OMG I'm so sick of this 680i crapbox

D

drpeppir

Limp Gawd
Joined
Mar 21, 2004
Messages
165
Ok been running this system in my sig for a while, and then all of a sudden when I play a media file it hardlocks for no reason with the speakers making a loud humming noise! This is getting ridiculous, does nvidia not care about the quality of their chipsets -- I cannot even fucking upgrade to 780i b/c stability issues are reported on that POS as well!

In addition to that going to X38 will cost ~1300-1500 since I water cool all parts on my systems and have 3 displays I would need 2x 3870X2's as well.

I guess I am stuck with this shitbox for a couple more months (8800 GTX's aside)
 
drpeppir said:
does nvidia not care about the quality of their chipsets
Click to expand...

No.

They know they can sell their chipsets because of SLI. They make a lot of money off convincing people that they are gonna get a huge boost from SLI and then forcing them to buy their shitty chipsets to do it.

Wish Intel would work something out where they an stick the Nvidia SLI chip or whatever it is on their mainstream boards not just Skulltrail.
 
NKDietrich said:
Wish Intel would work something out where they an stick the Nvidia SLI chip or whatever it is on their mainstream boards not just Skulltrail.
Click to expand...

Isn't that what Intel is trying to do by not giving Nvidia access to it's Common Serial Interface (CSI) license? Without that, Nvidia wouldn't be able to create and sell core-logic sets compatible with Intel's processors (Nehalem and beyond). :D
 
KrylonArt said:
Isn't that what Intel is trying to do by not giving Nvidia access to it's Common Serial Interface (CSI) license? Without that, Nvidia wouldn't be able to create and sell core-logic sets compatible with Intel's processors (Nehalem and beyond). :D
Click to expand...

Possibly. I'm waiting to see how all those politics boil down. Should be interesting.
 
NKDietrich said:
No.

They know they can sell their chipsets because of SLI. They make a lot of money off convincing people that they are gonna get a huge boost from SLI and then forcing them to buy their shitty chipsets to do it.

Wish Intel would work something out where they an stick the Nvidia SLI chip or whatever it is on their mainstream boards not just Skulltrail.
Click to expand...

the compiler is on the card itself. the limitation is software only.
 
Oh common MAN TEH 680i SLI ROX!!!!!!!!!!

Yeah they pretty much suck. That's why I have some dead/dying ones hanging on my wall. I've been through so many of those pieces of shit at this point.
 
drpeppir said:
Ok been running this system in my sig for a while, and then all of a sudden when I play a media file it hardlocks for no reason with the speakers making a loud humming noise! This is getting ridiculous, does nvidia not care about the quality of their chipsets -- I cannot even fucking upgrade to 780i b/c stability issues are reported on that POS as well!

In addition to that going to X38 will cost ~1300-1500 since I water cool all parts on my systems and have 3 displays I would need 2x 3870X2's as well.

I guess I am stuck with this shitbox for a couple more months (8800 GTX's aside)
Click to expand...

with your overclock, it was practically inevitable. i run a q6600 on my 680 board at 333fsb only and lock the ram to 800mhz.
 
rennyf77 said:
the compiler is on the card itself. the limitation is software only.
Click to expand...

Which is why 7800/7900 series cards could have their drivers hacked and have SLi enabled on Intel boards. I remember people running SLi on 975 boards this way, but the last I heard Nvidia put a lock on the drivers so it made it harder to run SLi on non Nvidia boards.
 
rennyf77 said:
with your overclock, it was practically inevitable. i run a q6600 on my 680 board at 333fsb only and lock the ram to 800mhz.
Click to expand...

You don't have to lock the RAM to 800MHz. The 680i SLI reference boards are pretty good about memory speeds up to DDR2 1200MHz. The non-reference boards on the other hand are not so good about that.
 
Dan_D said:
You don't have to lock the RAM to 800MHz. The 680i SLI reference boards are pretty good about memory speeds up to DDR2 1200MHz. The non-reference boards on the other hand are not so good about that.
Click to expand...

and all this time i was nerfing my dominators for no reason.
 
