Nvidia acknowledges DIMM failure on 680i motherboards (sort of)

[x800professor]

Important message regarding memory and the EVGA 680i motherboard

NVIDIA has investigated end user reports of high performance DIMM failures on the NVIDIA nForce 680i SLI-based platforms. During this process we have been in close contact with DIMM manufacturers and the DRAM manufacturers they rely on to understand the failure scenario. By working with our community, we believe that the observed failure is a breakdown of the silicon in the DRAM caused by the prolonged application of 2.4V on the voltage rails of the DIMMs.

NVIDIA’s own internal testing has observed this failure on multiple motherboards using different chipsets (both NVIDIA and non-NVIDIA chipsets). This issue is not directly related to motherboards using the NVIDIA nForce 680i SLI MCP or other chipsets.

If you are using this type of memory and are experiencing this issue, NVIDIA recommends contacting your memory manufacturer or system manufacturer for additional information and warranty information.





I've had 2 sticks die on an EVGA 680i motherboard while only running at 2.2v. They were both intended to run at 2.2v. They were both corsair XMS2 and from two different matched pairs.

[newtekie1]

I had the actual motherboard die, it simply stopped booting with memory in it. If I put memory in it, it would turn on for a second then turn off, like there was a short somewhere. However, without any memory the board would start up and stay on just fine. I tried different memory, different slots, and nothing worked.

I wonder if that has anything to do with this problem? The RAM was only running at 2.1v and apparently there are a lot of poeple complaining that these boards simply don't last if you are using RAM voltages higher than the standard 1.8v.

[x800professor]

I had the actual motherboard die, it simply stopped booting with memory in it. If I put memory in it, it would turn on for a second then turn off, like there was a short somewhere. However, without any memory the board would start up and stay on just fine. I tried different memory, different slots, and nothing worked.

I wonder if that has anything to do with this problem? The RAM was only running at 2.1v and apparently there are a lot of poeple complaining that these boards simply don't last if you are using RAM voltages higher than the standard 1.8v.

I've had two graphics cards die along with 2 sticks of ram. I have to wonder if the board had anything to do with the cards. They were both x1950XTs. My 8800GTS works just fine now. Those boards are garbage, plain and simple.

[newtekie1]

The new revision seems to be much better, but the original revisions were plagued with problems. I am on my 3rd board.

[x800professor]

The new revision seems to be much better, but the original revisions were plagued with problems. I am on my 3rd board.

My 2nd. Although I've had two RMAs already. Here's my board's numbers.

[newtekie1]

Mine are identical to yours.

[x800professor]

Mine are identical to yours.

This board seems to be ok. Well, aside from the fact that one of the X1950XTs died on it. It also has a non-functioning fan header, but I don't care.

[newtekie1]

It still has the annoying surface mounted components on the back of the board behind the processor though, which means most 3rd party heatsinks won't work without modifying the mounting hardware. :cry:

[x800professor]

It still has the annoying surface mounted components on the back of the board behind the processor though, which means most 3rd party heatsinks won't work without modifying the mounting hardware. :cry:

The original thermaltake big typhoon works. All you really need is some sort of non-conductive buffer material for other things (at least most of the time). So it isn't THAT big of a deal, but it is annoying. You'd think that they'd figure that an enthusiast class board would end up with enthusiast class CPU cooling attached to it. Of course, that would be giving them too much credit.

[tkpenalty]

*COUGH COUGH COUGH* Someone isnt happy... Man, is this even a high end board? Like prepare to RMA your RAM and GPU again T_T...

[x800professor]

*COUGH COUGH COUGH* Someone isnt happy... Man, is this even a high end board? Like prepare to RMA your RAM and GPU again T_T...

Everything is ok thus far. I have had zero stability issues that weren't due to the heat (it's waaay hotter than it was a few weeks ago here). Of course, that doesn't mean anything. If the Motherboard is overvolting anything, it will take a while to degrade.

[D007]

Everything is ok thus far. I have had zero stability issues that weren't due to the heat (it's waaay hotter than it was a few weeks ago here). Of course, that doesn't mean anything. If the Motherboard is overvolting anything, it will take a while to degrade.

