Killed my BIOS chip on DFI NF4 - why??? Made no mistake........

[HARDCORECLOCKER]

:confused: O.K. - Here's the story:

After some time with my BH-5 I decided to go for another try with my TCCD.

Shut down the rig, unplugged the power cord and waited till all onboard lites were off.

Swiched the jumper to the position for 3,3V rail and replaced my BH-5 with the TCCD.

Cleared CMOS and started the system - long beeps, a constantly blinking power lite on front panel and 3 red onboard lites at the diode area.

Hm, rebuilt all like it was before but still no POST. Then tried 3 kits of RAM, long CMOS clear, removed the battery, unplugged all USB connections and all the other tricks but still no go. :mad:

Fortunately I have a second BIOS chip - threw it in and ooooooops, rig is running again !!!

Hm again - gave the first BIOS chip a second try - but again no go, definately killed. But why?

Replacing the RAM can not kill BIOS and when I swiched the jumper for the voltage rails there was no power to the system but the battery.

Never heard that it is recommended to remove the battery before setting the jumper........ :confused:

Conclusions:

1) Any ideas what happened here?

2) All You people with 3 red LEDs on and spinnin' fans but no boot try a second BIOS chip.

:toast:

[MadMikeSS]

Thanks for the heads up, unfortunately I doubt that most of us have a second bios chip floating around. :)

[IvanAndreevich]

You can always hotflash

[Skip]

yea just hotflash so you have that bios chip incase the problem happens again. i would suggest hotflashing as soon as possible.

[BLHealthy4life]

what is hotflash?

[Rezag1000]

what is hotflash?

I don't know, something life-saving perhaps?
:confused:

[MadMikeSS]

what is hotflash?

Something that women have when they go through menopause?

[NST6563]

my bios got corrupt when I changed some of the mem timings too low. luckily the bootblock wasn't damaged, but it was all buggered up for a while. At first I just got the beeps, then turned it off. Turned it back on and it came up to the bootblock saying checksum error, please insert floppy... turned it back off and walked away pissed (and to find a floppy drive). Came back and it powered on to the Lanparty splash screen as if it'd just had the cmos cleared. Went into bios and there were all kinds of droppings in the menus and chopped text and whatnot...so I promptly re-flashed and the problem went away.

[LowRun]

DFI and bios death, two good old friends :rolleyes:

Bios savior fever ahead :D

[conrad.maranan]

There are too many threads open at the moment dealing with DFI problems. We only need one active thread since most of the issues being discussed are already being discussed in another thread.

Please refrain from starting up threads on the same or similar topic from this point forward. In the future, multiple threads on relatively similar issues will be consolidated into one thread.

Please refer to the AMD section Guidelines (http://www.xtremesystems.org/forums/showthread.php?t=58976) for advice and suggestions on how to initiate and participate in discussions.

Thanks. ;)

[Theo404]

I totaly respect what your saying conrad, but I believe HCC is bringing in a new concept which would have probably been lost in the 30 odd page DFI thread. I agree compleatly that there is a trend arising to post willy-nilly in the AMD section, but I do think this has worth outside of the huge DFI thread. Do you enjoy trawling through 18 pages of posts looking for little gems like this? :confused:

[HARDCORECLOCKER]

I totaly respect what your saying conrad, but I believe HCC is bringing in a new concept which would have probably been lost in the 30 odd page DFI thread. I agree compleatly that there is a trend arising to post willy-nilly in the AMD section, but I do think this has worth outside of the huge DFI thread. Do you enjoy trawling through 18 pages of posts looking for little gems like this? :confused:

:D Subscribed - THX. That was exactly the reason I started a new thread.

:toast:

[icehot]

i feel worried about changing memory now if it can cause problems :( already fed up of computers atm, took me a week to sort out my dfi ultra-d :(

[Theo404]

Might be an idea to try and fully discharge the components before messing with them, I read in another thread (the name escapes me) that you should pull your power supply cable, wait a few seconds then press the power button (on the PC). Makes sense to me because I tried it and the leds came on for a second and the fans moved a little, suggesting there was still stored current somewhere.

[HARDCORECLOCKER]

i feel worried about changing memory now if it can cause problems :( already fed up of computers atm, took me a week to sort out my dfi ultra-d :(

;) Nah - think it was not caused by changing RAMs - has to do somethin' with the Vdimm jumper in my opinion.......

:toast:

[Theo404]

Ya, far more likley it was a surge or shock when you switched the vdimm regulating curcuit over to the 5v line.

