Hi "felinusz man. Got your PM man and the link and came and read the whole thread from beginning to the end and the few posts past yours here.
I am going to say exactly what I know and no more about this situation and will temper my personal biases against memory that requires so much voltage anyway. That said.
According to DFI-Street when changing J17 position you should do the following:
Changing from 3.3V to 5V: http://www.dfi-street.com/forum/sho...67&postcount=17
1. Go into Genie BIOS and manually set your vcore (cpu voltage). It MUST NOT be on Auto!
2. Save & exit. Return to bios upon reboot.
3. Verify that your vcore is correctly set. Save, exit, shut the rig down.
4. On the motherboard, move the 4v jumper in the upper corner of the board to the proper position (see manual/cd-manual/online manual for pics).
Boot up, go into Genie BIOS, do DRAM Voltage. Should now be adjustable above 3.2V.
Changing from 5V to 3.3V: http://www.dfi-street.com/forum/sho...305&postcount=4
1. boot into bios and lower vdimm to below 3.3v
save & exit
2. boot up and see that the voltage 'stuck' to the new setting correctly.
save & exit and shut down
3. change vdimm jumper back to 3.2v position.
4. boot up and make sure vdimm is where it is supposed to be in the bios.
keep in mind that sometimes the voltages for vdimm or cpu 'stick'. Its very random and rare but I have witnessed it myself. Its a quirk. Clear CMOS and the voltages will return to their defaults and can be reset.
Ok back on the 3.2Volts settings and Jp17 is on the 3.3Volt rail for voltage supply and now flash the bios above and use ONLY the Yellow slots.
Note: here! Do not really care about how fast you can now go. If there is going to be any testing done and it is> you need to quit feeding the dang memory into a slot you think is killing memory until such time as testing is done. That is all I can say about that for you who have seemed dead memory.
Now that should get most back into safe territory even if they go back to the 5V mode according the procedure for change as shown above.
Now I have not seen this but have a tester that has had recording Oscope on the Vdimm and has not seen a spike as of yet. No he has not tested 15 motherboards but he still has seen no voltage spike.
Oskar is aware of this thread and is looking deeply at the bios as I type this. I am now awaiting UTT as I have NONE ; mainly because I don't really like it but nevertheless am getting some for testing. When UTT arrives will setup and get running with Bios 510-2 as linked above and put sticks in Yellow slots (1 and 3) and start to running at High Volts and at as much FSB as it will do and let run prime or some other proggie to load the system and see if any problems surface from Yellow slots (1 and 3) and also to be able to test any beta bios Oskar hands me for test against his findings.
Now other than that I have no frikken clue and will not speculate. I have said all I know and will say. No guessing and jumping to any mind blast I may have come to me.
Thank you, "felinusz" for giving me heads up man. Did what I said I would do when you asked.
RGone...
Quote:
Originally Posted by felinusz
Has anyone tried contacting DFI about this, aside from the person who started a thread on the DFI forums (the thread that was locked)?
I just PMed RGone on DFI-Street about this thread, hopefully he'll come over here .
__________________
Back and forth from board to board it seems.
As for the cap, the ESR rating is also important - the best is to get caps that match those already on the board, or replace all of them. The problem is that if you put in any old cap, it may become more stressed than the others if it's ESR rating is lower than the others. BTW, a lower ESR is better for supply purposes. ESR = Equivalent Series Resistance.
Anyway, that's not why I posted here. I posted an important finding - it came at the end of page 6, then RGones post came so I don't think anyone has seen it. In a nutshell, Vdimm is not being set correctly on power up. http://www.xtremesystems.org/forums/...4&page=6&pp=25
__________________ 06-18-2005 According to RGone at DFI-Street: Cold boot problem? You need to email: DFI Direct
I have seen the theory on the SPD chips also from Uwackme on dfi-street
Ozsnoal I think is onto a different cold boot problem that I am not sure is related. On the other hand maybe it is related to unusual voltages (high or low) on boot-up.
I am watching this thread because I am running my VX on the 5V rail 24/7 @ 3.3V crunching BOINC (Climate Prediction) and don't want dead ram.
Maybe I'll do the 3.3V sense mod on my PSU and go back to the 3.3V rail after reading this thread.....................................