I hate this mobo... 1.5 volts and I cant get a conroe above 3.2 ghz withouth shooting up to insane temps.. the fact that i have to run @ 1.5 v in the first place just to get to 3.2 ghz is enough to make me want to slam it into a wall.. friggin lost my box so newegg wont return it.. gr8... target practice anyone?

[x800professor]

I hate this mobo... 1.5 volts and I cant get a conroe above 3.2 ghz withouth shooting up to insane temps.. the fact that i have to run @ 1.5 v in the first place just to get to 3.2 ghz is enough to make me want to slam it into a wall.. friggin lost my box so newegg wont return it.. gr8... target practice anyone?

It really depends on how hot your room is and what kind of cooling. With a big typhoon and a cool room, I can hit 3.6 easily on less than 1.5v. By the way, it doesn't give the CPU the voltage you ask for. It is usually about .03-.05v lower. So, 1.5 is really 1.45-1.47v.

[Ketxxx]

Damn i would of taken a blowtorch to the POS by now, then mailed it back to the manufacturer with a lil note reading something along the lines of "do not make such shit again"

[D007]

It really depends on how hot your room is and what kind of cooling. With a big typhoon and a cool room, I can hit 3.6 easily on less than 1.5v. By the way, it doesn't give the CPU the voltage you ask for. It is usually about .03-.05v lower. So, 1.5 is really 1.45-1.47v.


thats a good point actually.. it is lower, i forgot about that. but the temps still kill me in an air conditioned house that stays very cool with 2 120mm fans and 2 80mm fans running full speed including the cpu and gpu fans at 100%.. Im buying water cooling very soon i guess.. hopefully this friday I can afford it after my car payment lol.. hey can anyone recommend a good water cooling for this type of set up? something that is capable of future useage as well for like the quad cores and all that? im going to get some more thermal paste and redo that as well.. i didnt like the factory paste that came on the cpu, looked cheesy..

[newtekie1]

Water cooling might be a bit much, however putting 1.5v though a processor on the stock Intel cooler isn't a good idea.

[D007]

Water cooling might be a bit much, however putting 1.5v though a processor on the stock Intel cooler isn't a good idea.

id rather run at 45 than 55 anyday lol.. the cooler i can run the better ya know.. im still going to keep the fans on it as well just for safetys sake..

[x800professor]

id rather run at 45 than 55 anyday lol.. the cooler i can run the better ya know.. im still going to keep the fans on it as well just for safetys sake..

On the other thread you said you had a stick die. Get CPU-z and see if you have the same board I have. If not, you have the old revision that seems to be the worst motherboard ever.

[dsdsdk]

#0, your title is misleading. The board isnt the cause of ram failure. Not rev AR or A1... Besides, memory vreg hasnt been changed when they made the new revision. So saying the revised board is better is bs. People have measured ram-voltage with DMM and it was spot on.

Running the ram out spec is the problem. Micron D9 are 800mhz chips, rated for 1,8v. And ram "labelers" run them at 1066+ with 2,2v+. They're bound to fail in a lot of cases.

And the fact that 800mhz are running fine in most cases, when 1000mhz+ are failing, seems odd that it should be the board killing them off.

[D007]

On the other thread you said you had a stick die. Get CPU-z and see if you have the same board I have. If not, you have the old revision that seems to be the worst motherboard ever.


no no no, the mobo it's self is having issues with killing high end memory.. i just had my 1st memory issue last night under circumstances i wouldnt think i should of had it in.. IE: 4-4-4-12 at 2.0 v running 3.0 ghz.. that never happened to me before..

[D007]

#0, your title is misleading. The board isnt the cause of ram failure. Not rev AR or A1... Besides, memory vreg hasnt been changed when they made the new revision. So saying the revised board is better is bs. People have measured ram-voltage with DMM and it was spot on.

Running the ram out spec is the problem. Micron D9 are 800mhz chips, rated for 1,8v. And ram "labelers" run them at 1066+ with 2,2v+. They're bound to fail in a lot of cases.

