEC that puts out the specifications for RAM. JEDEC specifies that DDR3 should be 1.5v.

http://en.wikipedia.org/wiki/DDR3_SDRAM

DDR3 memory provides a reduction in power consumption of 30% compared to DDR2 modules due to DDR3's 1.5 V supply voltage, compared to DDR2's 1.8 V or DDR's 2.5 V.

I've bolded the pertinent part.

Regarding "factory overclocking," since DDR3 is supposed to run at 1.5v and a typical "trick" for overclocking is to increase the voltage... :hmm: I'm sure you can come to your own conclusions.

 

[Spikesoldier]

snipped from this thread: http://forums.anandtech.com/showthread.php?t=2140386




really the whole 1.65v vs 1.5v thing is just that the 1.5v memory is made of better modules than same kit rated at 1.65v.

just because something is 1.65v doesnt mean its for bloomfields only. the whole perception with nehalems and running high dram voltages frying cpu's is a function of vdimm and vtt.

basically with nehalems according to intel, you never want vdimm - vtt = .5v or more.

i.e.

vtt 1.21v
vdimm 1.65v

1.65 - 1.21 = .44 ok


vtt 1.21v
vdimm 1.75v

1.75 - 1.21 = .54 too high


vtt 1.33v
vdimm 1.75v

1.75 - 1.33 = .42 ok


this difference of vtt and vdimm voltage is what causes the memory controller on nehalems to become 'fragile' to vdimm increases when using custom voltages.

obviously when you use 1.5v modules or even better, 1.35v, you wont be running into this problem unless you are absolutely trying to squeeze everything out of it.