[g0dM@n]

Introduction:
Alright, I have a set of TwinMOS TMII 2x512 (UTT 50D). I had a problem running one of them VERY often, and it would RARELY boot (but when it did it would keep working as long as I never cut the power to the mobo completely either by the power plug or by the PSU switch). I also have a set of OCZ VX PC4000 2x512 that I had a similar issue with where one of them wouldn't post AT ALL.

So here I am with two 2x512 UTT sets where only one from each set posts. You would say why not use one from each, but why should I? I was so angry that I was going to RMA both, but before doing so I PMed SteveOCZ about my OCZ VX. Well, it looks like there is a fix to this.

Well, it turns out that that the DFI NF4 has a cold boot issue. What happens is that when you pull the power plug out of the PSU (or turn off its switch), and wait a few seconds until the lights go out on your motherboard you have just caused yourself a problem. Now, when you plug the power back up to the PSU (or turn the switch ON) the lights on the mobo come on again. Now your ram is going to get 2.7v and boot up with that; thus, not giving your UTT ram to boot up at 2-2-2 at whatever FSB you have it at. The thing is that some of this UTT stuff won't boot at low voltage at tight timings, and since this board tries to boot 2.7vdimm at POST on a cold boot you think that the ram is dead.

The Procedure:
The fix is this you basically put in a single UTT stick that WILL boot at cold-boot... then go into bios, set it to 3.0v or so, and then relax your timings to about 2-3-2-8. Save and exit BUT DO NOT POWER DOWN YOUR PSU!!! Let the computer reboot so that you know it's working, and then hold down the power button until the computer turns off. Turning it off with the power button on your case (or the one on the mobo) will still keep the lights on the mobo lit; therefore, there is no cold boot now! Throw in the other UTT stick in as well, and boot up the computer with 2x512. It should DEFINITELY work if that UTT is not dead.

From now on you will have to repeat this procedure every single time that you do a cold boot b/c that motherboard is going to try to boot again at 2.7volts. If you don't believe me, then whip out your DMM and check it out. I don't have a picture of where the vdimm checkpoint is, but you can try the pin 7 trick on your dimm slot to check.

Infinitevalence's Procedure (you may find his more helpful if mine doesn't work)

Warning:
Considering that you're going to have to add ram while the mobo has power be VERY careful. DON'T WEAR ANY JEWELRY ON YOUR ARMS AND MAKE SURE YOU'RE GROUNDED. The motherboard still has power while you're doing this even though the computer is off.

Special Thanks:
I'm sure plenty of people will be thanking me for this, but many thanks goes to the OCZ guys.

Thank you SteveOCZ for the heads up in that PM you replied to me with, and here is a link that he showed me about this:

http://www.bleedinedge.com/forum/showthread.php?t=10643

Let's thank RyderOCZ for the tutorial that he posted, and OzSnoal for finding this issue.

P.S. I had a talk with PMS Fishy about this and his board cold boots at 2.8v. He also mentioned some other funny stuff, so maybe you could ask him about this as well. This is why everyone should have a DMM.

[torin3]

Originally Posted by g0dM@n

From now on you will have to repeat this procedure every single time that you do a cold boot b/c that motherboard is going to try to boot again at 2.7volts. If you don't believe me, then whip out your DMM and check it out. I don't have a picture of where the vdimm checkpoint is, but you can try the pin 7 trick on your dimm slot to check.

Forgive me if I'm a little fuzzy on this, but would this possibly be able to be addressed by a future BIOS fix?

[g0dM@n]

In all honesty, I am not in position to give you an answer that will in fact be true... but if you ask for my opinion here it is:

I think that the best way to find out if a future BIOS will fix it is by taking out your bios chip and seeing what happens when you plug the power into your PSU while the chip is out. If it does the same thing even with the bios chip out, then I would say that a BIOS flash cannot fix it. If it doesn't do the same problem, then I would say that a BIOS flash CAN fix it.

I hope no one thinks my idea is stupid. As a matter of fact, I'm going to try it as soon as I get home from work. I'll leaving this sh*t hole in a half hour, and it takes me 40 min (ish) to get home.

I'll let you know. I would recommend you asking this at dfi-street, though. They will have a better answer than me.

[infinitevalence]

I have found that if you are having this problem some times you can power on and off a few times and get it to boot. At least thats been my experience with my BH-5.

[Vrykyl]

g0dm@n - any luck testing to see if your dfi nf4 board overvolts the vdimm like mine does?

[torin3]

Originally Posted by g0dM@n

I hope no one thinks my idea is stupid. As a matter of fact, I'm going to try it as soon as I get home from work. I'll leaving this sh*t hole in a half hour, and it takes me 40 min (ish) to get home.