[Skip]

it could just be a fluke or something?

[Perc]

Might be an idea to try and fully discharge the components before messing with them, I read in another thread (the name escapes me) that you should pull your power supply cable, wait a few seconds then press the power button (on the PC). Makes sense to me because I tried it and the leds came on for a second and the fans moved a little, suggesting there was still stored current somewhere.

yeah you are right. i always unplug the system flip the psu's switch and wait for all the lights to go off on the mb before even sticking my hand in the case! i would be willing to be that it was some sort of surge that caused the problem not just changing the memory.

peace perc,

[Dumo]

yeah you are right. i always unplug the system flip the psu's switch and wait for all the lights to go off on the mb before even sticking my hand in the case! i would be willing to be that it was some sort of surge that caused the problem not just changing the memory.

peace perc,Absolutely...the best precaution. I think OPB posted the warning way back....

[mad mikee]

1. Oskar said awhile ago that yeah you can do bad things if not COMPLETELY discharged so maybe a factor.

2. Try hotflashing (boot, pull working chip, and replace w/ one you wanna flash , then flash the 'new' chip) including the boot block (just to be sure) the 'bad chip and see if that brings it back.

3. Thinking it was prolly 5v jumper since I have swapped memory in this board literally hundreds of time w/o issue (other than occasionaly loosing a cap off the dimm :upset: )

[HARDCORECLOCKER]

yeah you are right. i always unplug the system flip the psu's switch and wait for all the lights to go off on the mb before even sticking my hand in the case! i would be willing to be that it was some sort of surge that caused the problem not just changing the memory.

peace perc,

:D Hey folks I know that and did it, turned off sys, swiched off the PSU and unplugged the power cord, waited for 2 minutes and all the lites were out of course.......

:toast:

[Theo404]

I dont think anybody actualy got what I said. :D

Of course you have to power down the system and pull the PSU chord, everyone does that. The point I was trying to make is to push the power button on the system after you've done that! Give it a try. on my system the leds came on for a few seconds....ie. still juice in it.

[HARDCORECLOCKER]

I dont think anybody actualy got what I said. :D

Of course you have to power down the system and pull the PSU chord, everyone does that. The point I was trying to make is to push the power button on the system after you've done that! Give it a try. on my system the leds came on for a few seconds....ie. still juice in it.

:D Now I see Theo - sorry for misunderstanding. Will do that next time. THX.

:toast:

[|SiLA|]

it happened to my friend to 2 weeks ago..he was benching with his new geil ultra-x(bh5) then since it was sucking bad he changed back the rams with the gskill..and suddenly he couldn't boot 3 leds on(it means can't "find" the ram FYI) told him maybe it was the mem controller died..then i also told him to use single channel using the second orange slot..then worked..then worked also in DC :)

dunno where the explanation is on why happened but it's scary :D :stick:

[vlad]

Something that women have when they go through menopause?
:ROTF:

Sorry about that.

So far when ever i've made changes i've unpluged everything and discharged and i havent had any problems. "Knocks on wood"

[The Mofo]

This is one of the reasons why i use a BIOS Savior. Do you think the corruption happened after you forced a CMOS Reset after you moved the jumper over?

[HARDCORECLOCKER]

This is one of the reasons why i use a BIOS Savior. Do you think the corruption happened after you forced a CMOS Reset after you moved the jumper over?

:D Yep - because using BH-5 on 5V rail before the Vdimm was set to 3,72 when I turned the rig off.

Then I swiched the jumper to the 3,3V rail and remebered my last setting 3,72. That was the reason I cleared the CMOS to prevent the TCCD from gettin' 3.2 volts and within that all somethin' went wrong.

:toast:

[SuperFreak]

I hope to get my RMA dfi back this week. I came home one day from work and found any load on my cpu caused it to crash. I went into the bios to check this out and the bios was garbled up. I tried to flash it but each time I would flash and save optimal setting and go back in to make chages it would garble up afain tried a few different bios - same thing until it hung up during a flash. then I got the 3 rid lights and the power light on the front of case flashing and it would freeze before it boot or run a floppy.

[The Mofo]

I wonder if forcing the CMOS reset fubar'd everything up.. Ive had a similar problem once and traced it back to not pulling the power connector from the psu/mb out before forcing cmos reset.

[EMC2]

--- HCC ---

Did you try to reflash the "bad" BIOS chip after getting things working?


I would agree and "preach" to always push the pwr button a couple times after disconnecting pwr, before changing any hardware.

