Jump to content
Sign in to follow this  

Cold Boot? Memory Died? [email protected] = who you need to contact

Recommended Posts

Yes, that was me - I tried with the booster.


The problem is that my RAM was already dodgy. All the booster really proved to me was that no matter what the voltage applied to the DIMM I still would cold boot.


What is still an unknown is why the VX ram is dying...

On the once side of the fence, high voltage RAM is getting the blame and on the

other side of the fence the motherboard is being blamed for killing RAM.


Angry I appreciate your feedback, you may be correct but until there is an official standpoint, and some HARD facts as many like to put it. Remember Anandtech gave VX and DFI a good rap - should they take it back? Let's just say, as far as I am concerned this is an unresolved issue.


A little rant...

Guess who sits with the problem?! I bet that in the end it will all be boiled down to the user who was stupid and should have read the 'disclaimer' and didn't cool the ram and should not have used the 5V jumper and should have burned in the ram for a little longer and should have had a power supply that was rated at xxx and should have come to the forums before plugging in his ram and should have made sure that the one setting in BIOS is correct and should have changed to the 3V3 JP17 settings and made sure that they followed some cryptic way before doing it and should have had the correct bios and should have plugged his ram into the correct slots and should not have used .... and if not - you're dooomed and it's all your fault! Have I left anything out?


/UV On, Tweak fun!!


Signing out...

Share this post

Link to post
Share on other sites

I doubt seriuously if any of the Winbond ram was intended to have 24/7 tolerances at 3.2-3.7 volts regardless of what any module maufacturer might determine.

Share this post

Link to post
Share on other sites

Deeyam I really want to stay our of this willy nilly thread of what ifs. HEhehehe.


"OzSnoal" as E_R would say, we wood lak some hard facts madam. I remember this near same scenario once before with many of the NF2 boards before the *appearance* of a demise of BH-5 type memories. Which while they were apparently demised; there was little memory trouble as I experienced it. Onward. A huge number of the NF2 boys modded their power supplies and also the boards to get Vdimm in the 3.4 to 3.6 range and killed the old original standbye BH-5 in times past by a year or so.


Now enter the resurrection from the dead of the BH-5-like memory and high volts and death again. It makes no sense to me to expect memory to run all the time with voltages that are very near to causing death in the first place and only a few ticks or less from running smooth and completely out of spec.


And yes this no matter how many think it should run that way. I don't know which side of a fence to place the failure. I have seen nothing that indicates to me to run out and buy memory that requires high voltages to run stable at this time. Now that much I really do understand. The rest is just they said, he said, no really they said, h*ll man didnt' you hear what they said? And on and on and I suspect will be a pattern for time to come. Because the memory is being pushed to within an inch of its' life and that never bodes well in the long run in my feeble MS red-nake mind.


Have fun as you are now returned to your regularly scheduled programming.



Share this post

Link to post
Share on other sites

Well you've been sucked in...... You need a sense of humour to deal with all the whiners - eh?


Thanks for the info on the other NF2 + ram debacle. I hadn't heard of that.


Rgone, if you would recommend OCZ ram which would it be? If OCZ would be kind enough to swap it out (I would need to contact them about it) for different make and model. I paid 224Euros/$275 for the VX (I see it for $227 at newegg).


PM me if you like.

Share this post

Link to post
Share on other sites
what you should be doing is getting rid of that VX/UTT based memory as it seems to be the cause of the cold boot. try any other low-voltage memory and see what you come up with.


You're joking, right? You're NOT suggesting that OCZ makes crappy or incompatible ram and that I never should have bought it in the first place, are you? Is OCZ aware that you're advising people to dump their products? Is that an OFFICIAL statement by DFI?


I paid a significant amount of money for that ram and based my decision to go with OCZ VX ram on both past experience (I've owned OCZ ram for about the past 4 years) and reviews I've read on the DFI NF4 Ultra-D and OCZ VX combination. Are you now suggesting that those reviews are or were bogus?


