EVGA 680i SLI board is killing ram.

[Solarfall]

i have no problems with 2.4v like many of you gues have reported that this is the lethal voltage... guess i'm one of the lucky ones. then again my memory is kind of rare outside Finland.

[Hard H20]

why do you say that

Perhaps because it would end up costing them a TON of money? Dinos if you believe the nVidia & eVGA PR/Cover Our A$$ departments findings, I've some oceanfront property in Afghanistan for sale if youd like a few acres.

I never ran my 8500C5D above 2.2v, ever because of all the problems that had been reported on the eVGA board killing modules. My 2 seperate eVGA 680i boards ate 2 sets in less than 1 month. On a related note my XBX2 has a set of 8500C5D running fine now @ the same voltage (2.2v) for a couple months. I have 2 new 680i boards and 1 new set of 8500C5D and 1 set of 6400C4 to install this weekend. They fail again, I think I will just join theystolemyname and play with legos it would by far be cheaper.

Now hindsights 20/20 if we all had our voltage @ 2.4v and over I would chalk it up to electron migration and call it a day. But, it can't be the 680i that's eating the modules like tictacs. Right!

Let me go find my Nomex coveralls because I know the nVidia fanbois will be along shortly with flamethrowers engaged.

[alpha0ne]

Perhaps because it would end up costing them a TON of money? Dinos if you believe the nVidia & eVGA PR/Cover Our A$$ departments findings, I've some oceanfront property in Afghanistan for sale if youd like a few acres.

Funny that I also have a few acres too that are not needed any more, you want some dinos ???

Hey H20 I want your case

[theteamaqua]

i have no problems with 2.4v like many of you gues have reported that this is the lethal voltage... guess i'm one of the lucky ones. then again my memory is kind of rare outside Finland.

http://www.xtremesystems.org/forums/...&postcount=107

this guys said his 3r board jsut died ... lol no offense .. but u dont sale a mobo until u actually do thorough testing

asus, gigabyte maybe dont have taht customer service compare to EVGA, or vdroop a lot. but at least they tested . so taht there is no need for customer service

honestly i think most of us who bought the 680i AR revision board paid EVGA to be a beta tester

i have never seen any board have such a huge problem by any of the companies

[fhpchris]

I wish I had one of these lucky overvolting boards.

2.5v is weaksauce.

[theteamaqua]

well just to show u how overvolt it was

i set 2.2v in bios for 24/7 use ... i was able to get 5-4-4-10 auto rest @ 1200 with my OCZ 8500 SLI ...

[Crimsonite]

I couldn't agree more, The expert killed three cpu's on me and since then I will not be the first to try any new Tech. The 680i boards on the market today are all reference boards and the top tier board makers are just getting ready to release their designs. The early adopters are the beta testers at their own risk, the manufactures will remain mute and the fan boys will insist that it is always the users fault. Welcome to high tech

That is so true. Samething happened with Honda and their crappy tranny. They denied the failing tranny on all accounts and was later faced a class-action law suit. Not long after that they announced a major recall on 657,000 units of V6 cars for faulty tranny that will fail and come to a complete stop. The best part is they're giving rebuilt-trannies even in brand new cars...lol...

[pumbertot]

That is so true. Samething happened with Honda and their crappy tranny. They denied the failing tranny on all accounts and was later faced a class-action law suit. Not long after that they announced a major recall on 657,000 units of V6 cars for faulty tranny that will fail and come to a complete stop. The best part is they're giving rebuilt-trannies even in brand new cars...lol...

what, they bundled a transexual with their cars? thats just sick dude.

back OT well still running happily @ 2.5V now for past week, im determined to kill this RAM so I can justify the n00bs. lol. nah like some said I just think you get lucky with your DIMMS, some can handle being run well over spec(1.8V) some just cant regardless of what voltage and speed the companies are binning them at.

[dinos22]

well just to show u how overvolt it was

i set 2.2v in bios for 24/7 use ... i was able to get 5-4-4-10 auto rest @ 1200 with my OCZ 8500 SLI ...

common man use a freaking DMM

[sdumper]

Has anybody actual tried to measure the voltage to see what the problem was or are we just assuming because we can hit high speeds that the board must be overvolting?

