NF4 Cold Boot Thread

[Forza971sbh]

So you said.

 

I've used it for over 3 months now and have not had a problem with mine though.

 

In fact it is only people with OCZ that seem to be complaining. Now I realize there are more people here with OCZ than other brands, but that does seem awefully coincidental.

 

yes it is!

 

had to rma my ballistix too ....

 

problems began the day I put them on my sli-dr... and it had been running for more than one year on a K8T, K8n, K8n neo2.... without any problems...

[Boppo]

yes it is!

 

had to rma my ballistix too ....

 

problems began the day I put them on my sli-dr... and it had been running for more than one year on a K8T, K8n, K8n neo2.... without any problems...

So you put them on the 5v rail and it burned out your RAM immediately?

[Forza971sbh]

So you put them on the 5v rail and it burned out your RAM immediately?

 

I can't tell you if it took 1 day or one week to kill it... but I noticed instability really soon after the 1st power-up of the rig'

 

that's why I "memtested" it and noticed that it produced loads of errors.... spent 3 month of tweaking to finally realize that it couldnt handle DDR200 (100MHZ!!!) @ awfully relaxed timings: 3-4-4-10

 

note that I've never "overvolted" it by the bios: 2.9v is the maximum voltage I ever used (that ram is rated 2.8v)

[Boppo]

Ah, that would be a different thing then I think. I believe the accusation here is that when using the 4v jumper to allow more than 3.2v to the RAM that the power is unstable and thus burns up the RAM.

 

Whether this is true or not I do not know. I do know that I see all kinds of accusations here though that are far more often than not just based on rumors and innuendos.

[Psylenced]

Itd funny you mention it, I enabled my 5v rail for the DIMM slots before I even installed my mobo for the first time, knowing that VX 4000 gold was going in. now the OCZ website and the sticker on the dimms connotates that it will run 2-2-2-8 at 3.3v. OCZ warranties their RAM up to 3.5v. I dont think they would warranty their ram at 3.5v if it had a high possibility of burning out at these voltages.

 

My ram has never seen voltages above 3.3v, because I could initially get it to run at what I thought was stable at 2-2-2-6 1T.

 

However, upon installing the OS on the brand new hardware (this was a completely new virgin rig, incl. the HDD) I startednoticing issues in games. Dungeon Siege 2 would intermittently lock up. BF 2 as well. so I loosened the timings and it went away for a while.

 

Then about 5 days ago (perhaps slightly longer) I was playing DS 2, and had an error so bad that it corrupted my save game files and DS 2 stopped running entirely because of corruption to the game files, the error mentioned something about CRC failure and my filesystem. I was furious, I was more than halfway through the game (and this is a rather long game mind you). So I reduced all my settings to optimized defaults, what im running now, with minor changes to accomodate my SLI system and disable onboard audio/marvell lan, firewire. Now I am still locking up in DS 2 intermittently.

 

The point is, is that if OCZ warranties their ram up to 3.5v, and I never volted it over 3.3v, imagine how bad the power spikes to the DIMM slots must be. At first I thought it was a cold boot issue, because my symtoms were very similar with the exception that one of the dimms would boot the system, while the other wouldnt, and eventually it got so bad that I cannot even boot the system anymore with both DIMM's in. I get the single long beep denoting a memory error. If I have just the one problematic DIMM in and I try to boot I get the one long memory error and (if I recall correctly, although I was so frustrated at this point I may be wrong) 3 LED's.

 

My system ran stable for a month with the settings in my sig according to 8 hours of prime and 24 hours of memtest looping test 5. Yet I still locked up intermittently DS 2 and BF 2. Then it suddenly all went in the crapper.

 

DFI is obviously aware of these problems, because its Expert series motherboard promotes revised power delivery features. Question is, why would they release a new revision of one of their most popular motherboards and not address the large community using the faulty ones. This is where my problem lies. If DFI had come out and said ok guys we screwed up, send us the faulty mobo's in for an "Expert" series exchange, or even a revised and fixed edition of my corrent board, I would not be angry in the least. Possibly a little annoyed but hey . happens. My anger is directed at the fact that they are obviously trying to sweep the issue under the rug and then remedy it with another solution, yet charge us 260 bucks for it.

 

If I had known this board was so unstable when I bought it, I wouldnt have. I was expecting some issues from the onset, with this kind of hardware you should, but I certainly wasnt expecting issues that could not be addressed via a bios patch, and instead I am stuck with a faulty piece of hardware that cannot be returned because it is running within its "specifications".