I THOUGHT that DFI and OCZ worked rather closely together on the VX / NF4 Ultra-D/SLI/etc. combo and since it's the ONLY motherboard that I am aware of that supports the VX specification, one would naturally assume that the two companies at least exchanged pleasantries.


Somehow I suspect though that it takes a LOT more than just pleasantries for a motherboard manufacturer to support a very specific ram specification. Or is that also a mistake on my part (aside from the mistake I made buying these products and expecting them to actually function as a cohesive unit)?


You're opening up a BIG bag of worms here, AG, and I sincerely hope those words of advise you wrote don't come back to haunt you, i.e. --- This just in ---- "DFI just announced that users of their NF4-based motherboards should stop using OCZ VX ram".


Or should I interpret what you wrote as just YOUR opinion?


Anybody out there from the OCZ side of this discussion care to offer a comment or opinion?




DFI NF4 Ultra-D, OCZ PC4000 VX 1GB, AMD A64 3500+ Winchester, Sapphire X850 XT, Thermaltake PurePower 680 watt (NATIVE 24-pin).

Share this post

Link to post
Share on other sites

Oh how some people can put in words what my mind was thinking........and ive never met you before...


Lol...i was thinking the exact same thing Old_Fart.


OCZ are on there way to comment i hope


As is AG i hope


Topboy out and waiting

Share this post

Link to post
Share on other sites

ok let me rephrase what I said as it might come across that i was blaming OCZ or any other mfg of UTT memory (which I am not)


I am not blaming OCZ nor any other mfg. What I should have said is that you all need to keep in mind that the memory (UTT) might be able to handle high voltage and mhz, as our board is able to handle, but that it still might fail anyway (using high voltage etc).


we (DFI) cannot control the heat generated by using this much voltage. That is a by-product of using excessive voltage on a memory module. Test it with any board where you put high voltage to any part (gpu, cpu, DIMM module) and you will see...high voltage means high heat, and that can ultimately kill any microelectronic part. This should be logical assumption (and i apologize if you don't recognize it as logical, but it truly is, that is why microprocessors are rated for a certain voltage).


Even UTT stuff that is supposed to be able to take high voltages is not guaranteed (actually the memory mfg might guarantee it, but DFI cannot guarantee what a memory module might or might not do when you apply extreme amounts of voltage to it) to last very long, no matter what the warranty is.


if you are not sure, or cannot accept this as an answer, then you need to stop using the memory/mobo combination (high voltages). This is the only real answer I can give. When I set up my boards with high voltage memory and I crank up the voltages to them, I accept the fact that the lifespan of the hardware might be greatly reduced. I have found this out first-hand with a 3200+ Winchester that used to do 2660Mhz, but now will not even clock past 2375Mhz no matter how much voltage i apply to it. My only recourse is to stop trying to pump up the mhz/voltage on it and use it normally.


That is why i suggested you should return your memory and get something that does not require high voltage. It was not intended to slam OCZ or any other company that offers high voltage memory. It is a logical point of view from me that using such voltages generates lots of heat, and heat + voltage = lowered life span of a product (especially memory/cpu/gpu).


also keep this in mind:




"who knows how these modules will hold up down the road. With a lifetime warranty on memory running 3.5Volts possibly 24/7, we might be seeing a lot of burnt out modules one or two years from now. Don't forget this is the first time we are seeing 3.2+ volts on motherboards and 3.5+ volts on DRAM as standards, so who knows the real lifespan of these products."


the part about 3.2v on the motherboards is not necessarily true, as our NF2 boards had 3.3v as an original option (LPB/Ultra Infinity NF2), but in a sense it is correct because no other mobo mfg offered such voltages, and before this new wave of UTT-based memory, no memory mfg offered such extreme voltage warranty on their modules.


No one knows how long a component that is being overclocked will last. No one can say for sure, 100% fact and guaranteed, that an overclocked, over-volted component will last 5 years, 5 months, 5 weeks, 5 days, or 5 hours.