I'm not being sarcastic I'm just curious.

[TouGe]

Has anybody actual tried to measure the voltage to see what the problem was or are we just assuming because we can hit high speeds that the board must be overvolting?

I'm not being sarcastic I'm just curious.

There was someone over at the EVGA forum that monitored the voltage and found no over-volting or voltage spikes on any of the slots he was using.

[X.T.R.E.M.E_ICE]

Hi. I now have a Gigabyte 680i board rev.2 i think, and i have no dead ram so far (5 weeks)the over voltage problem has been fixed by nvidia now i think.

[Malmedahl]

I was was of those "running fine here" guys , that is until today. What has been reported re: the ram is happening to me also now. Although I don't think the ram has been killed. I think the board ha gone wacked. Like one stick works, and one stick doesn't, then all of a suden the reverse, then both sticks work great, then none, all over the place. Once in windows it seems fine no matter what. But when you cold boot, here go the long beeps and have to re-save bios and try the black slots and on and on. voltage, overclocking etc, don't seem to come that much into play with all this crazy behavior. I think it's something on the board going wacky, but I couldn't begin to guess what exactly. I even tried a bios chip from a BFG board abd it behaved exactly the same way. I wonder if the striker has thi sort of flakeyness being reported?? I might give the asus a try to see if it will settle down at least. Thinng is, (and I was warned about this) all was well for a few weeks then bam! Oh well, it goes with the territory I guess. Expensive territory at that.

Sorry for the OT.


I was just reading trhough this thread when i saw this problem of yours, it matched exactly mine i had on SS and DFI SLI-DR.
My thing was that im using the board "up side down" BTX its maybe called and then the DIMMs end up horizontal underneath the CPU and the evap mount.
There were a very thin little gap venting in air into the evap and socket area, thus creating condensation that stealth sneaked into the bottom of my first DIMM beneath the CPU.

You couldnt see it until after a few days of problems and the copper on the sticks and inside the DIMM´s had tiny small corroded spots wich i saw by coinsidence when using a little LED pocket light.

[Kin Hell]

Considering all the cr4p I had with another platform, my EVGA 680i A2 has been fine untill recently. I thought I gotten over the problems by flashing back to P27 from the P28. P28 let me get an E6600 ES CPU all the way to 3.9GHz & benching with 3DMark06. Most impressed...for about 2 days & then things started falling over. Right now, Im having trouble running a 1GB single stick of Geil PC2 4300 with 1.9Volts up its 4ss! If I put a single stick of OCZ SLI 8500 in the board, I have to use 2.3volts & get random BSOD's revolving around NVNRM.SYS codes.
Personally, I am thinking the mobos Memory Voltage Regulator is shafted or/and it burnt my Ram. The 2GB OCZ 8500 kit never gave C1 Error codes at 1.9volts. No matter how many times I reset CMOS or flash this board back to P27, I get a C1 Error code unless I ramp up the boards VDIMM. Now its got so bad, using the Geil @ 266Mhz intead of the OCZ @ 400, I'm still having to ramp up the boards default voltages but when all this was new, never a C1 error ever!
Just checking the supplier, they are right out of stock!
Ya gotta laugh............

[bench261]

i remember when the thread 1st started and everybody pretty much blamed the guy. but yeah this board is frying ram. my 10th anniversary sticks needed more and more juice as time went buy to boot up. one d9 stick needed 2.3 just to boot at ddr 6400 speeds. problem is, it gets hot and becomes unstable in the super long run.

if i shut down the comp, it would not boot unless i removed one stick, and then i would need to set it at 2.3 volts and reset all my settings, and then i need to reset the comp and lower it back to 2.1 volts. it will operate at 2.1 volts but wont boot up without 2.3 for 1 stick. so i bought a cheap set of d9's and sending my current one to RMA, if it happens again then its definatly the board.

evga said it was the ram that degrades with too much voltage, like 2.4, except i always had it at 2.0

[Leeghoofd]

well you set 2 volts in the bios but how much volts or spikes does it put out ?, my Striker board also overvolts, if I set 2.2 volts I get about 2.3. Okay that ain't much, but my OCZ don't like too much voltage or I get weird erratic behaviour... so I set 2.125 to get 2.2...and have some flow over the dimms that will never hurt...