I'll let you know. I would recommend you asking this at dfi-street, though. They will have a better answer than me.

Well, it makes sense to me . I've been meaning to get over to dfi-street, but just haven't yet. It is still a while off before I'll be getting my monster DFI system together. Local taxes were 50% higher than expected, and then some insurance premiums hit all at the same time. The $1500 threw the budget a little out of kilter.

I'm looking forward to seeing what your result is. Thanks!

[{PMS}fishy]

Originally Posted by Vrykyl

g0dm@n - any luck testing to see if your dfi nf4 board overvolts the vdimm like mine does?

Mine over volts by .15v untill I run over 3.3v, and then it under volts close to .2v.

Also, when the MB is in standby vDIMM is 2.5, when it starts to post it reads same as the 3.3v rail, then drops to the value set in the BIOS. Im running the 3.3v jumper.

I have a feeling that the MB posts for a few seconds the max voltage supplied to the vDIMM mosfets. Im going to switch it to 5v and test that theroy. That could be an explaniation as to why these MBs kill ram.

[g0dM@n]

Did you test it yet, Fishy?

[{PMS}fishy]

Originally Posted by g0dM@n

Did you test it yet, Fishy?

No. Friday.

[g0dM@n]

Alright. I just fully tested mine... 4v jumper used:

Set 3.4v in bios, save, exit, reboot --> computer boots up at 3.451 vdimm, and all the way into windows. Shut down computer and it vdimm jumps down to about 1.4v, and then gradually down to .05v (it probably goes down to zero verrrry slowly). Now, I switch off PSU and wait for the mobo lights to turn off, and then turn the switch on... BOOM vdimm at 2.51v. I turn on the computer and it jumps to 2.54v and I get my long beep. My mobo tries to boot at 2.51-2.54vdimm... GREAT!

One of my ram from my 2x512 twinmos set can't boot at that low voltage, and the same goes with one of my 2x512 OCZ VX 4000 set.

Okay, so I go to test my other theory about what would happen if you try this same thing without a BIOS chip. Okay, so I get into bios and set it to 3.4v, save, exit, reboot... computer turns on and vdimm reads 3.45v again. I now push the case power button and the computer turns off, but the mobo still has power and the mobo lights are still lit. The vdimm gradually decreases til near 0v. Okay, so now I take out the bios chip, and push the power button. The computer turns on (obviously no POST screen b/c I have no bios chip in there), and my vdimm boots up at 3.45v just like without a bios chip. I now turn off the computer via the power switch on the PSU and wait for the mobo lights to turn off. I turn the PSU switch back on, wait for lights to go on, and then push the case power button. The computer now turns on and once again boots up at 2.51-2.54v.

This leads me to believe that a bios flash will NOT fix the problem. What is it that tells the mobo to boot at XX.XX voltage? Considering the fact that the mobo knew to boot at 3.4v when I had no bios chip, where is that info coming from? Wherever it's coming from is where the problem probably lies.

[torin3]

Originally Posted by g0dM@n

This leads me to believe that a bios flash will NOT fix the problem. What is it that tells the mobo to boot at XX.XX voltage? Considering the fact that the mobo knew to boot at 3.4v when I had no bios chip, where is that info coming from? Wherever it's coming from is where the problem probably lies.

Thanks for doing the testing. Looks like I should go with RAM that deals well with low voltage. Maybe somebody will figure out a hardware mod that will deal with this.

[g0dM@n]

I'm just wondering where the information comes from to tell the motherboard to boot at X.XX volts.

[Flip-Mode]

Very interesting question indeed, looking forward for an answer.

[Janus67]

this sort of ticks me off since I just sent back a stick of Twinmos for my friend because of this exact problem, I hope this gets fixed in a BIOS fix or something, because that is one heluva problem for DFI board owners.

[g0dM@n]

Yes, it happened with one of my twinmos in a set, and one ocz vx 4000 in a set. I had to RMA them both b/c I don't feel like dealing with that. The new twinmos has no problem.

[Super Nade]

So, you took out the BIOS and it didn't boot? Why is that so unexpected?

Check this thread out at xtremesys, which talks about VTT tracking problems as a source of this cold-boot issue. One could perform two Volt mods to eliminate this (not the tracking problem though).

http://www.xtremesystems.org/forums/...ad.php?t=66624

felinusz's experience with the VMod:-
http://www.ocforums.com/showpost.php...3&postcount=35

[kukyfrope]

This solution worked to revive my Redline, once, then it went dead again after I shut down the computer but didn't disconnect power to the board :\ Really sucks that the DFI board is like this.