The deal with the mem change over... use those multiple settings stores ("BIOS reload" screen) to store good working settings and a "default" super slow setting (like 2.7 Vmem, 1.4 Vcore, 100Mhz FSB, 6x multi, 2.5,4,4,8 mem) for change overs and recover mode. Assign function keys to them. Then before changing mem types out, go into the BIOS and do a "load from this setting" the super slow setting to insure you can boot up afterwards.

Also, turn off the "auto save last good boot" in the same menu immediately after you have booted the "super slow" setting. That way your default "safe setting" is sure to boot anything safely, and the CMOS won't get written to during a bad OC setting boot. You can also select the "safe setting" with a jumper without having to clear the CMOS supposedly (say supposedly because the only time I tried it was before I had my CMOS set up as described, lol).

Peace :toast:

[Theo404]

So do you think that the 'auto save last good boot' feature could be the cause/final straw of the numerous DFI SDS? I say 'the last straw' because I guess if the board starts to fail, becoming more and more unstable on each boot, once it gets criminaly unstable then it tries to rewrite 'good' settings on a bad boot, and because the system is unstable it borks the bios. Interesting....

Edit: To elaborate, could this feature also cause compleatly sudden death, ie. user trys highly overclocked settings, bios panics due to instability and trys to 'flash' good settings and borks itself??

[GoldenTiger]

I too use a BIOS Savior... ever since my Abit AN7 had a corrupted BIOS and I had no way to restore it without RMA'ing the board at the time, I bought one for my main rig. For $20-30, and the fact that it can be used with newer motherboards (the model I have works fine on a DFI Lanparty NF4-Ultra D too, as they both use the same BIOS chip capacities/types), it's a very good investment. When overclocking, things can go wrong quickly with the wrong settings regarding the BIOS... it's a good backup.

[i found nemo]

yea i do that all the time (turn power on with cord undone) works like a charm

[HARDCORECLOCKER]

--- HCC ---

Did you try to reflash the "bad" BIOS chip after getting things working?


I would agree and "preach" to always push the pwr button a couple times after disconnecting pwr, before changing any hardware.

The deal with the mem change over... use those multiple settings stores ("BIOS reload" screen) to store good working settings and a "default" super slow setting (like 2.7 Vmem, 1.4 Vcore, 100Mhz FSB, 6x multi, 2.5,4,4,8 mem) for change overs and recover mode. Assign function keys to them. Then before changing mem types out, go into the BIOS and do a "load from this setting" the super slow setting to insure you can boot up afterwards.

Also, turn off the "auto save last good boot" in the same menu immediately after you have booted the "super slow" setting. That way your default "safe setting" is sure to boot anything safely, and the CMOS won't get written to during a bad OC setting boot. You can also select the "safe setting" with a jumper without having to clear the CMOS supposedly (say supposedly because the only time I tried it was before I had my CMOS set up as described, lol).

Peace :toast:

:D No - did not try a hotflash, if somethin' goes wrong I have no BIOS anymore, ordered a new BIOS chip for 15 bucks, not really expensive.

I always set manually save settings before changing anythin' - here not the problem because I often replaced RAMs before. This issue was caused by some bad voltage I think, has to do somethin' with the Vdimm jumper.

It doesent matter if I push the start button after turning off the rig, there is the MACH II GT within the boot circiut and only the MACH gives a signal to the board, so it only would cause that the MACH is free from electricity....

:toast:

[mrlobber]

HCC, I think we need to get to know each other better...

... because now we're both on the same boat :D

Just made approximately the same mistake.

Wanted to see how some TCCD borrowed from a friend clock on my DFI.

Vdimm for my VX was set to 3.4 with the 5v jumper, of course.

Loaded optimized defaults first, though.

Took VX out, put TCCD in, moved the jumper... and cleared cmos simultaneously as well.

A note for others: power cord was unplugged and power switch was pressed at least for a couple of times to remove any excess power remaining in the circuits.

Pulled everything back, pressed the power button...

... BANG! 3 red lights and nothing. NO POST.

I guess, I need to refresh my hotflashing skills... my nF2 rig is still running :D

[HARDCORECLOCKER]

HCC, I think we need to get to know each other better...

... because now we're both on the same boat :D

Just made approximately the same mistake.

Wanted to see how some TCCD borrowed from a friend clock on my DFI.

Vdimm for my VX was set to 3.4 with the 5v jumper, of course.

Loaded optimized defaults first, though.

Took VX out, put TCCD in, moved the jumper... and cleared cmos simultaneously as well.