When you run anything out of spec, you take a risk that it will not last. When you run it within spec, you are guaranteed a certain amount of time that the component will last, and if it fails within that time, you are guaranteed a replacement.


as I have said many times, we offer the options of high voltages and high mhz and all sorts of overclocking goodies, but we do NOT offer any warranty or guarantee that your board/components will achieve ANY overclock at all. Period. We only guarantee that your board will perform at stock speeds and stock voltages. That is all we can offer...if we didnt, there would be a zillion RMA's because you 1, didnt get a 5Ghz overclock, or 2, you pumped 2.2v to the cpu and 3.6v to the memory and your hardware died or you still couldn't get 5Ghz overclock, etc...whatever the case, guarantee of any kind of overclock is suicide to us.


We can only guarantee that the options to run such voltages, timings, settings, etc are going to be on our board, and again, we only guarantee that the board will perform to spec @ stock settings.


hopefully this clears up any ideas that I was blaming any one specific mfg.

Share this post

Link to post
Share on other sites

Cool thats good to know AG. Thanks for clearing up that information.


Ok i think what we need to do now is all sit tight and wait for the boys at DFI and anyone else involved to come back after their testing and post the results.


I know this will take time and look forward to the results.



Share this post

Link to post
Share on other sites

Actually, I haven't really tried to over-clock my ram at all. I can't even get it to run reliably at stock speeds.


As I understand it, the stock speed for OCZ PC4000 VX ram is 250MHz and it's suppose to perform at that speed with timings set to 2-2-2-8 and at 3.3 volts. If I'm wrong in any of these numbers, then someone please correct me but I believe the above specs are directly from OCZ's website.


The following IS directly from their website ----


"The PC-4000 Gold VX is made to operate above the traditional memory voltage to increase speed and stability to ultimately allow the memory to thrive under high voltage."


Thrive. It doesn't say tolerate, it says thrive. I'm pretty sure that even OCZ engineers are aware that increasing the voltage to their chips will also create a corresponding increase in the heat they generate. Ya think?


Also from their website ----


"OCZ EVP® (Extended Voltage Protection) is a feature that allows performance enthusiasts to use a VDIMM of 3.5V ± 5% without invalidating their OCZ Lifetime Warranty."


Lifetime warrantee. Sure, I'll try and get at least 3 years out of this system but in all likelihood, I'll be surprised if it lasts till the end of the year at the rate it's going. OCZ does in fact stand behind their products and regardless of the circumstances, I never had a lick of trouble exchanging the ONE defective ram module I purchased 2.5 years ago that simply failed for no apparent reason. That was from a batch of about 10GB of ram (mostly 512MB modules).


Also and unless my math is wrong, a 3.5V +/- 5% tolerance works out to a range of 3.325 to 3.675 volts. Giving my Thermaltake PurePower 680 the benefit of a doubt, I'm relatively confident that the voltages it's supplying to the motherboard are quite stable. I've had a pair of fairly expensive Fluke digital multimeters and a very old Simpson analog multimeter attached to the 3 (important) rails for about 2 weeks now and I haven't seen anything to convince me that my power supply is the problem.


Not having the specs on the voltage regulators or discreet components (resisters, capacitors, etc.) used on the motherboard that are suppose to regulate those voltages (at the chip level), I really don't have a clue as to how stable the voltages to the ram, CPU and nForce4 chipset are but I'll also assume their within tolerances.


More from OCZ ---

"The Gold Voltage eXtreme product line is 100% hand-tested to ensure OCZ quality and is optimized for use on Athlon 64 based motherboards. OCZ VX Gold Edition memory modules are rated to handle up to an incredible 3.5V without invalidating the OCZ Lifetime Warranty so that performance enthusiasts can tweak their systems without worry."


Personally, the highest I've set my VX ram to is 3.3 volts, so I haven't really pushed it at all. But, that "tweak their systems without worry" part of the last sentence seems like a pretty bold statement, especially since I haven't been able to tweak my system at all. I'd also REALLY like to know who's motherboards their using to do all that hand-testing and optimizing of their ram with. I wonder if they're experiencing cold boot issues too?