Too bad Nvidia ever released this crappy chipset upon us ... too many issues...

[dinos22]

i remember when the thread 1st started and everybody pretty much blamed the guy. but yeah this board is frying ram. my 10th anniversary sticks needed more and more juice as time went buy to boot up. one d9 stick needed 2.3 just to boot at ddr 6400 speeds. problem is, it gets hot and becomes unstable in the super long run.

if i shut down the comp, it would not boot unless i removed one stick, and then i would need to set it at 2.3 volts and reset all my settings, and then i need to reset the comp and lower it back to 2.1 volts. it will operate at 2.1 volts but wont boot up without 2.3 for 1 stick. so i bought a cheap set of d9's and sending my current one to RMA, if it happens again then its definatly the board.

evga said it was the ram that degrades with too much voltage, like 2.4, except i always had it at 2.0

nothing has changed buddy

680i does not kill more on average than other chipsets.........the numbers are just not there to cause any concern...........have you got any idea how many 680i chipset motherboards are floating around

and have you ever really read his post.........he's a n00b and fried his RAM because he most likely did not power off the system properly while changing sticks

[Leeghoofd]

Be carefull Dinos if DerekT sees your reply !!!

[bench261]

nothing has changed buddy

680i does not kill more on average than other chipsets.........the numbers are just not there to cause any concern...........have you got any idea how many 680i chipset motherboards are floating around

and have you ever really read his post.........he's a n00b and fried his RAM because he most likely did not power off the system properly while changing sticks

i dunno about that, i never heard much consistency about 965 or 975 boards killing ram, but the evga site has a thread where all sorts of ppl have this issue, and nvidia is blaming the ram. to say this guy is a noob, is calling all those guys noobs, and most of them are running the ram within spec. 3 months after i received an rma'd board for that p25 bricking bios, one d9 ram got "degraded" at 2.0 volts.

believe me, ive read the threads over there including the ones where people dont know whats wrong with the comp, when its really a "degraded" stick. cuz i was blaming the mobo directly for the vid card for having no display and having all sorts of error messages on the bootup screen, when it was on a halfway working ram.

its easy to call ppl noobs or think they're a moron until it happens to u, cuz thats exactly how i used to think. the reference board just had too many problems being rushed out until all the revisions were done, and even then they don't like to acknowledge a problem that makes them liable to something, or finally acknowledges it when a fix is made, or just secretly fixes it. there is certainly an issue though, and it's either the old reference board or d9 ram or the combination of both. i guess i'll find out in 3 months. between the makers of d9 ram and a board that has had fsb walls, x-fi crackling, hd corruption, poor early OC on quad cores, bricking bios, coil buzzing, i wouldn't believe what nvidia says cuz it just screams poor testing or QC

[bench261]

well you set 2 volts in the bios but how much volts or spikes does it put out ?, my Striker board also overvolts, if I set 2.2 volts I get about 2.3. Okay that ain't much, but my OCZ don't like too much voltage or I get weird erratic behaviour... so I set 2.125 to get 2.2...and have some flow over the dimms that will never hurt...

Too bad Nvidia ever released this crappy chipset upon us ... too many issues...

i dont know, cuz it took forever for nvidia just to fix the correct temp readings through bios fixes, so i wouldnt trust what the bios says anyway, and for all i know it could be the minor changes within the updated bios that caused a spike...if it existed. i mean some bios caused cpu temp to go up. is it cuz its actually hotter from more voltage even though its at the "same" setting, or is the temp reading just off? like the ram voltage...who friggin knows

[X.T.R.E.M.E_ICE]

he's a n00b and fried his RAM because he most likely did not power off the system properly while changing sticks

I wouldn't call 8 years of over clocking making me a noob, and i have been playing with computers for 13 years now. EVGA 680i board to this day(13 years) is still the only board that has fried any my ram and i have had a few 680i boards.
Please state all the other boards that are frying ram?Some links would be great? thanks.