And the fact that 800mhz are running fine in most cases, when 1000mhz+ are failing, seems odd that it should be the board killing them off.


uhm, the board "is" the cause of memory failure in most cases..lol.. you can do the exact same thing with virtually any other board and not have your memory fry.. 2.1-2.2 v is not a huge overclock for memory
( like 2.4 is ), it has actually become a very standard memory overclock and should be expected from a board that boasts about having "extreme overclocking ability" it should live up to the hype, not melt memory sticks..

my memory issue is more in regards to having all kinds of corruptions and having to reformat my hd'd on a regular basis due to this.. it's makin me nutz.. i had one memory "issue" one that i never had before under the exact same circumstances.. so why would i have this issue now all of a sudden? just seems like maybe things get cumulatively worse, i hope thats not the case but idk..

I have the A-2 board,( p26 bios ) which i think is the new one thank God.. but still i had to reformat like 3 or so times in the last month (30 days).. that seems excessive to me.. i had werid boot errors saying "insert system disc" for a while but then it just went away lol.. thats weird in its self to me.. problems never just go away, they hide and become worse..lol..

honestly if i had the box still i'd of rmad it and got a striker extreme.

[x800professor]

#0, your title is misleading. The board isnt the cause of ram failure. Not rev AR or A1... Besides, memory vreg hasnt been changed when they made the new revision. So saying the revised board is better is bs. People have measured ram-voltage with DMM and it was spot on.

Running the ram out spec is the problem. Micron D9 are 800mhz chips, rated for 1,8v. And ram "labelers" run them at 1066+ with 2,2v+. They're bound to fail in a lot of cases.

And the fact that 800mhz are running fine in most cases, when 1000mhz+ are failing, seems odd that it should be the board killing them off.

...and that's exactly why EVGA tech support told me that my RAM died because of my board and others have had the same problem with the same type of ram. Yep, you got it.

[x800professor]

#0, your title is misleading. The board isnt the cause of ram failure. Not rev AR or A1... Besides, memory vreg hasnt been changed when they made the new revision. So saying the revised board is better is bs. People have measured ram-voltage with DMM and it was spot on.

Running the ram out spec is the problem. Micron D9 are 800mhz chips, rated for 1,8v. And ram "labelers" run them at 1066+ with 2,2v+. They're bound to fail in a lot of cases.

And the fact that 800mhz are running fine in most cases, when 1000mhz+ are failing, seems odd that it should be the board killing them off.

By the way, D9s are 667, not 800. Even running them at 800 is overclocking them, technically. You find me another discussion board where everybody's D9s are failing at 2.2v on different boards and maybe I'll agree with you.

Oh yeah, one of my sticks died at 1.8v.

[dsdsdk]

GIVE ME A REASON. why should the board be frying ram? I've gave you a reason explaining it isnt the board. All you say is "nope it's the board...".

Just because the general public "believes" it's the board, you join right in.
And people with problems post in forum, those without problem dont post that often..

Tons of people are running highend ram just fine. And boards dont come off the productionline different that often. And if it does come off the prod-line different, why would it be the vreg for ram every time?? there's no logic to what you're saying.

"insert system disc".. remove cdrom, remove diskette, remove usb, check IDE/sata cables, jumpers. What harddisk do you have? Raptors and DM10/Maxline have some compatability issues..

X800, so one type of ram has issues with the board, and others dont. How can it be the board?!

[D007]

GIVE ME A REASON. why should the board be frying ram? I've gave you a reason explaining it isnt the board. All you say is "nope it's the board...".

Just because the general public "believes" it's the board, you join right in.
And people with problems post in forum, those without problem dont post that often..

Tons of people are running highend ram just fine. And boards dont come off the productionline different that often. And if it does come off the prod-line different, why would it be the vreg for ram every time?? there's no logic to what you're saying.

"insert system disc".. remove cdrom, remove diskette, remove usb, check IDE/sata cables, jumpers. What harddisk do you have? Raptors and DM10/Maxline have some compatability issues..