Dan_D said:
Oh common MAN TEH 680i SLI ROX!!!!!!!!!!

Yeah they pretty much suck. That's why I have some dead/dying ones hanging on my wall. I've been through so many of those pieces of shit at this point.
Click to expand...

Dan_D. Why do I always agree with you? ;)

But yeah they really suck. The chipset overheats even when you got a waterblock, it still dies.
 
This does not make me happy since I have 2 of these boards and well both of them are crapping out on an overclocked q6600 and a stock e6600. Dam 680i

Time for an Abit IP35-e
 
LOCO LAPTOP said:
Dan_D. Why do I always agree with you? ;)

But yeah they really suck. The chipset overheats even with you got a waterblock, it still dies.
Click to expand...

I've fried 3 water cooled, one of which was an EVGA 680i SLI Black Pearl with pre-installed water blocks.
 
Mine is ok......I guess it was made on Wednesday.:D

Sorry about yours.
 
Shut up!!!! I got one, and it never gave any trouble (yet).

Seriously, reading this thread makes me super nervous, because anytime something goes wrong, I keep thinking, it happened to me too.

Fortunately everytime it is just some app, or driver, or cut cable or something.

I really hope this is just beginner's luck (my 1st build), and I will be able to use it till it becomes obsolete.

I have also signed up to upgrade it to 780i on the last day, in case things start go on it.

Oh well, like they say, if it ain't broke...
 
I'm on my 2nd one right now, approaching the 6 week mark, so i should know if its going to die or not soon. I run a q6600 with sync'd memory at 400FSB. Hopefully though, this one will hold out, it seems to actually have fairly decent NB temps.
 
veedubfreak said:
I'm on my 2nd one right now, approaching the 6 week mark, so i should know if its going to die or not soon. I run a q6600 with sync'd memory at 400FSB. Hopefully though, this one will hold out, it seems to actually have fairly decent NB temps.
Click to expand...

What are your exact north bridge temps? Just curious.
 
SpeedyVV said:
Shut up!!!! I got one, and it never gave any trouble (yet).

Seriously, reading this thread makes me super nervous, because anytime something goes wrong, I keep thinking, it happened to me too.

Fortunately everytime it is just some app, or driver, or cut cable or something.

I really hope this is just beginner's luck (my 1st build), and I will be able to use it till it becomes obsolete.

I have also signed up to upgrade it to 780i on the last day, in case things start go on it.

Oh well, like they say, if it ain't broke...
Click to expand...

I couldnt agree more.
I added triple-SLi and the day I did, I got the nvidia power sentinel pop-up, I knew the PSU was stout enough........I thought, well here goes my board.......the prophesy has been fulfilled.........ended up being a bad connector..........so far so good.
Only thing I did do, was add active NB cooling from the first day I got mine.
 
Fuck 680i. I hated my old board. It was an LT even, which supposedly was more reliable than the SLI variant. It ran fine @ stock, but pushing FSB was a pain in the ass.

Dumped for P35 with much haste.
 
I could have told you not to buy an NVIDIA chipset. The only chipsets you ever buy to go with Intel CPUs are Intel chipsets. They just work. It's like syncing an iPhone on a Mac with the Mac versions of iTunes, Apple Address Book, and Apple iCal. You don't ever see those people complaining; all the complaints come from the people using Outlook and the Windows port of iTunes. No one knows Intel CPUs better than Intel, and no one's chipset will work better with an Intel CPU than Intel's.
 
InorganicMatter said:
I could have told you not to buy an NVIDIA chipset. The only chipsets you ever buy to go with Intel CPUs are Intel chipsets. They just work. It's like syncing an iPhone on a Mac with the Mac versions of iTunes, Apple Address Book, and Apple iCal. You don't ever see those people complaining; all the complaints come from the people using Outlook and the Windows port of iTunes. No one knows Intel CPUs better than Intel, and no one's chipset will work better with an Intel CPU than Intel's.
Click to expand...

yeah he hit it on spot.

and looking at his sig, i think you should listen to him.
 
