DFI Admits 5Volt Vdimm Problem

[topboy]

I hate to be the bearer of bad news but this needs communicating with our users

Read and weep

http://www.bleedinedge.com


UPDATE

I have just spoken to DFI RMA Department about my board........

Guess what was wrong....The chipset had blowen

Just as i was told by DFI USA.....Taiwan know that the 5 volt jumper is causing issues and could blow the chipsets on these boards....looks like mine was 1 of those.

ALSO.........BEWARE

When you RMA your board....DFI REPAIR the board NOT Replace it.

Guess i just wasted 300 bucks on a board thats now been hacked by some guy in a warehouse somewhere in a business park.

Good bye DFI.....I will never use you again....and now you have 1 customer who is on the war path to bring down you name and very bad products

Topboy

[pazza316]

Told ya!

My contact told me this ages ago!

[[XC] 4X4N]

Since this was a major selling point about the board, you think that they would have tested it better before releasing it. Now lets see if they do anything about it. I miss my 260, 2-2-2-6.

[The Mofo]

Hasnt affected my board and ive been running +5V from day one.

[pazza316]

Yeah lets see! They could repreive themselves for me by admiting the mistake and releasing a new board without the prob!

[bachus_anonym]

I don't buy it just yet...
There's a LOAD of folks out there, including myself, that have NEVER had any problems with 5v jumper. I would say, there must be a very special circumstance (type and amount of memory used perhaps) that might be causing trouble. It just cannot be said that 5v has flaws if you see that many use it with no drama at all.

If I don't see it coming from Oscar Wu's mouth OFFICIALLY, on DFI-Street, with specific, technical explanation, then I do not buy it

[G H Z]

I have run the 5v jumper ever since I got this board with no problems whatsoever, and I would also say that this is easily the best board I have ever run. 271 mhz 32mb Pi stable 2x512 BH-5 @ 1.5-2-2-3 1t.

It suck that people are having problems with this board but to me with an enthusiast board such as this it is just something that will happen. Its the same with high performance cars, they need extra care to stay in top form.

I wouldn't even care too much if I lost some Ram on this board considering how well it overclocks. The bottom line for me is I hope this does not change the products DFI delivers. I want them to stay with a serious overclocking mindset when the plan out and deliver a board. With this being a problem now I wonder how their philosophy might change.

[topboy]

Well this is what DFI Tech Support communicated directly to me on the phone.

So in my eyes this makes it official.

Topboy

[bachus_anonym]

So in my eyes this makes it official.

Not really... It will be official If we hear a chief DFI engineer admitting that. Until now, it is NOT official.

[shadowing]

I would agree with that. Just because it is a Tech Support guy that said this does not make it 'official'.

It has to come out of the mouths of the guys who engineered it.

[soul]

Bah, I've been running off the 5v rail since I got the board (granted, it's only been a few weeks at most), and haven't had problems. My memory is running within the warranted voltage range, and my board hasn't been modded in any way to void the warranty... if there IS a problem that DFI has known about from the start, they can replace my system for me when it dies. Until then, great board.

[ben805]

I suppose we should stick with the 3.3v jumper trick and ramp up the 3.3v rail on our PSU then? man if this is true then this sucks! my PCP&C 510 SLI only goes up to 3.5v on the rail so it ain't going to work....anybody try the OCZ memory booster on these DFI NF4 board?

[topboy]

Do you really think Oskar Wu is going to come on here and bash his own board....I think not.

All i am stateing here is the coversation i have just had with DFI's Technical support.

I feel it is only resonable to inform people of the information that i was told.

Yes until DFI make a statement, you may say this is not official, however if this is going to save a few people the nightmare of RMAing their boards because of this problem until DFI realease a fix, then i hope i have helped.

Topboy

[charlie]

5v from day#1, 2 different boards... still alive and kickin'

C

[Major]

I think the blame for most board failures from running the 5v rail can be placed on user error, lack of cooling, or lack of knowledge. The very the idea of running 4v on your memory should be enough of a risk to alarm someone to the idea that this might cause hardware failures and damage, and as such should be considered done at your own risk ! This is common sense stuff if you ask me.