X800, so one type of ram has issues with the board, and others dont. How can it be the board?!



ok first off it's the chipset on the board leading to the frying of the ram.. answer me this batman.. was it happening on older boards? no.. and what are you saying here? lol.. I don't see you posting one single technical spec with all this technical talk..

so you go ahead and post voltages and some engineering numbers based on your expirements and ill tell you what i know from life experience.. this was not happeneing on other chipsets..

Never said I was a technician but you sure are trying to sound like one.. so if you are.. show me technical reports.. otherwise your joining the same crew you seem to like to put people in for "jumping on the band wagon" tell you what, when you see the friggin doctor bill i have from my back being wrenched all over leaning into this case trying to work with it's plethora of problems you tell me it's not the chipset.. and I want to see your "proof" it's not.. all I see are words from a guy who likes to talk so far..

bottom line, it's the manufacturers fault for not working with memory makers to build a product that actually worked.. instead they threw out a half assed product they knew would have issues and hoped for the best.. had they waited and solved the problems we wouldn't be having this discussion.. the 680i and vista are in my box of "things i do not like" right now (final).. but i have to deal with this mobo for the time being.. but it's not my fault or problem they didn't do their research.. what do you work for nvidia or something? lol.. your arguments are senseless to me.. i cant even see how you can say it's not this chipset causing problems and it's not the fault fo the makers in general for not fixing this problem before release.. they did EXACTLY what microsoft did with their rushed vista product and they got EXACTLY the same result.. and unstable product.. it's their fault.. idc if it's the board or the chipset or the memory, or santa friggin clause, we should not be the ones dealing with it. they should be.. what do i look like a guennie pig to you?

[dsdsdk]

batman, are you joking.. the chipset frying ram??! so it detaches from the board and attacks the dimms? Or is it the mean ones and zeros the chipset is sending to the ram?

Answer me this, did 1066mhz ram exist "during" older-board-period?
How many people have 1066mhz+ ram in a 965/975 compared to 680i?

If it's the chipset, why hasnt it killed my PC2-8500 1066mhz 5-4-4-12 2,2v yet? Do i have a good chipset? My rev. AR board ran fine with my ram, and so does my new A1....

Why would the board only fry highend ram? seems odd doesnt it? Maybe some ppl are just unlucky and have their lowend ram die.. But the general consensus at evga forum is that the problem is with highend ram.

And no, I dont work with any one.

why do i have to prove anything? you're the ones accusing, you need to back up your claims!

[dsdsdk]

i'm not a tech.. i'm just not easily caught by "hype". But I do know how to apply logic to a situation and I can prolly measure dimm voltage (but that wasnt the problem). And I can tell you that error-data cant cause ram to die. However it can cause reboots and freezes.

[D007]

batman, are you joking.. the chipset frying ram??! so it detaches from the board and attacks the dimms? Or is it the mean ones and zeros the chipset is sending to the ram?

Answer me this, did 1066mhz ram exist "during" older-board-period?
How many people have 1066mhz+ ram in a 965/975 compared to 680i?

If it's the chipset, why hasnt it killed my PC2-8500 1066mhz 5-4-4-12 2,2v yet? Do i have a good chipset? My rev. AR board ran fine with my ram, and so does my new A1....

Why would the board only fry highend ram? seems odd doesnt it? Maybe some ppl are just unlucky and have their lowend ram die.. But the general consensus at evga forum is that the problem is with highend ram.