Yeaa I admit this was the first nvidia chipset ive ever bought; i have always stuck with intel but in 06 when this system was purchased sli looked too promising-- so i made a comprise... it backfired-- i will have to stick with ati/intel from now on for graphics and cpu respectively hopefully with what intel is boasting it will be intel/intel but who knows what sort of consequences that may bring!
 
I'm just going to lay out a few ideas:

1.) Not that I'm outright questioning anyone's 'skillz' here, but anyone who attempts to overclock a board housed around the 680i SLI (which all still cost roughly $200 each), fries it, and then attempts to buy another board expecting different results multiple times would normally indicate a user error. I only say this because, well, I only see about a dozen people in this thread who seem to keep on frying boards, which is very much unlike the other threads dedicated to 650/680 boards that contain dozens of pages of users who seem to be having plenty of success and are happy with their boards (not that I'm ruling out blind conformity, which is possible but rather unlikely...)

2.) Last I checked, pumping too much voltage into the 6x0i series was counter-productive to overclocking the FSB (which was why everyone was frying their RAM and the whole board basically bursts into flames). Both 650i boards (Asus & Gigabyte) I have used overclock effortlessly on stock voltages, except CPU, to at least 425 MHz with my E6400 before I have to drop the LDT frequency before I top out at roughly 475-480 MHz FSB (it seems to be the CPU's limit).

3.) While I know everyone here is all about trying to go the distance in the never-ending pissing contest of who's-got-the-fastest-pee-cee, not every piece of silicon can break a record and hold it for months-to-years-on-end. Heck, even [H] only tests their overclocks for stability for only 24 hours tops; when the hell does a day of overclocking become a testament to durability over time? Seriously, when was the last time you heard of someone who tested a maximum overclock and held it for, say, six months? "Why would they do that, by then a faster board will be out and the lump of silicon I have will be outdated," you might say. Well, fine then, you'll have to live with the fact you're taking the risk of buying a board that might not last six months being overclocked all the time.

4.) My P965 board died after 8 months at a mild 350 MHz FSB for no visibly discernible reason, while the 865PE board I bought almost 4 years ago (with it's equally unamazing overclock of 233 FSB on a Northwood P4) is still running strong, all the while an NForce 4 SLI AMD board I passed along to a friend of mine is still working after nearly 3 years and is currently holding a 250 FSB overclock. Owning an Intel chipset doesn't guarantee durability, overclocking, or quality. Thought you ought to know.
 
well then i guess its safe to say that 680i Chipsets suck at overclocking with higher voltages.

i mean my P35 can take 1.5 Vcore and 1.4 Northbridge Voltage, along with 2.25 Volts on the RAM and the damn thing doesn't even get that hot.

Intel chipsets are good for overclocking (excluding the 915/925 series), as you said with your 865 chipset
 
yojimbo said:
I'm just going to lay out a few ideas:

1.) Not that I'm outright questioning anyone's 'skillz' here, but anyone who attempts to overclock a board housed around the 680i SLI (which all still cost roughly $200 each), fries it, and then attempts to buy another board expecting different results multiple times would normally indicate a user error. I only say this because, well, I only see about a dozen people in this thread who seem to keep on frying boards, which is very much unlike the other threads dedicated to 650/680 boards that contain dozens of pages of users who seem to be having plenty of success and are happy with their boards (not that I'm ruling out blind conformity, which is possible but rather unlikely...)
Click to expand...

I've only had one memory module fry because of a 680i SLI board and that's because the north bridge burnt up. BTW I only bought two 680i SLI boards one of which I'm still using. The rest were either returned to the retailer or RMA'ed through the manufacturer. One EVGA 680i SLI board I got for free and it still fried. Three of them were water cooled (two regular A1's and one EVGA Black Pearl) and I always kept my voltages to a minimum. As for the people that are happy with their boards I guarantee you most of them are using Core 2 Duo processors and not Core 2 Quad's or Core 2 Extreme quad core CPUs. Those that run quad core processors will likely have boards that die faster.

yojimbo said:
2.) Last I checked, pumping too much voltage into the 6x0i series was counter-productive to overclocking the FSB (which was why everyone was frying their RAM and the whole board basically bursts into flames). Both 650i boards (Asus & Gigabyte) I have used overclock effortlessly on stock voltages, except CPU, to at least 425 MHz with my E6400 before I have to drop the LDT frequency before I top out at roughly 475-480 MHz FSB (it seems to be the CPU's limit).
Click to expand...

