ATTN: DFI nf4 users * UPDATE *

[rise]

**** UPDATE 5/29 ****

well, more of a "heres whats happening" post. as mentioned below, the cold boot issue, as suspected, should not be a difficult bios fix and currently they are sending some beta bios to users who are experiencing this exact problem.

the 4v jumper problem is still somewhat inconclusive, but testing at DFI labs indicates that the boards are NOT spiking the voltages and that the regulators are indeeed working properly. the problem seems to be mostly with UTT and ch5 and users expecting to pump that much voltage through with out a problem. if they do, then they better make sure they know what they are doing, have adequate cooling and be prepared as with all oc'ing for the consequences.

from ag-

1. electron migration - when an IC is bombarded by not only high voltages but high heat for extended periods of time, sometimes they break down and will fail to work at previously normal values.

the 5/28 update below has some more info and a link to the thread they designated to be the official vdimm issue page. Oskar Wu hisself is taking some user boards in to check the problem and sending them new ones

the main thing is, as mentioned in the manual, that pumping over 3.3v can result in components being damaged.

i will say i am very happy now with the way dfi and its employees at street are handling this now. they did what everyone wanted to and come out and say "yes, we hear you, we're working on it" which they were slow to do. of course they are busy with the x2s, a new official bios, and testing for them is alot more involved than a regular user. the concern ove the locking of the threads was valid until they gave some response, which they have.

and how great is it to see Oskar Wu in the forums taking members boards to taiwan to see if he can recreate the problems.

**** UPDATED 5/28 ****
word from ag at dfi.street
-cold boot bios fix
-voltage spikes? what voltage spikes?

**** UPDATED 5/27 ****
reply from rgone at dfi.street


as a trly to this---

from bleeding edge

just spoke to Frank Wong at DFI USA about the major problems people are getting with the 5Volt Vdimm jumper setting on these motherboards.

His response is very disturbing.

Mr Wong basically informed me that DFI R&D in Taiwan know about this issue and have been testing this setting with different major branded ram, and that the 5 volt setting is causing and i quote "Major Issues".

Mr Wong said that R&D in Taiwan reccomend the jumper is set to the default 3.3 volt setting on the board.

At this time they DO NOT have a fix for this.

Mr Wong also said "Our R&D has also found that the Vdimm jumper when set to 5 volts can cause some of the affected boards chipsets to blow"

So there we have it straight from the hourses mouth.

If you own 1 of these boards, and have the jumper set to 5 volts, change it now back to the 3.3 volt setting.

some more chat on it at extreme

thankfully as a tccd/tcc5 user it doesn't apply to me, but i know a ton of you use the vx and bh5 chips.

edit--btw, read up a bit and make your own conclusion as there is no official response from DFI but rather this conversation with a dfi tech rep. obviously not everyone is having this problem but it may be better to err on the side of caution until we hear something conclusive from dfi.

its more than a litlle disconcerting to me to see a thread on this locked at dfi.street i'll be interested to see why.

 

[Promethply]

That's probably why ASUS doesn't allow radical voltage settings on their NF4 mobos.

 

[rise]

i know it never appealed to me, it oc's like a champ with out the insane volts. and i'm never a fan of more volts.

 

[cornholio2]

I'm glad I went with ASUS A8N-E first since I want to know how fast AMD 3700+ San Diego is. It's stable on stock speed and no need to overclock. Games flow like water.

 

[Mucker]

The Ultra-D is still a superb board especially if you are a TCCD or low volt ram user. I agree with rise about volts and keeping them low. The board is feature rich and provides alot of bang for buck. I'll take 300 HTT on low volts anyday and look forward to putting a dual core on it down the road....

m

 

[rise]

mucker, i think i'm going to cut up the credit cards tomorrow with x2's coming and the g70's around the corner sli-dr, x2 4800+ and a g7800gtx or two. thats nitro right there.

maybe just the x2 and add a 6800gt

 

[imported_Bleh]

I was using the jumper and running 3.5 volts to my Value Vx to run at 250 mhz 1.5-2-2-5 but once I read this i set it back to original jumper settings and 3.2 volts. 227 mhz 1.5-2-2-5.

 

[Mucker]

LOL No sheet. The never ending upgrade cycle that I am in needs a good neutering! (ouch) I promised myself to hold off for 1 year, however, I may have to check into Upgraders Anonymous.....or better yet just buy more stuff!!!!!

m

 

[rise]

glad you saw this bleh. its not a guaranteed fry but not worth it imo.