[Kin Hell]

I wouldn't call 8 years of over clocking making me a noob, and i have been playing with computers for 13 years now. EVGA 680i board to this day(13 years) is still the only board that has fried any my ram and i have had a few 680i boards.
Please state all the other boards that are frying ram?Some links would be great? thanks.

I can provide verbal proof that a DFI RD600 board fried 1 stick of Geil 4300. There's a link somewhere on XS about some of the fun/bs back then!
Funny thing was, it was RD600 board #4 that killed the ram, just before it killed itself. All the previous boards had C1 Errors out of the box & anyone associated with Support was telling me it was the RAM SPD that was prolly causing the C1 Error. Another funny thing was that RD600 board #4 NEVER had a C1 error untill is killed one of Geil 4300 Sticks. SO who was talking 4rse where?

I changed to EVGA 680i, got to 3.9Ghz (435fsb) on the P28 Bios & managed to run 3DMark06 up there. 2 days later, I was getting SB Audigy2 Pops & bangs across the bus but felt it was the chipset amplifying its pain via the sound card. Since all this happened around 12 weeks ago, it's been a slippery slope since, to the point where I was still getting BSOD's with 1 stick of OCZ 8500 SLI using 2.4v which it is rated at. My supplier told me those kind of volts were required by the Ram! - Not true, as the board did post with 2 x 8500's installed with default of 1.9v @ 800Mhz.
My board was an "A2" according to CPUz & is suppose to be free from being capable of frying Ram. Will be interesting to see what Rev I get back from RMA. As far as I am concerned, I RMA'd one dead board & 2 x 1GB sticks of dead OCZ 8500 SLI RAM. Crikey, the Ram had that "toasty-burnt-out" smell about it.

Since my first Overclock took place on an Amiga 4000/040 back in 1992, it's taken all these years for me to fry anything & the common denominator for me are the Memory Controllers. They just don't appear to be handling the Frequencies/Voltages demanded & the Digital PWM's, now appearing as standard, may also have some influence on failure as well.
I got so pi$$ed of at my supplier saying it's your "PSU" all the time, I Emailed PC P&C in the USA, opened my wallet (whilst bending over) & prised £319.55 out of it to import a 1KW Quad SLI affair with its own Tester.
This PSU has been Daddy to 4x RD600 Boards & 1x 680i board & has lived on a UK 250v Mains loop with it's own dedicated "Electrical Filter". The home I live in was totally re-wired only back in 2004. (Home Cinema & Hi-Fi nut here)

I don't think the boards are being assembled with adequate parts. Inferior & cheaper components lead to failures. 4x RD600 boards & 1x 680i (eventually) kinda says, sh1t loads.

With regard to bench261 saying:

if i shut down the comp, it would not boot unless i removed one stick, and then i would need to set it at 2.3 volts and reset all my settings, and then i need to reset the comp and lower it back to 2.1 volts. it will operate at 2.1 volts but wont boot up without 2.3 for 1 stick.

This is exactly what I saw over the last week prior to RMA.......Shutting it down & then getting a C1 Error just for turning it back on is just pathetic!

Meanwhile, I'm back on my Amiga 4000T - so nice to be "Bill-Free" again!

[Leeghoofd]

Please state all the other boards that are frying ram?Some links would be great? thanks.

One users claims his P5K dlx is a ram eater too becasue 3 peeps had death rams this week...

Speederlander already replied to my response saying probably heat killed one of his 2gb stick, also the other user ran his Gskills at over 900mhz at 4-4-4-12( in relation to his earlier posted SuperPI times ) so he must have used also more than 2.1 stock volts to get there, in case you overclock things can break

It's a risk we are all glad to take otherwise run everything at stock, if it kills ram then then you can call it a ram eater !!! Heat and electrical components don't go together over 2.2 volts I would recommend a fan over the ram...