And now for the bad news ---


At the moment my system is running at 200MHz FSB/HTT, timings are at 2-2-2-6 and the ram voltage is at 3.2V. I put the 4V jumper back to it's default position and my VX ram simply won't run at any voltage below 3.2 volts anymore. That leads me to believe it's probably on it's last legs (meaning that very soon I'll be contacting OCZ to replace it).


Which leads me to the obvious ----


OCZ seems very confident that their ram can handle a great deal more voltage (with the obvious addition of a GREAT deal more heat) and still survive. They don't say that it'll last 2 years, 2 months, 2 weeks, 2 days or 2 hours. They simply say it's warranted for as long as I own it. Whatever is causing these problems (cold boot issues, fried ram issues, high taxes, whatever), I could really care less. All I want to see is a solution that works.


I'm NOT asking for anything more than the performance I paid for. The reality of all this is that I would be perfectly happy if I could get my system to run (and BOOT) stable at 250MHz FSB at 2-2-2-8 and 3.3 volts. For the very brief period I actually saw it running at those speeds and settings, I snagged a 3DMark2001 SE score of nearly 29K with everything (including ATI's 5.3 drivers) at stock settings. I'd call that acceptable, given my A64 seemed perfectly happy at 2.5GHz and 1.5 volts. My core temperature was at 28C within 60 seconds of that score, by the way (thank you Danger Den).


I'm not interested in over-clocking to extremes. I am, however, very much interested in running at stock speeds (DDR500) and being able to boot my system every time I hit the power button and NOT have to hit the reset button 3 or 4 times to get into Windoze. When I get there, I want to USE my system for the purposes I built it for (mostly video editing) and I don't want to worry about whether my ram is going to die next week while I'm using it.


Somebody (take your pick) screwed up and I'll be damned if I'LL take the blame because I chose the components I read about that appeared to offer very good performance at a reasonable price. I'm not trying to set a new land speed records or pipe freon into every orifice of my Lian Li case. All I want is good, reliable performance that I can depend on --- EVERY DAY. That's what I paid for and that's what I expect.


Is that too much to ask?




DFI NF4 Ultra-D, OCZ PC4000 VX, AMD A64 3500+, Sapphire X850 XT, Thermaltake PurePower 680 and other assorted stuff not worth bragging about.

Share this post

Link to post
Share on other sites

It seems something is wrong with the dram driving method somewhere...sorry i can't prove it....


my second pair of bh-5 pc3500 start to freeze after i pop in my venice for 2 day..I ran the ram with 3.2v for a week before up the voltage to 3.3v + 5v supply..


venice seem much more tolerant w 0510-2 bios with bh-5. it will never freeze in memtest #8..but i lost a gauge to find how the bh-5 would be comfortable with before die....


my 1st pair went dead after it freeze in memtest #8 but i continue to burn it in memtest #5....thought was burning in phenomena...


so 2nd pair i was extremely cautious and gauge from memtest #8 phenomena....


venice is too tolerant and no errors till 250Mhz w 3.4v + 0.03v ( it allows though )...


after a day, windows keep freezing or retarded as it should be....screen freeze for a couple of seconds intermittently...I reboot and memtest it.....tonnes of errors coming out from it and can't even do anything like CL2 200 Mhz anymore....this is just a 2 week ram only.......


went into bios and rekey all setting manually again..ok for few rounds then same problem....


I quickly took it out and pop in tccd...


leave the ram around for a day and only pop it in my spare gigabyte nf4 board to test again.....CL2 210Mhz no problem continuously for 2 day memtest at 2.8v...I would vmod it to test w 3.0v again ...


my SLI-D never allow my bh-5 to run at CL-2 looser timing on this board and i had to crank up to CL-1.5-2-2-2 to run it with winnie...


not a complaint here as i am trying to deal or work around it but would like AG to take note of it or maybe try to replicate it for the benefits of clearing this weird clouds over this area....


That is the senario i face with....

Share this post

Link to post
Share on other sites
This topic is now closed to further replies.
Sign in to follow this  

  • Create New...