Again the problem with the 680i SLI reference boards has to do with their shitty voltage regulation and poor voltage component selection. Also the BIOS seems to overvolt anything left on auto at times. Non-reference boards suck ass because the manufacturers fuck the BIOS up by altering the way they set tmings on the memory modules. They virtually ignore the SPD values of the board and set the RAM to certain settings just a slight bit tighter than the SPD values. This was confirmed by OCZ on their website. They had to release their own version of the Striker Extreme and P5N32-E SLI BIOS ROMs because ASUS fucked them up so bad. So while Gigabyte and ASUS designs are better in some ways than the reference boards they still suck ass in a lot of ways. Memory compatibility is a big part of why. The reason why these manufacturers have horrid memory compatibility is because their memory tweaks are designed to improve performance over the non-reference design. They did this so they could win benchmarks more than likely. Unfortunately it comes at the cost of memory compatibility problems. Companies like OCZ have had to deal with a ton of headaches because of this type of tweaking. As I already said they had to release their own unfucked version of the BIOS because of this. Another problem squarely rests on NVIDIA's shoulders for fucking their own reference BIOS versions up after version P28 by forcing Spread Spectrum settings to be disabled in order to improve their poor quad core overclocking ability even on A1 revision boards. This BTW caused me no end of grief as it killed my system because I was using a RAID controller in the middle slot which my had the array my OS was installed to. Because of this I got data corruption and finally the system wouldn't even detect the RAID controller as a result of this crap. I ended up having to post a thread on this forum asking for a previous version of the BIOS to get my machine working again.

Good job NVIDIA. BTW newer BIOS versions have NOT addressed this issue either. The RAID contoller issues still plague boards flashed with BIOS P29 or higher TODAY. This problem began 4 months ago or longer. Totally unacceptable in my book.

yojimbo said:
3.) While I know everyone here is all about trying to go the distance in the never-ending pissing contest of who's-got-the-fastest-pee-cee, not every piece of silicon can break a record and hold it for months-to-years-on-end. Heck, even [H] only tests their overclocks for stability for only 24 hours tops; when the hell does a day of overclocking become a testament to durability over time? Seriously, when was the last time you heard of someone who tested a maximum overclock and held it for, say, six months? "Why would they do that, by then a faster board will be out and the lump of silicon I have will be outdated," you might say. Well, fine then, you'll have to live with the fact you're taking the risk of buying a board that might not last six months being overclocked all the time.
Click to expand...

Smart people overclock smarter than that. I've had hardware that has lasted years overclocked. The key is to test for maximum stability under the most varied conditions. Its' up to the owner to take the risk and make sure they do it right. If you don't you pay the consequences. In any case if you think overclocking 680i SLI reference boards kills them you are mistaken. I've run them at stock and overclocked and I've seen deaths of those things in three weeks time. You simply can not rely on most 680i SLI reference boards. Fortunately after 11 of them I know how to tell the good ones from the bad ones and reasonably predict their longevity.

yojimbo said:
4.) My P965 board died after 8 months at a mild 350 MHz FSB for no visibly discernible reason, while the 965P board I bought almost 4 years ago (with it's equally unamazing overclock of 233 FSB on a Northwood P4) is still running strong, all the while an NForce 4 SLI AMD board I passed along to a friend of mine is still working after nearly 3 years. Not even owning an Intel chipset guarantees durability, overclocking, or quality. Thought you ought to know.
Click to expand...

There is some random chance to it but if you really know what you are doing you can be reasonbly certain that the hardware will last. The key is to make sure that you don't push the hardware too far.
 
rennyf77 said:
just speculation, but i'm willing to bet nvidia only made that chip to force intel to buy something from them.
Click to expand...

licensing dude!

think of the NF100's as dummy hardware locks, this would enable SLI without making modifications to the nvidia drivers since the drivers would see an nvidia chipset.
 
