NF3 Ultra-D 939/AGP

[Major_PITA]

Likewise I have the resistor.

[Hoot]

I've been in Electronics since 1970 and an Engineer since 1981 and I resent being referred to as a Whiner because I bought a motherboard that was not ready for release yet. Being an Engineering type, I am not afraid to take a solder iron to a new motherboard.

 

I removed R716, not C716, nor Q716. This 15K ohm resistor appears to be a S.I.T. (Select In Test) resistor that fine tunes the Vcore regulator output to match what is called for in the Bios. That in itself is a bit of an oxymoron since it's obvious DFI doesn't test these boards to spec. or else they would not have shipped everyone with the same resistor value...

 

Here's what I measured at the socket with a high quality DMM and using the devices in my signature, while idling in the bios screen:

 

Bios Value = With Resistor / Without Resistor

VID Pass = 1.328 / 1.440

1.400 = 1.328 / 1.440

1.425 = 1.355 / 1.468

1.450 = 1.382 / 1.496

1.475 = 1.410 / 1.522

1.500 = 1.437 / 1.552

1.525 = 1.464 / 1.579

1.550 = 1.494 / 1.606

1.550 + .1 = 1.538 / 1.656

1.550 + .2 = 1.636 / 1.759

1.550 + .3 = 1.681 / 1.807

 

My guess is this resistor value should not be an infinite open, but selected to make the actual voltage agree with what you called for in the bios. If you're picky, you can install a 20K 10T trimpot in its place and fine tune the value for a perfect match, or you can leave it alone. The good news is with it gone, you should not have an undervolted Vcore condition when recovering from a CMOS reset anymore. My particular San Diego did not mind 1.328V but I'm aware that some peoples CPUs did.

 

Hoot

[terror_adagio]

The good news is with it gone, you should not have an undervolted Vcore condition when recovering from a CMOS reset anymore. My particular San Diego did not mind 1.328V but I'm aware that some peoples CPUs did.

 

Hoot

 

Indeed. With my X2 and using the default voltage, which is underclocked, I was able to get into the BIOS, change stuff around. Even format and get Windows installed.

 

It was when I was attempting to set up Windows for the first time, installing from the actual OS where I encountered problems.

 

So while it isn't a problem for some, it definitely is a problem for others.

 

More from DFI:

 

There are some hard ware design problems

with the board. The board needs to be re-worked, furthermore it is also

required updating BIOS to sort the bug out. One resistor must be taken

away, please refer to the picture. Again, I apologize for the

inconvenience may cause you.

Sincerely Yours truly,

Tech Support

[DaLurker]

Indeed. With my X2 and using the default voltage, which is underclocked, I was able to get into the BIOS, change stuff around. Even format and get Windows installed.

 

It was when I was attempting to set up Windows for the first time, installing from the actual OS where I encountered problems.

 

So while it isn't a problem for some, it definitely is a problem for others.

 

More from DFI:

 

There are some hard ware design problems

with the board. The board needs to be re-worked, furthermore it is also

required updating BIOS to sort the bug out. One resistor must be taken

away, please refer to the picture. Again, I apologize for the

inconvenience may cause you.

Sincerely Yours truly,

Tech Support

 

Does that mean when they come out with a reworked board, we can all RMA our Rev A. boards and get new ones? I sure hope so because I don't want to be buying "beta" hardware.

[terror_adagio]

Does that mean when they come out with a reworked board, we can all RMA our Rev A. boards and get new ones? I sure hope so because I don't want to be buying "beta" hardware.

 

You can RMA your board. If the new one will be better.... who knows.

[ixtapalapaquetl]

That's some good work there hoot. Thanks for the info.

[AbsoLoot]

I resent being referred to as a Whiner because I bought a motherboard that was not ready for release yet...... If you're picky, you can install a 20K 10T trimpot in its place and fine tune the value for a perfect match, or you can leave it alone.

 

 

Earlier on when I said "crybabies", I didn't mean those who can find a problem with the board, but those who can't live with those problems. A good tech like yourself can find a problem on any board, just like you can take a brand new car to a good mechanic and they will find things wrong with it. I was referring to those on here who don't care what is wrong, or how to fix it, only that something is wrong and they are owed because they want $250 worth for their $125.

 

Sorry if I called any grown-ups crybabies.

[brakezone]

Well, if you're going to do it, I'll hold off. I have 5 different soldering irons for different jobs, including an SMD station for rework. I was hoping it would address one of my problems, but I don't have the "Cold boot flu" on my unit, at least at stock speeds, but it could account for being unable to run 300 HTT, like I could do on my Neo2 Platinum. I'll watch for your results.