A note for others: power cord was unplugged and power switch was pressed at least for a couple of times to remove any excess power remaining in the circuits.

Pulled everything back, pressed the power button...

... BANG! 3 red lights and nothing. NO POST.

I guess, I need to refresh my hotflashing skills... my nF2 rig is still running :D

:banana: Ha - I'm not the only one !!! Bad luck often comes twice - though it hit two different people this time........ :p:

So what do You think about it? Has somethin' to do with the jumper, heh ?

Maybe we should report OSKAR - not all user have a second BIOS chip like myself or a second rig to do a hotflash....... :slapass:

BTW within all that trouble I noticed that there even is no use to move the jumper back to the 3.3V rail - I had the same temps with 2,8 Vdimm on 3.3V as on 5V rail, but have 2 x 120mm fans blowing straight to that area....



:toast:

[mrlobber]

So what do You think about it? Has somethin' to do with the jumper, heh ?

Maybe we should report OSKAR - not all user have a second BIOS chip like myself or a second rig to do a hotflash....... :slapass:

Yeah, definitely the jumper is involved somehow. All I know is Oskar warned that when changing the jumper, you should act almost as when clearing bios - with all that power unplugging and so on. Obviously, there is a connection between the two (naturally, should be so that the board knew when it has the high Vdimm available), but it has some nasty backdoors and these links are getting opened if they (bios and the jumper) are reset simultaneously...

The most stupid thing is, I read this thread already, but somehow thought it doesn't apply to me :p: :stick: At least now we have 2 cases proving a theory :D

[HARDCORECLOCKER]

Yeah, definitely the jumper is involved somehow. All I know is Oskar warned that when changing the jumper, you should act almost as when clearing bios - with all that power unplugging and so on. Obviously, there is a connection between the two (naturally, should be so that the board knew when it has the high Vdimm available), but it has some nasty backdoors and these links are getting opened if they (bios and the jumper) are reset simultaneously...
The most stupid thing is, I read this thread already, but somehow thought it doesn't apply to me :p: :stick: At least now we have 2 cases proving a theory :D

;) Yep - think that's the point here. Sorry to hear You even read my thread before it happened to You.......... :(

BTW...., never ever change Your avatar - I love it !!!

:toast:

[The Mofo]

Maybe if you didnt force a clear on the BIOS, it wouldnt cause the corruption? Only thing i can think of is there is some kind of voltage spiking the BIOS chip when you force a cmos clear.

How come everyone is moving the +5V jumper back over to default? You dont have to crank it up if you dont have to.

[Unrealcpu]

This is pretty funny how everyone is RMAing their DFI boards this week. Mine died on me last week LOL

GO DFI

[c7775]

its obvious because youre too hardcore :nono:

[mrlobber]

its obvious because youre too hardcore :nono:

Never thought of myself that way :D

[mrlobber]

Ok, update on my situation:

my DFI is half dead.

That is, it posts with memory in memory slots 1 or 2, and it doesnt post with memory in slots 3 or 4 (closest to the cpu). Tested with 3 different memories (VX, BH-5, TCCD), 2 bios chips (beta 3.26 and official 3.10), both Vdimm jumper configurations.

Apparently, it seems that something in the Vdimm circuit is damaged.

So long for this DFI. Was a pretty good board for me. Time for another one...

[HARDCORECLOCKER]

:D Also update: Got a new 2nd BIOS chip now, who knows whal will happen in future........ :stick:

:toast:

[computerpro3]

:confused: O.K. - Here's the story:

After some time with my BH-5 I decided to go for another try with my TCCD.

Shut down the rig, unplugged the power cord and waited till all onboard lites were off.

Swiched the jumper to the position for 3,3V rail and replaced my BH-5 with the TCCD.

Cleared CMOS and started the system - long beeps, a constantly blinking power lite on front panel and 3 red onboard lites at the diode area.

Hm, rebuilt all like it was before but still no POST. Then tried 3 kits of RAM, long CMOS clear, removed the battery, unplugged all USB connections and all the other tricks but still no go. :mad:

Fortunately I have a second BIOS chip - threw it in and ooooooops, rig is running again !!!

Hm again - gave the first BIOS chip a second try - but again no go, definately killed. But why?

Replacing the RAM can not kill BIOS and when I swiched the jumper for the voltage rails there was no power to the system but the battery.

Never heard that it is recommended to remove the battery before setting the jumper........ :confused:

Conclusions:

1) Any ideas what happened here?

2) All You people with 3 red LEDs on and spinnin' fans but no boot try a second BIOS chip.