I've had a DFI NF4 for several months now, it has been faultless in every way. I've used the 5v rail for Vdimm without issues.

my 2 cents

edit: I think that DFI had the best intentions by providing this option on the the NF4 for the enthusists market, what they didn't anticipate were peeps with limited or no experience OCing using the option.

I'm up to 4 cents so I'm going to step down now

[bachus_anonym]

Do you really think Oskar Wu is going to come on here and bash his own board....I think not.

All i am stateing here is the coversation i have just had with DFI's Technical support.

I feel it is only resonable to inform people of the information that i was told.

Yes until DFI make a statement, you may say this is not official, however if this is going to save a few people the nightmare of RMAing their boards because of this problem until DFI realease a fix, then i hope i have helped.

Topboy

Post it at DFI-Street, then... We'll see what happens

[Malves]

DFI will never say what went wrong. But SLI tells something when you think about 5v for vdimm and design of the board.

[topboy]

Post it at DFI-Street, then... We'll see what happens

Posted.....and DFI have allready responded....Mr Frank Wong, DFI USA Technical Support quoting R&D in Taiwan.....

Topboy

[Major]

see page 9 in the DFI NF4 user manual !

DRAM 4V Select

Attention:
The DRAM voltage as originally supported by the spec is up to 2.85v only. Jumper JP17 allows you to select beyond the defined spec. Although this function is supported, selecting a higer voltage may cause unstable current supplied to the system board which may subsequenly cause damage to the CPU and DIMM's

I think that pretty much says it all !

Sorry one more point, I wonder how many boards have been damaged by using the 5v rail for Vdimm while peeps were using 20pin PSU's and not using the extra molex power plugs suppiled on the mb ?? clearly this would be the USER and not the design of the mb.

[G H Z]

I was wondering the exact same thing major. How many of these guys had the problem when running a 20 pin, unapproved PSU on this board.

[topboy]

see page 9 in the DFI NF4 user manual !



I think that pretty much says it all !

Sorry one more point, I wonder how many boards have been damaged by using the 5v rail for Vdimm while peeps were using 20pin PSU's and not using the extra molex power plugs suppiled on the mb ?? clearly this would be the USER and not the design of the mb.

Well i have been using all the required power connectors ( molex and 3.5 floppy along with the 24 pin and 6 pin and my PSU is a enermax 600 watt that has the 24 pin ATX supply.

So as DFI said to me they advise to change this jumper back to 3.3 volt as this is a known problem that R&D have found in taiwan

Topboy

[G H Z]

Well i have been using all the required power connectors ( molex and 3.5 floppy along with the 24 pin and 6 pin and my PSU is a enermax 600 watt that has the 24 pin ATX supply.

So as DFI said to me they advise to change this jumper back to 3.3 volt as this is a known problem that R&D have found in taiwan

Topboy

And you lost some ram on this board? Maybe you could tell us exactly what happened in your case

[topboy]

And you lost some ram on this board? Maybe you could tell us exactly what happened in your case

I lost 1 stick of VX4000 with this board straight away when running at 3.4 volts when i forst got this board and ram.

I then had this RMAd by OCZ....great work OCZ....and have been running at 3.2 volts with the 5 volt jumper just fine for 2 months.

Then the board failed....died 3 days ago....no changes made at all to the config....it just died.

This is when i RMAd the board and talked to Frank Wong tonight about the problem.

As the news articles states at BE " 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"

I belive this is what happened to my board

So when i get my new board back...as recomneded by DFI Tech support, i will be running at 3.3 volts Vdimm jumper until they fix the problem.

Topboy

[bachus_anonym]

I have said that long time ago... I think, DFI NF4 is just not for everyone, period.

DFI says that in the Manual (like Major pointed out) and everybody knows that playing with high volts can cause trouble.
So, my advice... If unsure and not confident enough, switch to 3.3v, get some TCCD and problem solved, right? Since 5v is what everybody says causes trouble.

[G H Z]

Well based on what topboy just said if true, that the 5v jumper problem is only affecting some boards chipsets then it does sound like something they need to work out.

Still the best board ever made imo.