Cold boot issues on DFI, i need testers

[Tony]

what is the cold boot issue exactly ?

There is a huge thread Eva on XS that explains, its basically issues booting with UTT at high voltages.

[Nugz]

Something i find weired with Just my 1 bad stick. Not the good one.
No matter what mother board i ues, the bad stick, after running a little bit, it has to cool down to about/ or around 20c or room temp befor it will boot again.
Tryed in.
IC7 v1.0, 3.2e D0
AS8, 530j
P4C800-E Delux v2.0, 2.8c M0
K8NS-Ultra 939, CBBID 0451 3200+
DFI NF4 Sli-DR Rev.AA0, CBBHD 0449 3000+

Usually wait about 20 mins and the bad stick is good to go. Stable @ 2-2-8-8-16-2-2-2 245 3.3v

I tryed the 3.3v mod. but like it's going to help. when i tryed it in other mother boards and the same thing.

ALL THIS HAPPEN "AFTER" I PUT THEM IN THE DFI NF4.

They worked fine in the IC7 @ 250 3.4v 2-2-2-5 A-N-A-D-D
Shut down and start back up fine.
Same in the K8NS-Ultra 939 245 @ 3.3v
then along came the DFI NF4. and now one stick is FUBAR. for sure.
Seeing as i got some 3200 Gold and tryed them on all the MB's ya see above and they work fine.

edit:: tryed the 3.3v mod = still have to leave the power supply on for the 2 sticks to work. Cant shut down the PSU, then try to boot, dont help. even after waiting 20 mins for the bad stick to cool. + side tho . dont need a fan over the dimm area.

Anyways
Peace Nugz

[maxoverclock]

bigtoe,

ocz, siad to trade the vx i had with some tcc5 pc4800. but some with pc4800/5000 have had problems so with the problem with the utt being worked on i am thinking that i should keep the vx sticks....what do you think...

i bought some vx sticks a 5 weeks ago and they ran perfect for almost 10 days then would not post. returned them to monarch for replacement....

i bought some mushkin redlines and they ran perfect (even better timings than the vx sticks). they ran for 9 days and then i shut down at night and they would not post in the am.

i have the new replacement sticks vx sticks that post fine right now, but i am sure that if i run them in the same thing will happen. i can ask ocz for an extension on the rma and do some testing for you with the new bios on the sticks that work right now....

max

[Tony]

If you have new modules, don't run the 5v jumper, run off the 3.3v with the tweak as i have a feeling its a lot better.

Im now seeing the board is doing something to the modules, so i will get back with DFI ASAP.

I have a fixed eye on this thread, sit tight i will email all those who have mailed me already.

[maxoverclock]

if the 5v rail is doing something to the sticks than it is specific to utt since my old bh-5 has been running the 5v rail for a long time without any problem....

i'll switch to the 3.3v rail and burn the new vx sticks in. my rma is good for another week and a half so i can experiment with them for a few....

max

[Tony]

NEW FINDINGS

Possible reason and maybe a work around, this may have been already posted but my info may elaborate a little.

If you switch your pc off on the wall or at the PSU after you shut down, upon switching the AC power back on you will have 2.6 to 2.8V supplied to your ram.Upon trying to boot the board looks to go thru 3 suspend loops before Vdimm is applied to the memory...this may cause the cold boot issue.

If you shut down and leave the wall switch and psu on the board will completely shut down...no voltage will be applied to the dimms. Upon a reboot the vdimm is applied immediately...the board boots as it should do.

So info i need.

How are you shutting off and what do you need to do to boot properly?

If you have one bad dimm now i doubt this will help you, but its worth a shot.

[maxoverclock]

i always shut down using standard window shut down. psu and wall plug were always left on/in...

[chainhoe]

right, i was wondering why my old TwinMos UTT wont run on 5V rail anymore. i'v got 4 sticks of old BH5 and they all run fine, been on 5V for about 2 months now

[THunDA]

NEW FINDINGSIf you switch your pc off on the wall or at the PSU after you shut down, upon switching the AC power back on you will have 2.6 to 2.8V supplied to your ram.Upon trying to boot the board looks to go thru 3 suspend loops before Vdimm is applied to the memory...this may cause the cold boot issue.

If you shut down and leave the wall switch and psu on the board will completely shut down...no voltage will be applied to the dimms. Upon a reboot the vdimm is applied immediately...the board boots as it should do.

So info i need.

How are you shutting off and what do you need to do to boot properly?

If you have one bad dimm now i doubt this will help you, but its worth a shot.

Bigtoe..

I havent had cold boot problems but it seems my tccd has been getting real instable the last few days ( Its been wanting more and more voltage ).. Last night I was testing with a dmm and noticed exactly what you explain above..

If I can help in anyway let me know..

[Nugz]

uesing a DMM. i mesured the Vdimm after turning on the PSU 2.5v, and can't boot with both sticks.
Boot the good stick, in slot2 the volts start out @ 2.3v then 2.5v then few seconds later 3.33v.
shut down, leaving the Power supply on. 0.0v and i put the second stick in and instent 3.33v and both sticks boot and run fine 3.3v @ 245.