On a second note it's nice to have ram at 1200mhz but I wouldn't run those settings 24/7...some users are a bit blinded by some posts of our magic overclocking friends who turn anything into gold... don't try it at home folks unless you are perfectly aware what you are doing

[Brother Esau]

I can provide verbal proof that a DFI RD600 board fried 1 stick of Geil 4300. There's a link somewhere on XS about some of the fun/bs back then!
Funny thing was, it was RD600 board #4 that killed the ram, just before it killed itself. All the previous boards had C1 Errors out of the box & anyone associated with Support was telling me it was the RAM SPD that was prolly causing the C1 Error. Another funny thing was that RD600 board #4 NEVER had a C1 error untill is killed one of Geil 4300 Sticks. SO who was talking 4rse where?

I changed to EVGA 680i, got to 3.9Ghz (435fsb) on the P28 Bios & managed to run 3DMark06 up there. 2 days later, I was getting SB Audigy2 Pops & bangs across the bus but felt it was the chipset amplifying its pain via the sound card. Since all this happened around 12 weeks ago, it's been a slippery slope since, to the point where I was still getting BSOD's with 1 stick of OCZ 8500 SLI using 2.4v which it is rated at. My supplier told me those kind of volts were required by the Ram! - Not true, as the board did post with 2 x 8500's installed with default of 1.9v @ 800Mhz.
My board was an "A2" according to CPUz & is suppose to be free from being capable of frying Ram. Will be interesting to see what Rev I get back from RMA. As far as I am concerned, I RMA'd one dead board & 2 x 1GB sticks of dead OCZ 8500 SLI RAM. Crikey, the Ram had that "toasty-burnt-out" smell about it.

Since my first Overclock took place on an Amiga 4000/040 back in 1992, it's taken all these years for me to fry anything & the common denominator for me are the Memory Controllers. They just don't appear to be handling the Frequencies/Voltages demanded & the Digital PWM's, now appearing as standard, may also have some influence on failure as well.
I got so pi$$ed of at my supplier saying it's your "PSU" all the time, I Emailed PC P&C in the USA, opened my wallet (whilst bending over) & prised £319.55 out of it to import a 1KW Quad SLI affair with its own Tester.
This PSU has been Daddy to 4x RD600 Boards & 1x 680i board & has lived on a UK 250v Mains loop with it's own dedicated "Electrical Filter". The home I live in was totally re-wired only back in 2004. (Home Cinema & Hi-Fi nut here)

I don't think the boards are being assembled with adequate parts. Inferior & cheaper components lead to failures. 4x RD600 boards & 1x 680i (eventually) kinda says, sh1t loads.

With regard to bench261 saying:



This is exactly what I saw over the last week prior to RMA.......Shutting it down & then getting a C1 Error just for turning it back on is just pathetic!

Meanwhile, I'm back on my Amiga 4000T - so nice to be "Bill-Free" again!

My observation is also that the Memory controller is the culprit, reason being I can run 3325mhz @475FSB 8 hrs Prime no sweat but the memory controller I am thinking cannot maintain at those high frequencies because I get really bad Lan NIC performance even though I am on a dedicated Server PCI-E NIC but when I scale back down to 450FSB NO MORE ISSUES!

We need to petition EVGA to get D.F.I to start making their Motherboards!

[Kin Hell]

My observation is also that the Memory controller is the culprit, reason being I can run 3325mhz @475FSB 8 hrs Prime no sweat but the memory controller I am thinking cannot maintain at those high frequencies because I get really bad Lan NIC performance even though I am on a dedicated Server PCI-E NIC but when I scale back down to 450FSB NO MORE ISSUES!
We need to petition EVGA to get D.F.I to start making their Motherboards!

nVidia chose to use the same Memory controller for 680i as they did on the nForce4 Chipset which controlled DDR ram. I'm convinced it's this area causing the burnouts. I also recall a similar issue back on the Rambus days. The memory controller could be upgraded to achieve better overclocks without burn-outs. The replacement controllers were of a better grade but cost more to purchase. It might be worthwhile checking out the assembly plants in Ninja land; - Cr4p parts help to keep costs down.

With regard to getting DFI to make boards, don't they already do a 680i LT board? Going by my past experience with DFI, if your petition was started, Im afraid I'd be a dead man, if my life depended on signing it!