And no, I dont work with any one.

why do i have to prove anything? you're the ones accusing, you need to back up your claims!

ok anyone can say the words "ones and zeros" memory is alot more complex than that.. memory deals heat based on frequency and voltage.. common knowledge.. if those voltages arent being regulated properly you can expect to fry your ram.. if those frequencies arent being allowed to flow properly i would assume it could cause issues leading to the loss of your ram..

back up my calims lol.. you gonna pay for my board now? otherwise i owe you no explanation what so ever.. it is simply what it is.. it doesnt work right.. like i said idc what it is causing it.. bottom line is they should of fixed it before it came out like this.. it's funny how you can so easliy "think" you must be right over the thousands of claims of issues with people reporting this exact same issue... you like my brother lol.. just can't admit it when your wrong.. defend it all to the bitter end.. me show me im wrong and ill back off and retract.. u saying "this isnt happening" just seems oh forget it lol.. this is pointless.

and still i don't see the technical data backing up your claim.. im done having an E-peen conversation with you.. i don't come here to argue. I come here for solutions, if you want to argue with someone find a child willing to do so. bottom line, the issues exist when they never did before.. coincidence? yea right..lol.. my friggin calims are backed up with me being head first in my system trying to fix it day in and day out.. are you with me at my house at night now? lol.. i think it's best if we just agree to disagree and move on.. fair enough?

I don't mean to come off as rude but if you were dealing with the issues im dealing with, you would be at your wits end as well.. not just memory issues.. lol i wish it was just memory, i could handle that.. massive data corruption, constant reformats, unstable clocks.. incorrect multipliers, the spp mcp is defaulted to critical level" why'?.. theses are facts and you know it as well as i do people are going through it.. im happy for you that your not.. but i am..
I just think it was put together badly.. maybe im wrong, maybe it's drivers.. all i know is.. i aint happy one bit.. just because you want to be right doesn't mean my system will magically start to work properly.. hey if that works please start.... now...lol..

well time to go home from work lol.. yall have a good night :D

[dsdsdk]

Fine, lets agree to disagree..

Regarding your issues. First off, uninstall ntune as it messes with your bios. There has been numerous reports of it causing problems (evga forum). The critical levels are defined in ntune, but they arent critical.

Incorrect multipliers? Are they shown in bios or is it in a program. Cause if it's a program, then the program is at fault. Is your cpu running slower than it should be?

Have you flashed to a newer bios. That should solve the data corruption . Take a SS of the event viewer under Control panel - administrative.

[dsdsdk]

And when installing nforce drivers, make sure you say No to Network Access Manager.

[D007]

only thing i see that i didnt do was say no to the access manager..dang and i thought i was being safe for not installing ntune.. i knew it was causing me issues. it made me crash on several occasions.. seemed like it would not let go of saved settings and would try to override any other settings i had made.. just my take on what it "appeared" to me to be doing. things ran strangely as it got worse and worse then poof.. nada...lol.. well i hope uninstalling the network access manager will help.. thanks for the info by the way..

yea i have the p26 bios but i was seeing some people say you should load bios defaults "before" flashing but i loaded defaults after.. won't hurt i think though, not from what i gather on evgas site.

oh i dont think my cpu is running slower than it should be at all.. seems fine to me anyway.

my multiplier kept defaulting to 7 for some reason.. idk why it would do that. i mean default is 8 right? i thought it was anyway. but it hasnt done that in a bit.. the fact that it had ever done it at all made me nervous though..

i hope removing the network access wont cause any conflicts now that its been installed with the other programs.. you know how they tend to depend on eachother sometimes with certain programs.. like outlook and windows.. i honestly don't know about this one.. just a blind hope thats not the case and i end up with a crap load of more instabilities.. but off i go to delete more nvidia stuff lol.. :D

[D007]

Decided to rewire my whole system after finding a disconnected sata cable.. theres an issue for ya..lol the sata connections have poor grip.. im using metal clip in connections so idk why it would come disconnected.. i rewired the whole system and used the other 12 v connectors for power to my hd's.. not those bulky 12v ones.. idk the terminology again for these cables. like the sata cable of power..lol.. the wide one.. anyway things run a hell of alot smoother now i mean like alot alot.. it wasnt this causing corruption on all the past instances but it could of very well been the reason why it happened the last time. that was after the bios update so idk.. but i ran everything again with as little flex in the cables as possible... it made a hell of an impact so far with stability which was my biggest issue.. lack of stability leading to data corruption and boot downs leading to reformats.. i kept pumping more and more power to it and it would make it more stable for some reason lol.. but maybe it was making up for poor connection.. i also took off my whole cpu heatsink fan and reseated it.. idk man.. i might be very happy if this keeps running the way it is right now.. I'm going to go get rid of network access manager now.. wish me luck lol.. ;)

[newtekie1]

Well my set of G.Skill started to act flakey and giving errors when set to the exact specs they are supposed to run at, so I RMA'd them. Did this board start to kill them and maybe I just caught it before total failure? I think so.