Need to change the BIOS start up Vdimm. or something.


Peace Nugz

[OzSnoal]

bigtoe, the 'new findings' I posted over at the DFI forum a few days ago, I thought you knew about it already - otherwise would have mentioned it. Screen shots of me measuring Vdimm and a movie of how my board powers up when AC has been removed, with one stick of RAM.

http://www.dfi-street.com/forum/show...2&page=2&pp=15

The problem is, that I have also seen when I do a normal shutdown (i.e. do not remove AC) and wait 30 seconds and power up again (which means that both dimms will have the CMOS specified Vdimm), I still get a boot problem. My machine has to cool off/the stick cool off(similar to what Nugz says further up).

Will you try that in your tests? Find a stick that fails to 'AC removal' cold boot and then try boot into windows, shutdown, wait 30 seconds and power up without removing AC.

[mikead_99]

Hmmm, interesting findings, not quite what I see with cold boot though. I run my system on an ups, and only kill psu when removing/replacing components in the case. I've run into cold boot on shut-down/restarts. Both from Windows, and even when escaping from memtest runs......100% stable memtest runs. I guess in my case it has technically shown as more of a "warm-boot" issue. Removing a stick has recovered this more often than not.

Still not testing on the UTT Test bios yet. As mentioned in my pm, I will only be able to start this later today. I will use only +3.3V with the +5V trick and just see what happens from there. I've got 3 of these boards here now, all Ultra-D's. One pops up with the reboot 3 LED freeze fairly regularly, one just seems hosed with Venice and 2-2-2 so I won't use it at all, and the other is a cherry waiting to be popped. I'll start with the one that I mentioned first and move on to the new one once I get some data.

One a side note, where is a convenient spot to clip a lead on to measure VDimm? I'm real curious to what I'll see with the warm-boot board.

[Tony]

measure vdimm on the missing cap left hand side of the dimm slots top of motherboard.

[OzSnoal]

PIC 2 shows that. Your vdimm will travel from 0V to the voltage set in the BIOS immediately when you power up. This will always be the case when you have not removed the AC.

I have started having 'Cold Boot' problems always when removing AC, now I can cold boot problems even if I shutdown and power up without removing AC and if I let the PC stand for awhile (1 hour), it will power up, BUT never if I removed AC though.

[tictac]

i found this in ABIt uGuru Power on sequence... not sure it will help or not...
anyway here you go...

1)enable atx power supply
2)atx power supply ready
3)ddr voltage ready
4)setup PWM for VCore
5)Assert PWM for VCOre
6)Check CPU VCore
7)VCore Ready
8)Initial Clock Generator IC
9)Chipset Voltage Ready
10)AGP Voltage Ready
11)3V Dual Voltage Ready
12)VDDA 2.5V Voltage ready
13)GMCHVTT Voltage ready
14)Check CPU Fan Speed
15)Assert all power ready
16)AWARD BIOS take over booting job

Hope Oskar Wu can fix it... GL

[Crazykooter]

I am starting to believe its more then just a utt issue. I have my utt coming monday but I have been using tccd which the dfi has required me to use more volts then normal IMO. Could this be a voltage controller issue instead of a bios issue ?

EDIT: The reason I say this is I have 2 DFI UT Ultra D boards. When I put my SD 3700 In older one it takes 1.72 to run stable @ 2860 mhz. With the new board I can run 2880 mhz stable @ 1.6v. That is a lot of diff for 2 boards exactly the same. We will see if it starts getting to the point I need more vcore soon I suppose. I think this has to do with the big picture of the cold boot issue.

[ninja_xxx]

I am starting to believe its more then just a utt issue. I have my utt coming monday but I have been using tccd which the dfi has required me to use more volts then normal IMO. Could this be a voltage controller issue instead of a bios issue ?

EDIT: The reason I say this is I have 2 DFI UT Ultra D boards. When I put my SD 3700 In older one it takes 1.72 to run stable @ 2860 mhz. With the new board I can run 2880 mhz stable @ 1.6v. That is a lot of diff for 2 boards exactly the same. We will see if it starts getting to the point I need more vcore soon I suppose. I think this has to do with the big picture of the cold boot issue.

what version is the one that use's more vcore....

[Crazykooter]

what version is the one that use's more vcore....

They are both A02 versions. Thats what makes me wonder how the exact same board can have so much variance as these do. It has to be the voltage regulator. Its what makes me think the cold boot issue is coming from the voltage reg as well. It may be that UTT is ultra sensitive to any changes in voltages and causing them to fail. I hope im wrong because I have some utt coming next week.

[Snark42]

I am starting to believe its more then just a utt issue. I have my utt coming monday but I have been using tccd which the dfi has required me to use more volts then normal IMO. Could this be a voltage controller issue instead of a bios issue ?

EDIT: The reason I say this is I have 2 DFI UT Ultra D boards. When I put my SD 3700 In older one it takes 1.72 to run stable @ 2860 mhz. With the new board I can run 2880 mhz stable @ 1.6v. That is a lot of diff for 2 boards exactly the same. We will see if it starts getting to the point I need more vcore soon I suppose. I think this has to do with the big picture of the cold boot issue.