[Forza971sbh]

DFI is obviously aware of these problems, because its Expert series motherboard promotes revised power delivery features. Question is, why would they release a new revision of one of their most popular motherboards and not address the large community using the faulty ones. This is where my problem lies. If DFI had come out and said ok guys we screwed up, send us the faulty mobo's in for an "Expert" series exchange, or even a revised and fixed edition of my corrent board, I would not be angry in the least. Possibly a little annoyed but hey . happens. My anger is directed at the fact that they are obviously trying to sweep the issue under the rug and then remedy it with another solution, yet charge us 260 bucks for it.

 

If I had known this board was so unstable when I bought it, I wouldnt have. I was expecting some issues from the onset, with this kind of hardware you should, but I certainly wasnt expecting issues that could not be addressed via a bios patch, and instead I am stuck with a faulty piece of hardware that cannot be returned because it is running within its "specifications".

 

totally agree!

 

but you can be sure of one thing: if this "expert" board really hits europe, I'm pretty sure the sales manager will have to "deal with" hundreds of anger letters from unsatisfied users remembering him how clever it is to tell your customers "we hear you", "we're faulty" and "we will find a solution"

[Psylenced]

Well looking back on the whole thing, it sounds just a little silly dont you think? This is one of the most powerful overclocking boardson the market, yet it doesnt support 1 of 2 of the most powerful RAM vendors on the market (Corsair), and now issues are arising with one of OCZ's most popular and powerful RAM series? At this rate I might as well have bought a Dell system, since this motherboard is dictating my cookiecutter system specs anyhow, its like having a prebuilt system from a major PC vendor. Arent there ISO or ANSI specs that are supposed to stop this kind of incompatibility on computer hardware, if so is DFI just not following them? Or are they trying to create a new standard at the expense of their customer base.

[Psylenced]

Oh it should also be noted that to correct the power fluctuation issues, the new "Expert" series SLI-DR requires an 8 pin EPS connector on your PSU.

 

So in addition to shelling out 260 (current retail price) for this mobo, expect to shell out another 200 for a server end PSU that actually has an EPS connector like the OCZ PowerStream 600 LE.

[AndyOCZ]

So you said.

 

I've used it for over 3 months now and have not had a problem with mine though.

 

In fact it is only people with OCZ that seem to be complaining. Now I realize there are more people here with OCZ than other brands, but that does seem awefully coincidental.

 

 

What I know for a fact is that Redline has died as well as OCZ. We just sold tons more, so it appears we have a bigger problem.

 

Not all baords do this. Lend your memory to someone with a CH5 UTT eater and see how long they last.

 

You might never have an issue, although your overclocking and running habits may affect reliability.

[Psylenced]

Ok so are you guys ready to get REALLY confused?

 

I reset my 5v vdimm jumperback to its non 5v setting... and guess what.

 

both sticks of RAM now work again. What the hell is the deal here. With the 5v jumper on I failed memtest and prime left and right even at optimized settings one a single stick of my dual channel kit. I pop the jumper back in the default position and now both sticks run again, and are passing memtest and prime at 2-2-2-6 at 3.2v?!

 

This does nothing except maybe get me out of an RMA mind you. I still think its total BS that I cant run my RAM at specification.

 

I guess the real test comes when I unplug my system this afternoon and haul it to work with me, only to plug it in and get the memory beep of death again even with the reduced voltages.

 

Heres to hoping at least that the poor folks at OCZ dont have to replace another pair of DIMMs.

[Boppo]

What I know for a fact is that Redline has died as well as OCZ. We just sold tons more, so it appears we have a bigger problem.

 

Not all baords do this. Lend your memory to someone with a CH5 UTT eater and see how long they last.

 

You might never have an issue, although your overclocking and running habits may affect reliability.

I am not denying there is a problem. Actually I am trying to get a grasp of what it really is. I happen to be a very inquisitive person by nature.

 

Now, making a blanket statement that DFI is guilty of "designing a board that fries RAM habitually due to dirty power delivery" seems a bit overboard, which really was my point. Especially since it appears that the cold boot problem and the RAM killing problem are two separate issues altogether.

[Psylenced]

I am not denying there is a problem. Actually I am trying to get a grasp of what it really is. I happen to be a very inquisitive person by nature.

 

Now, making a blanket statement that DFI is guilty of "designing a board that fries RAM habitually due to dirty power delivery" seems a bit overboard, which really was my point. Especially since it appears that the cold boot problem and the RAM killing problem are two separate issues altogether.

 

Not really from my understanding. The cold boot problem is an issue of ram voltage not being applied properly at boot up after shutdown, but I could be wrong, and am certainly willing to conceed the possibility.

 

However even if I conceed to this point, then there are two serious problems with the mobo instead of one, unless the cold boot issue is one that can be squarely laid at OCZ's feet. But I dont hear of other mobo's having cold boot issues with VX.