[x800professor]

GIVE ME A REASON. why should the board be frying ram? I've gave you a reason explaining it isnt the board. All you say is "nope it's the board...".

Just because the general public "believes" it's the board, you join right in.
And people with problems post in forum, those without problem dont post that often..

Tons of people are running highend ram just fine. And boards dont come off the productionline different that often. And if it does come off the prod-line different, why would it be the vreg for ram every time?? there's no logic to what you're saying.

"insert system disc".. remove cdrom, remove diskette, remove usb, check IDE/sata cables, jumpers. What harddisk do you have? Raptors and DM10/Maxline have some compatability issues..

X800, so one type of ram has issues with the board, and others dont. How can it be the board?!

Let's see, it gives the correct voltage most of the time but occasionally spikes. You have any voltage measurements cover several days? Oh yes, all of the owners of this board that have watched their ram die at stock speeds have had bad ram! Yes, that's it! There's no connection between the mobo and the dead ram, it's just one big coincidence! People who work at EVGA know that board is crap.

[D007]

Well, just had a physical memory dump while playing age of empires 3.. said " NV4" in the memory dump area.. hmm wonder what nv could mean..lol.. nvidia maybe? i ran memtest to 1000% overnight and had no errors, the system is running like clock work.. max temp 60c under 100% load. everything seems perfect then bam.. blue screen of death.. man.. idk.. i dont even know what to say anymore..lol..

[newtekie1]

GIVE ME A REASON. why should the board be frying ram? I've gave you a reason explaining it isnt the board. All you say is "nope it's the board...".

Just because the general public "believes" it's the board, you join right in.
And people with problems post in forum, those without problem dont post that often..

Tons of people are running highend ram just fine. And boards dont come off the productionline different that often. And if it does come off the prod-line different, why would it be the vreg for ram every time?? there's no logic to what you're saying.

"insert system disc".. remove cdrom, remove diskette, remove usb, check IDE/sata cables, jumpers. What harddisk do you have? Raptors and DM10/Maxline have some compatability issues..

X800, so one type of ram has issues with the board, and others dont. How can it be the board?!

Runnin the RAM out of spec isn't the cause, obviously since I just had RAM start to die on me that was run 100% in spec. There are several reasons the board could be causing this problem.

1.) Irregular voltage, and voltage spikes. Even tiny spikes can cause damage given enough time.

2.) How many volts the motherboard is providing is one thing, but at what amperage is another. If the motherboard is giving 2.0v to a DIMM, but the amperage is double what it should be the DIMM will fail over time from too much power. There is a big difference between pumping 4w through a DIMM and 8w, even though it doesn't seem like it. The DIMM might be able to handle that for a while, but over time it will degrade.

That is just my two theories as to why this board kills RAM.

[newtekie1]

Well, just had a physical memory dump while playing age of empires 3.. said " NV4" in the memory dump area.. hmm wonder what nv could mean..lol.. nvidia maybe? i ran memtest to 1000% overnight and had no errors, the system is running like clock work.. max temp 60c under 100% load. everything seems perfect then bam.. blue screen of death.. man.. idk.. i dont even know what to say anymore..lol..

There is a thread somewhere around here that says the common memory dump errors and what causes them. I believe the NV4 dump is talking about an nVidia file, but is usually caused by the CPU being unstable.

[D007]

how could the cpu bee unstable? low voltage? i cant see how thats possible if it passed the full 3d mark 06 tests various times and ran memtest at full capacity for a whole night.. it's not like it spikes to dangerous temperatures.. ok if anyone sees this link pleas if you don't mind direct me to it, im going to go look for it myself now. fyi i keep getting one beep on start up every single time i start my computer..