[g0dM@n]

Originally Posted by Super Nade

So, you took out the BIOS and it didn't boot? Why is that so unexpected?

I did not take out the bios to see if it didn't boot. I took it out to see if it made a difference with the voltage issue on the vdimm.

I was checking with a DMM the whole time with and without a bios chip, and they are completely identical symptoms. The reason I took out the bios chip was to see if that was the part that was telling the system to boot at 2.5v. Had that been the case, then I would have been convinced that a revised bios could fix it. Since that is not the case, it may not be possible... and if it is it must be tough to do.

[g0dM@n]

Originally Posted by kukyfrope

This solution worked to revive my Redline, once, then it went dead again after I shut down the computer but didn't disconnect power to the board :\ Really sucks that the DFI board is like this.

It should not do that. If you have power running to the board you shouldn't have the issue unless you have a naturally fluctuating voltage. If the system is in standby mode, then it will not default to 2.5v at boot. It should stay the same as what you have set in bios, and what the computer ran at before you "shutdown" (not power down).

[Meathead]

So, if i ever have to travel with my computer, I will have to dick around with diff sticks of ram to get this board to work? I am kind of upset because all I hear is praise about the dfi board and now once my dfi board arrives, I see a ****load of problems with it.

[g0dM@n]

You would have to clear the CMOS, then go into the bios and redo your settings with a higer voltage, then save and reboot.

[Meathead]

That sounds a lot better. What happened to having to use a diff stick? Also, could you save your bios settings and then clear cmos and just set it back to the bios you saved? or does clearing the cmos delete all your old saved settings.

[g0dM@n]

@Meathead

You don't always have to use a different stick. Out of your 2x512 (or 2x256 or whatever), you have to find the one that can boot at least at the board's default voltage. When you find out which can, label it; if both do, then obviously you won't have to label either. Now, you can save your bios settings in CMOS RELOADED or whatever it's called.

When you kill the power to the board, the board then doesn't use the voltage in bios as booting. Let's say you had 3.4v set in bios. When you turn the computer off with the regular front power switch or shutdown mode from windows, the board never loses power. It's still in standby mode for a front power switch. When you boot up the PC that way it knows everytime to boot at 3.4v. The problem is when you cut the power to the board. When you cut the power to it, then the next time you turn the PC on it's going to boot at like 2.54v, and then after a few moments it'll jump back up to 3.45v (that's what my DMM read when I had 3.4v set in bios). You'll basically need ram that can boot at ~2.5-2.6v to be able to get into bios to fix it everytime you cut the power to the board.

If your ram can handle that low voltage for booting at its SPD settings, then you can just do what you said.

Clear enough? I may have rambled on too much... I'm tired. =/

Btw, my car got towed today in MANHATTAN, NEW YORK... almost $300

[Meathead]

Will the ram be at the default settings also? I highly doubt any of my ram can boot at 2.5-2.6 with 250+ htt. I got 2 sticks of pc4000 redline and a stick of 3500 bh-5 mushkin. Also, can you just clear the cmos, boot into bios and switch it to a saved bios or does it clear all the saved bios's?

[Janus67]

everything would be reset to default settings, it won't be trying to run at 2-2-2-5 (unless you don't reset the cmos). It doesn't clear all of the old saved bios's, they stay on there (I have had to reset mine several times and it is still there in cmos reloaded).

[Meathead]

So, it's not as bad of a problem as many have been saying it is? If you can just revert back to the old settings (the saved bios) then its a minor issue imo.

[g0dM@n]

No, it's not THAT big of a problem. Also, sometimes you won't even have the problem, as my current TwinMOS UTT 2x512 can actually boot at 250fsb 2-2-2-6 timings with the cold boot issue. Mine doesn't seem to have the problem, but it does make sense to say that not all UTT ram can boot with 250fsb 2-2-2-6 timings at default voltage.

[Meathead]

Yea, I'd agree. Thank you guys a lot for the help this really cleared up a ton of my questions.

[g0dM@n]

If you do have problems in the future do NOT hesitate to ask. Most of us have already gone through the crap that you probably won't have to.

[scooter787b]

I have 2x512mb of mushkin blue cl2, and the cold boot problem has gotten worse, i press the power button and the fans and lights change for a split second then nothing, so i sit their and keeping pressing the button till it works, ive been reading this thread and dont see a true true fix?

also lately the fans have been running longer after i shutdown my system, could this be related? Im typing this from another computer aswell :\

edit: also no oc is in place right now
edit2:nvm