After 3 dead boards, moved to p35+x38 and never looked back. Only reason why i have a nib rma evga A1 still sitting on the shelf.
 
Dan all I gotta say is HardOCP must value your opinion, cuz u def know whats up!
 
drpeppir said:
Dan all I gotta say is HardOCP must value your opinion, cuz u def know whats up!
Click to expand...

There is some information I forgot to include, I've updated that earlier long post to include the information I didn't include before.
 
Wow, how unfortunate for some...
Ive had my 680i board for about 9months now and its been chugging along nicely.
I have a 6420 crusiing along at 3.0GHz with all voltage set to stock levels w/ air cooling.

The only real issue taht I have found with the board itself is that the SPD detection sucksballs and/or doesnt work at all for my ram (OCZ 1066). Becuase of this, If I ever need to clear the CMOS, I have to stick a slow 667Mhz module in by itself, set the BIOS up for my OCZ ram, power it off and then put my OCZ back in. ;)

My first experience with an NVIDIA chipset some years ago (P5N32 or someshit) was horrible, couldnt get that pos to work right, and ended up goign with an Intel chipset. But I figuerd Id give NVIDIA another shot, and so far im happy taht I did.
 
bosorio said:
Wow, how unfortunate for some...
Ive had my 680i board for about 9months now and its been chugging along nicely.
I have a 6420 crusiing along at 3.0GHz with all voltage set to stock levels w/ air cooling.

The only real issue taht I have found with the board itself is that the SPD detection sucksballs and/or doesnt work at all for my ram (OCZ 1066). Becuase of this, If I ever need to clear the CMOS, I have to stick a slow 667Mhz module in by itself, set the BIOS up for my OCZ ram, power it off and then put my OCZ back in. ;)

My first experience with an NVIDIA chipset some years ago (P5N32 or someshit) was horrible, couldnt get that pos to work right, and ended up goign with an Intel chipset. But I figuerd Id give NVIDIA another shot, and so far im happy taht I did.
Click to expand...

Again the 680i SLI reference and non-reference boards aren't always horrible. I have a EVGA 680i SLI (122-CK-NF68-AR) board that is just over a year old now and it has never crashed, locked up or blue screened on me running the same instance of Windows XP Professional SP2 since it was built. That machine gets used a ton as well. My other board is a Striker Extreme version 1.00G with BIOS 1303. Though it overclocks for shit and can't go past 800MHz reliabily it never locks up and works like a champ with my Q6600. That combination has been turned into my Windows 2008 server and it has worked very well so far. I've had that board for about 14 months give or take and again though I don't like it due to poor overclocking, it is reliable and it has an awesome feature set. At stock speeds it has absolutely no quirks.

Quad cores, certain types of memory, and of course months of overclocking to hell and back can and usually does seem to take its' toll on these things. Typically I've found that dual core processor based machines last longer than 680i SLI boards paired up with quad core CPUs. I've also found that if you have NB temperatures regularly exceeding 50 or 55c watch out. You've got a time bomb.
 
i am another person with an nvidia chipset who consistently gets nervous reading these threads. I bought mine about 4 months ago, have an e6750 up to 3.5 on it with not so much as a hiccup.

I think I may just be one of the lucky ones. But i really like this evga a1 board.
 
Ill tell you what the 650i chipset isn't much better i dumped my DFI 650i ultra for a MSI P35 - FR and this sucker is nice. Cool NB temps and all with an E6550 @ 2.8GHz with crappy artic silver cooler it runs great very stable. one thing is I can tell the age of my 7900gs because no matter what pc it is in and what driver you use it still blue screens the pc when it goes into hibernation. But 9600gt looks promising for what I get it for.
 
Manaknight said:
i am another person with an nvidia chipset who consistently gets nervous reading these threads. I bought mine about 4 months ago, have an e6750 up to 3.5 on it with not so much as a hiccup.

I think I may just be one of the lucky ones. But i really like this evga a1 board.
Click to expand...

And if your luck holds there is no reason to change it out for anything else. Just be aware that many of them do have problems and having to replace it down the line (sooner than you want to) is a real possibility.
 