Same power supply was being used in these circumstances? Same BIOS? Only the boards were different or were some other components different?

[Crazykooter]

Same power supply was being used in these circumstances? Same BIOS? Only the boards were different or were some other components different?

Same bios same everything. All I did was swap out the board and flash the bios to 414-3. It seems to be the best bios for me so far.

[saiamne]

EDIT: Edited all over the place

>>I now need to know which module it is that causes the board to not boot.
>>Is it the dimm in slot 4...IE the orange next to the CPU
YES, the dimm in slot 4 causes me not to boot.

>> can you please also test whether both dimms alow a clean boot in single
My one DIMM 'normally in slot 4' cannot boot in any slots - yes, I know, email you.

>> Now running RAM off the 3.3V rail (Vdimm = 3.2V and measured with multimeter)
I cannot boot with both dimms, 200FSB, 3-3-3-8 @ 3.2V
I cannot boot with both dimms, 240FSB, 2-3-2-8 @ 3.2V
Only can boot with the one single dimm 'Normally in slot 2'.

>>can you please also test whether both dimms alow a clean boot in single channel mode.
>> Do these tests at 2.5-3-3 timings 2.8Vdimm 200fsb.
My 'Normally in slot 4 dimm' cannot boot in single channel in Slot 2 @200FSB, 2.5-3-3 @ 2.8V
The other dimm boots fine.

From what I've read here and DFI-S my problems are pretty much the same as Oz's...my "normally in slot 4" stick is now completely DOA...all booting and testing works fine for me using my "good" stick of ram (only after booting with my PC27000 and flashing to 510-2 bios was this at all possible, previous to that neither VX stick worked in any configuration)...all slot configurations work fine with my older PC2700 (2x512) ram from an older computer (they're Corsair xms2700c2 rev1.3, Mosel Vitalic modules)

Been running everything off the 3.3v since I got it back up and running this afternoon...haven't ventured into the 5v yet

I always do a "soft" off...windows shutdown only - not PS off/wall off

I think I'm in agreement with what a few others have speculated, that there is some sort of voltage spiking/instability when running off the 5v rail...but I'm this for the durration, don't wanna give up on my SLI-DR or my PC4000VX...will test as much as I need to in order to get this right for everyone

[Snark42]

erm.... I hadn't been having any cold boot issues but I just flipped the ATX switch off and left it for 30 seconds and then turned it back on and tried booting and it just never got to post. I tested the vdimm and this is what I got..
http://img.photobucket.com/albums/v4...rk42/vdimm.jpg
....as you can see, it's only at 2.54, which isn't enough to boot VX @ 200-2-3-3-8??

I turned it off by holding down the power button but left the AC and switch on. tested the vdimm...at 0. turned on and still only goes up to 2.54vdimm.

So do I just leave everything for 24 hours and hope for the best?

[OzSnoal]

That may help, as I need a cool off period - this may always have been the case, I just might not of noticed. Cause normally when I switch my pc off, I would not be using it for at least 1 hour.

Otherwise, if the cool off period doesn't work, try boot up with only one stick, alternating between the two until you do. Loosen your FSB to 200, drop volts to 3.2V, loosen ram timings to 3-3-3-8, shutdown, do not remove AC, plugin second dimm.

My Note:
It looks like my one stick of 'bad' ram, once it has been used (i.e. booted up), needs to have a cool off period. I am using the OCZ booster at the moment, running at 3.5V and still have cold boot problems. Personally, I think my one stick of OCZ VX has gone 'bad'. I say bad in quotes, because when I boot with it, and manage to boot with it, it works wonderfully. But it is a stick that once booted up with, I cannot cold boot again without some sort of resting period. It also will not boot the PC when being used as a single stick.

I think my one stick is bad, in a strange way - what caused it and why is a good question. We are all looking for that answer, and eventually we will find it.

[ninja_xxx]

They are both A02 versions. Thats what makes me wonder how the exact same board can have so much variance as these do. It has to be the voltage regulator. Its what makes me think the cold boot issue is coming from the voltage reg as well. It may be that UTT is ultra sensitive to any changes in voltages and causing them to fail. I hope im wrong because I have some utt coming next week.

nooooo did you change the version number cause i could swear it was something else before.... i have A02 also

at dfi street there say'n everything is ok... i don't know what to believe anymore... will i might get a rma on my psu so may ocz will give me the powerstream instead of the modstream then i could use the trick to get more volts

[Crazykooter]

nooooo did you change the version number cause i could swear it was something else before.... i have A02 also

at dfi street there say'n everything is ok... i don't know what to believe anymore... will i might get a rma on my psu so may ocz will give me the powerstream instead of the modstream then i could use the trick to get more volts

Ya I did change the version number. SOrry it was a typo. I have a powerstream 520w which gives me unreal stability on my rails. I will know more on tuesday if I will have the cold boot issue with mine as well. This twinmos utt has already had a issue on one DFI board. Since I have 2 boards I am gonna be able to test more then he did. I am hoping it all works out well for me. If not then I am just gonna have to find someone to sell me some real bh-5.