 

Hoot

 

So you absolutely cannot get your board to run 300htt under no circumstances hoot? Even with 100mhz ram and a low cpu multi? I know in your case you either have memory that doesn't work well with the board or the board itself is messed... well i don't know that is, just guessing.

[DeNs]

Hoot - thank you very much for testing that. It seems like the resistor will instead of being removed completely be replaced with one of a different resistance, but that's just a guess.

 

Thing is, even if it does make it overvolt, does it actually stabilise the vCore and make the fluctuations less wild? Mine's pretty much fine when it's under load (spot on really), but when idling it tends to go a little strange (around +0.05v usually).

 

If this really is the only issue, then I think i'll be keeping this board - my vCore is fine for what i'm doing, and the chip is more limiting than some of the other problems on this board anyway. However I guess i'll decide if/when the time comes.

 

Such a shame that my ram can do over 300HTT but my CPU will do no more than it

 

---dens

[NineBall]

The following picture was sent by DFI to several users including myself. My board has a resistor in the spot where the arrow is pointing to.....Does anyone have a board with NO resistor in that spot?

 

 

Just checked, I have that little resistor. But tell you what, that BIG 1.25" square

ic in this pic is actually 1/4" in real size. And that resistor, you need a big

magnifying glass to see it. There is just no way to safely remove it for 99% of

us. The tip of my old soldering gun would burn the surrounding items.

Looks like the only way to remove it is to put the point directly on the center of it

but soon as you remove the iron, the solder will harden... no way to grab the

sucker. If you have ever seen these mobo tooling machines you would see that

the hundereds of soldering points are done with hot "pins", not irons actually.

 

And from hoots volts with/without resistor, looks like now it overvolts by the same

undervolt margine. I got a SD chip so don't know if it matters when under.

Geez, wtf can be done now?

 

edit: oh and btw, why does that pic with NO resistor look so clean, like it was

NEVER there in the first place ????

 

ps. - thanks Hoot!

[Guest Lucky]

I've been in Electronics since 1970 and an Engineer since 1981 and I resent being referred to as a Whiner because I bought a motherboard that was not ready for release yet. Being an Engineering type, I am not afraid to take a solder iron to a new motherboard.

 

I removed R716, not C716, nor Q716. This 15K ohm resistor appears to be a S.I.T. (Select In Test) resistor that fine tunes the Vcore regulator output to match what is called for in the Bios. That in itself is a bit of an oxymoron since it's obvious DFI doesn't test these boards to spec. or else they would not have shipped everyone with the same resistor value...

 

Here's what I measured at the socket with a high quality DMM and using the devices in my signature, while idling in the bios screen:

 

Bios Value = With Resistor / Without Resistor

VID Pass = 1.328 / 1.440

1.400 = 1.328 / 1.440

1.425 = 1.355 / 1.468

1.450 = 1.382 / 1.496

1.475 = 1.410 / 1.522

1.500 = 1.437 / 1.552

1.525 = 1.464 / 1.579

1.550 = 1.494 / 1.606

1.550 + .1 = 1.538 / 1.656

1.550 + .2 = 1.636 / 1.759

1.550 + .3 = 1.681 / 1.807

 

My guess is this resistor value should not be an infinite open, but selected to make the actual voltage agree with what you called for in the bios. If you're picky, you can install a 20K 10T trimpot in its place and fine tune the value for a perfect match, or you can leave it alone. The good news is with it gone, you should not have an undervolted Vcore condition when recovering from a CMOS reset anymore. My particular San Diego did not mind 1.328V but I'm aware that some peoples CPUs did.

 

Hoot

 

Very good to know..well done.

I very much doubt that this is the main issue with this m/b. Im not an electrical person but the 1T and DRAM issues need to be worked out b4 the postal service should be overloaded with NF3UltraDs just to fix vcpu. The cpu voltage dicrepency is a very minimal problem that can be overcome by a simple adjustment in the bios to counter the undervolting. Id say a complete rev2 may be required to fix completly what is bugging a lot of peeps.

[NineBall]

This dfi thread on www.pcper.com forum explores the "cold boot" issue related

to other machines (its only 7 or so pages, the first page reads right out of some

of these thread posts here).

 

http://forums.pcper.com/showthread.php?t=296126

 

I think its not a dfi issue, but is a bios issue.

Even on my asus nf2 board, I would get the once in a blue moon "no signal" vga

thing where I had to reboot, but that was with the amd xp class machines.