i7 860 damage w/DRAM @ 1.66V?!

[slugman]

I have an i7 860 and Gigabyte P55A-UD4P board.

My RAM calls for a minimum operating voltage of 1.65V, so I had it set to 1.66V in the BIOS.

I just discovered that this can damage the memory controller or the i7, which has a max voltage of 1.65V?!

What are the chances I incurred damage?

As anyone else tested this?

Why is this not a bigger issue?

I am slightly (ok more than that) concerned.....


any input?

 

[aigomorla]

I came in here about to give u a little butt whipping...

but as i read your thread im seeing your talking about ram.

As long as your not pushing your ram too hard... meaning doing TON's and TON's of ram dependant applications non stop.. u'd be safe @ even 1.7 with active cooling.

1.65 is within default voltage range for some ram, so yes its safe.

Keep it below 1.7V if ur really paranoid.. or get an active fan on your ram.

 

[slugman]

LOL changed title, I see how it sounded like I meant CPU.

But I'm not worried about the RAM, it's operating range is 1.65V-1.85V. Rather, I am worried about the memory controller on the CPU!

 

[Rubycon]

There is a +500mV differential limit. So if your VTT is 1.1 and you set your ram at 1.65 you are technically outside of that limit.

Most folks when o/c will run VTT around 1.3 - 1.35V which means your ram can go as high as 1.8-1.85 volts! (many good DIMMS can handle it too but it's not necessary)

 

[slugman]

motherboard automatically set Vtt to 1.17V when I had it at 1.66V. Of course once I learned the danger, I clocked my RAM right down to 1.5V and relaxed the timings to 9-9-9-28 from 9-9-9-24. Now Vtt is at 1.10V.

I notice that easytune 6 says my RAM's power-on voltage is 1.62V. Should I increase Vtt to 1.15V so that even at power on it's within this limit?

EDIT NOTE: naturally, I am not using easytune 6 to set the clocks, just checking what I set in the BIOS while in windows.

 

[Rubycon]

I'm not familiar with that board (i.e. hands on) but if you're overclocking I'd set VTT/VDIMM/VCORE/UNCORE voltages manually. Most boards will ramp up volts when BCLK is increased at much higher rates than typically necessary. This creates more heat and even wear and tear on your cpu if you're shooting for the moon.

 

[slugman]

I had a mild overclock of 3.36 but not anymore. I've got what is (to me) a lot of money invested in this system and don't want to ruin anything.

Currently Vtt@1.17, RAM @ 1.62V, 12x(1600Mhz- not overclocked, what it's listed as) bclk 133.

Although the RAM suggests at least 1.65V I'm going to do stability testing and see if all is well.

Is a Vtt of 1.17V safe? I've seen somewhere that the i7 shouldn't go above 1.115v??

 

[Rubycon]

Safe is up to 1.4 actually so you are fine.

 

[aigomorla]

There is a +500mV differential limit. So if your VTT is 1.1 and you set your ram at 1.65 you are technically outside of that limit.

In the Process of getting debunked...

And theres enough creditable people on the debunking case for me to almost accept it.

 

[MJinZ]

I have an i7 860 and Gigabyte P55A-UD4P board.

My RAM calls for a minimum operating voltage of 1.65V, so I had it set to 1.66V in the BIOS.

I just discovered that this can damage the memory controller or the i7, which has a max voltage of 1.65V?!

What are the chances I incurred damage?

As anyone else tested this?

Why is this not a bigger issue?

I am slightly (ok more than that) concerned.....


any input?

None.

 

[slugman]

@MJinZ you mean no chance I incurred damage or none as in no input?

Though microcenter lists the min voltage as 1.65V, newegg lists the min voltage as 1.55V for the exact same RAM (right down to the model number).

http://www.newegg.com/Product/Produc...211409&Local=y

Assuming it passes memtest86+ 4.1 overnight (30 min and so far OK) would you guys assume that it's good to go at 1.56V?

Would linX/prime95 (ok more concerned about stuff I'll actually be working on) fail because of RAM when memtest has proven stable?

 

[MJinZ]

There is no chance you damaged it. The RAM can probably run at 1.7V and be just fine.

Setting VDIMM To 1.65V doesn't mean the RAM is getting 1.65V at all times. It could be getting anywhere from 1.63 to 1.67 depending on how good or poor your board is, and how accurate your BIOS is.

 

[slugman]

Well, thanks. To be clear, I'm not worried about the ram at all: its the memory controller in the CPU that's spec'd for 1.65V max. The RAM can handle up to 1.85V according to A-DATA. It was the controller spec I hadn't known about. I mean jesus, they should print that all over the RAM package before saying "designed for core i3/5/7 and AM3 at 1.65-1.85V" ... next to that it should read CORE SERIES 1.65V ONLY or something? I mean my mistake must've been made before?

 

[mbevolution]

does the ram boot with lower voltage (1.5v)? the kingston rams i have and the corsair xms ones i installed for a friend both recommended 1.65v but they boot absolutely without a problem at 1.5v, in fact the xms is running at 1.4v without a single problem.

 

[dajeepster]

@MJinZ you mean no chance I incurred damage or none as in no input?

Though microcenter lists the min voltage as 1.65V, newegg lists the min voltage as 1.55V for the exact same RAM (right down to the model number).

http://www.newegg.com/Product/Produc...211409&Local=y

Assuming it passes memtest86+ 4.1 overnight (30 min and so far OK) would you guys assume that it's good to go at 1.56V?

Would linX/prime95 (ok more concerned about stuff I'll actually be working on) fail because of RAM when memtest has proven stable?

or better yet... just go to the manufacturer's website

http://www.adata-group.com/?action=product_specification&cid=5&piid=91

 

[slugman]

yeah turns out I have the AX3U1600GB2G9-2G sticks, different from the AX3U1600GB2G9-AG sticks that micocenter lists on their site, though that's where I bought them. A-DATA isn't clear on this either, but mine are 1.55V.

I had text chatted with intel (without giving personal info, of course, haha) to see if 1.66V was OK and was told I had "almost certainly" damaged my memory controller. But maybe they just say things like that. If it's stable in Memtest86+ for 10 hours, that probably means memory controller is fine? I'm in 7 writing this now and it seems fine.... linx also works normally.

 

[alyarb]

heh you're fine. if anything was damaged there's a 90% chance it wouldn't even POST.