Working scenario:
Basically when you shutdown your PC and power it up again, Vdimm moves up from 0V to the voltage you set it to in the BIOS. (Measured with a multimeter).
Failing scenario:
HOWEVER, if you pull out the power plug, reinsert the plug, DO NOT switch on the PC yet. Hmmm, what is that 2.7V doing on Vdimm?? Switch on the PC hmmmm, still 2.7V? And of course you now have the cold boot problem. The Vdimm never reaches the value I have set in the BIOS.
FYIs...
The voltage on Vmem when you remove AC and then re-apply is caused by the way one of the ICs initializes at power up (along with how it's used). The actual value depends on your settings (it can come up as low as 2.45V).
The proper Vmem setting isn't set until after the BIOS is thru the POST sequence (when only 1 light remains lit)... so if your memory won't pass the "detect RAM" stage at the lower voltage, you get the "cold boot" problem.
It is a very easy fix in the BIOS (hey Oscar!)... the Vmem/Vcore configuration IC needs to be initialized before the detection phases of POST.
Now other than that I have no frikken clue and will not speculate. I have said all I know and will say. No guessing and jumping to any mind blast I may have come to me.
Thank you, "felinusz" for giving me heads up man. Did what I said I would do when you asked.
RGone, thanks a lot for checking this out dude!
Your take on this is appreciated, it's great to see good company representation like yours!
One think to add too though is that I've seen some dead Rams BUT their chips where OK......Their SPD chips were dead......After changing the SPD chip with another one, Ram was working perfect again.......Maybe the SPD chip can't handle the spikes or so and it's just give up
Quote:
Originally Posted by celemine1Gig
Hipro's theory of SPD-ROM chips dying sounds more reasonable to me. I had a look at my UTT sticks today and found out that the SPD-ROM chip used on mine ("TMC 24A02") is rated 2.5~5.5V. So on mine, there should theoretically be no problem with high Vdimm.
Just for fun:
Could anyone with "dead" RAM post the info printed on his SPD-ROM chip? Perhaps this way, we could find out quite fast, if the theory is any good.
Quote:
Originally Posted by Shroomalistic
Ive had the same problems but not on the dfi nf4. Mine was on the dfi ut 250gb. Killed 4 sticks of twinmos 3200 utt and 2 sticks of old school ch-5. I believe it was the spd chips also cause letting it sit for a day or 2 would let it run again but not for long. I could run memtest on them for days with no errors but as soon as I restarted they would boot till I pulled them out and let them sit for days in again. Now I have some newer twinmos utt and they have the larger spd chip in a different spot and havent had a problem with them. Maybe we should all check our dead sticks and see if they have the larger spd chip or the smaller one.
I still have my old bad TMII AA4T (CH-5/UTT) sticks and I have just received new TMSP 1A4T (BH-5/UTT) sticks so I decided to compare their SPD ROM Chips. As you can see they are different, both in sizes and in location. I hope this means something ... something good
So .... does OCZ Gold VX series got smaller or bigger SPD ROM chip?
This is how OCZ VX PC3200 on BrainPower B6U808 looked like when I bought it last year and took heatspreaders off - which now we know are Winbond UTT-CH5 I do not have those sticks anymore, though.
Just to give you some food for your brains
EDIT: It looks like UTT-CH5 based modules use smaller SPD chip compared to UTT-BH5 based sticks...
This is how OCZ VX PC3200 on BrainPower B6U808 looked like when I bought it last year and took heatspreaders off - which now we know are Winbond UTT-CH5 I do not have those sticks anymore, though.
The voltage on Vmem when you remove AC and then re-apply is caused by the way one of the ICs initializes at power up (along with how it's used). The actual value depends on your settings (it can come up as low as 2.45V).
Sure, doesn't mean that I agree with their HW design. It is simple enough to have circuitry that keeps components in states you want them to be in - without having software to actually do it. Obviously the shutdown code does it - there is 0V on Vdimm. There should just as well be 0V on Vdimm when you plug the AC cord in the back of the machine.
Anwyay, I hope DFI look into turning the Vdimm voltage up to the CMOS specified value before trying to detect the RAM.
__________________ 06-18-2005 According to RGone at DFI-Street: Cold boot problem? You need to email: DFI Direct
.
Linear Mode