After having 8 (isn't that what you wrote??) of these boards fail on him, Dan_D certainly has a right to come out at recommend against this product. Maybe he was pushing them harder than normal. Maybe there were other circumstances coming into play in his situation that affected the outcomes of his overclocking and board reliability. Whatever the reason, 8 out of 11 failures (approx) isn't that good of a track record.

At the same time, my non-reference board overclocks ram quite well, and I have a decent overclock on my quad core, too. All my temps are very good, and my system is exceptionally stable. Lockups are few and far between. (My HTPC is starting to freeze up on me occasionally, don't know what the heck... ) Anyways, my experience with my abit 680i board has been very good. But at first I was limited by the high PWM temps like many others. With paltry overclocks and barely higher than stock voltages my PWM temps were hitting the high 130c range in Prime95 torture tests with small FFT's!!! At 125c the thing starts the shut down process of your computer so you don't fry something. That scared the living crap out of me!! I did a simple heat pipe/heat sink mod with a little arctic silver and folding the thermal pads in half, and now my temps are very respectable on PWMs, my CPU, and everything else I run. I love all the expansion ports, USB and firewire headers, and the e-sata ports on the back are a really nice touch!

Like any electronics produce, a certain number out of every 10 produced are going to have problems. The 680i's overall, sadly reported a higher per capita failure rate than many other chipsets. Of course you gotta take into account that people who have problems are way more likely to post and complain than people who are happy. Why do I need to hop on forums if everything is working just fine?? (one might ask). Still, it seems more people report problems vs. the number that report successes.

I really hate to write negatively at all about a product I own and love, and that has been fantastic for me thus far, and I hate to agree with Dan_D even more!! :p jk! But I can't deny the problems that are frequently reported on all the forums I visit.

(see sig for rig deails:)
Q6700 3.33MHz
8GB ram @ 1000MHz with 1.95 volts
 
SirKronan said:
After having 8 (isn't that what you wrote??) of these boards fail on him, Dan_D certainly has a right to come out at recommend against this product. Maybe he was pushing them harder than normal. Maybe there were other circumstances coming into play in his situation that affected the outcomes of his overclocking and board reliability. Whatever the reason, 8 out of 11 failures (approx) isn't that good of a track record.

At the same time, my non-reference board overclocks ram quite well, and I have a decent overclock on my quad core, too. All my temps are very good, and my system is exceptionally stable. Lockups are few and far between. (My HTPC is starting to freeze up on me occasionally, don't know what the heck... ) Anyways, my experience with my abit 680i board has been very good. But at first I was limited by the high PWM temps like many others. With paltry overclocks and barely higher than stock voltages my PWM temps were hitting the high 130c range in Prime95 torture tests with small FFT's!!! At 125c the thing starts the shut down process of your computer so you don't fry something. That scared the living crap out of me!! I did a simple heat pipe/heat sink mod with a little arctic silver and folding the thermal pads in half, and now my temps are very respectable on PWMs, my CPU, and everything else I run. I love all the expansion ports, USB and firewire headers, and the e-sata ports on the back are a really nice touch!

Like any electronics produce, a certain number out of every 10 produced are going to have problems. The 680i's overall, sadly reported a higher per capita failure rate than many other chipsets. Of course you gotta take into account that people who have problems are way more likely to post and complain than people who are happy. Why do I need to hop on forums if everything is working just fine?? (one might ask). Still, it seems more people report problems vs. the number that report successes.

I really hate to write negatively at all about a product I own and love, and that has been fantastic for me thus far, and I hate to agree with Dan_D even more!! :p jk! But I can't deny the problems that are frequently reported on all the forums I visit.

(see sig for rig deails:)
Q6700 3.33MHz
8GB ram @ 1000MHz with 1.95 volts
Click to expand...

If yours is working for you, then great. On a personal note I didn't push mine any harder than anyone else I don't think. I've had these things die with no overclock, minor overclocks and major overclocks. I really started having problems with them dying after I got a Q6600. After I started overclocking the quad core that's when things really got tough. The board I have with a dual core E6600 is still running like a champ.
 
You must log in or register to reply here.
Back
Top