 

[Rustler]

Mucker I may go with you to Upgraders Anonymous.................LMAO, I never can leave it alone, I am looking forward to the dual cores. Nice this is all the stuff I have sold to upgrade has upgraded someone else......It is a vicious cycle....OMG I am an enabler for upgraders...........I feel so ashamed. That is why I wear this bag the Unkown Ugrader. Yeah I have my DFI rig set at the lower voltage settings.

 

[rise]

lol rustler, know what the real shame is? i have a 6800gt packed up ready to ship once i get paid. it keeps calling to me. "sli, sli"

 

[shira]

There have been concerns, apart from this jumper issue, that using a vdimm above 2.9 or 3.0 can lead to early CPU death. It was almost exclusively for that reason that I decided not to go with OCZ Gold VX memory, despite the impressive numbers it produces. When you factor in the fact that these these high-voltage RAMs create a sh_t-load of heat, and require active cooling, it doesn't seem worth the worry and hassle to use them.

 

[rise]

all things considered i think the tccd has always been the best choice. the extra cooling for the vx never appealed to me when tccd gives 275 and up relatively easiily and some getting 300-310 :shocked:

 

[bigtoe36]

It was almost exclusively for that reason that I decided not to go with OCZ Gold VX memory

You can always run 2-3-2 with 3V or so and still enjoy low latency at pretty high fsb's

 

[darkswordsman17]

Originally posted by: rise4310
all things considered i think the tccd has always been the best choice. the extra cooling for the vx never appealed to me when tccd gives 275 and up relatively easiily and some getting 300-310 :shocked:

Yeah, I had some TCCD running 1:1 317Mhz @ 2.8V on one of these boards.

 

[rise]

had? damn, where are they, hanging on the fridge like an A+ on a report card? thats sweet man.

my winnie is croaking me around 2500 but i'd be happy if i can get 290x9 1:1.

 

[JJ650]

There was something on DFI street about beta bios allowing the usage of the yellow slots for high voltage memory (like the VX and others) and dual channel setups. This was seeming to prevent early termination of the boards. Seems like the orange slots are for lower voltage and the yellow slots specifically setup for high voltage.

I have yet to try out the beta bios. WOuld like it to be an official bios before I decided to use it. I may, however, go ahead and try it out. I have the VX ram in the orange slots @ 3.2v and am worried about killing the slots like some people have.

Althought the ram is @ 3.2v, there isn't too mch heat coming from them.

 

[rise]

the jumper is the key. additionally, if you see the thread about cold boots over there you'll notice the bios, or some of them, seem to be doing some odd things with the vdimm, in particular, having volts present which shouldn't be. that guy ozkoal did a great job trouble shooting and investigating that issue.

i'm still on the 3.10 bios as they work fine for me 24/7 so i'd rather the others be the beta guinea pigs.

as far as the orange or yellow being for high/low voltages i'd be stunned as the manual, while warning of the dangers of using such high volts, don't specify using one or the other.

 

[JJ650]

I thought the same thing. Something should be said in the manual.
But, that manual is lacking if you consider the myriad of settings with virtually no explanation.

 

[rise]

very true. look at how many people aren't aware they should have all 4 connectors in?

frankly, i'm more disappointed in the way dfi.street handled it. if you're working on a problem you don't lock the thread. thats just BS. and the linking to the thread that had nothing to do with the issue is just as bogus.

edit---the linking is bogus unless this is a bios fix like the other may be which i doubt as people are saying it took out chipsets as well as the ram.

 

[Concillian]

Originally posted by: JJ650
I thought the same thing. Something should be said in the manual.
But, that manual is lacking if you consider the myriad of settings with virtually no explanation.

No kidding, it's not even clear which PCI-e slot is the x16 slot. You just have to assume that PCI-e #1 is the x16 slot.

Worst manual I've seen in a LONG time.

 

[rise]

some words from rgone

 

[Mucker]

DFI-Street is a strange place indeed. Moderated by 2 DFI employees and they don't even post any info there about the issue. However, they do post the problem at different forum sites and have links to those forums... WTF is that all about??? The more I visit there, the stranger I find it......

m

 

[Rustler]

I am not running my memory at high voltage 2.8 is what i have my Corsair set at, but DFI caters to the OC crowd and I like to moderately OC there are many that ar rabid OCers. I think we should let the chips fall where they may and see what comes out of this. I really like my DFI board the Ultra 4, I picked up another over at Hard Op forums for another SLI projcect. I personally think when you are using voltage that high on some components you are naturally gonnna stress some. Like overloading rounds with gunpowder.

 

[rise]

is this the "official" reponse? i'm sure more will come out.