:toast:


I"ve been looking into this, as it happened to me on 2 seperate occasions. First time, brand new build using Mushkin TCCD 1gb pack. The good stuff, lvl2 black. Board failed memtest, next day bios chip died. Or so I thought. I rma'd my memory too and scored a deal on bh5 to hold me over on RMA. reflahsed bios, bh5 works flawlessley with it. Tried TCCD again, failed memtest. I didnt leave it in long enough to kill the bios.

Second time, building a system for a friend who already had mushkin tccd, same memory as me. I warned him about the bios thing, he wanted to try it because he didnt wana get new memory. Worked fine for 3 days, then became windows unstable, still passed memtest though. Then got the bios checksum error, no amount of reflashing fixes it. Tried my bh5, reflashed, works fine. Hes going to buy bh5, but in the mean time, I haev him 100% stable for a week now with a SINGLE stick of 512mb mushkin tccd.

You also had tccd, and do a search, 99% of the people that killed their dfi's had TCCD. Interestingly enough, this boards bios was specially optimized from the get-go to run TCCD super fast; I believe oskar or bigtoe said that (dunno who, it was someone that knows their bios...). My theory is that one of these optimizations is simply running too aggressive for some of the TCCD memory out there, and corrupting the bios as a result.

Seriously, do a search, its scary how many people have the dead bios/tccd combo. I don't know why or even if memory could corrupt bios, but I'd put money on there being SOME kind of link there.

[HARDCORECLOCKER]

:D Indeed very interesting facts here - had some problems with TCCD right from the start with this board - booting into windows at 10 x 334 1:1 but no matter what settings I tried - didn't get it stable.

And the BH-5 running like charme......... :p:

:toast:

[computerpro3]

more from the 3400 @ 3.2ghz thread.....

we corrupted our bench drive right away, then we had alot of system problems. we thought the mobo was dead for over an hour. tried everything.. then eshbach threw in one of my sticks of bh5 and BOOM! it worked! put his ram back and it fired right up!

in the end we learned alot, but didnt run any 3d.

[quicksilverXP]

Not me... BH-5 is running horribly with mine with OCZ's new BH-5. I've done EVERYTHING anyone else in this thread would have done.

I first tried seeing how far I would get without enabling the 4v option, and I only got to 235 Mhz. So I did usual (turn it off, pull plug, wait a minute, clear cmos, change jumpers to 4v and then restart). When it restarts, it just turns on for about 10 seconds, then shuts off and the vdimm light goes off. So I do the same thing, and it finally restarts. I test with Memtest only at 200 Mhz and at 3.3v (just to check for instability at stock speeds), and I get Memtest errors galore.

Right now I am trying different bioses and timings and nothing seems to help. Odd... I never had this problem with VX's.

[computerpro3]

now that sounds like bad memory, I havent heard of any cases of limitied overclocking with memtest errors, hell, my TCCD was superpi stable at 280mhz before it went. Check it in another machine

[mrlobber]

Hm, some news kinda on this topic by OPB here:

http://www.xtremesystems.org/forums/showthread.php?t=59414

[situman]

could all these DFI problems be related to it being incompatible with Windows XP SP2?

http://www6.tomshardware.com/column/20050414/index.html

Apparently it can cause BSOD and in worse case scenerio, causes the system not to boot in general.

[LowRun]

Site is down at the moment, could it be that their web server is hosted on DFI + XP SP2 rig :p:

[Theo404]

That article, like so many over at 'toms hardware' is less than, how should I say it tactfully, the gospel truth.

As far as non-overclocking software and causing damage to hardware is concerned... :slap:

Back pedal a few posts and catch the link to a OPB thread about a known issue with is highly likely to be the explanation to this case.

[EMC2]

You also had tccd, and do a search, 99% of the people that killed their dfi's had TCCD. Interestingly enough, this boards bios was specially optimized from the get-go to run TCCD super fast; I believe oskar or bigtoe said that (dunno who, it was someone that knows their bios...). My theory is that one of these optimizations is simply running too aggressive for some of the TCCD memory out there, and corrupting the bios as a result.

Seriously, do a search, its scary how many people have the dead bios/tccd combo. I don't know why or even if memory could corrupt bios, but I'd put money on there being SOME kind of link there.

Just another data point...several weeks of abuse with TCCD and no problems of any kind. But then, I do not leave any memory timings on Auto and I have the "Auto store last bootable configuration to CMOS" disabled ;)

[Fzero]

would something like this help prevent this from happening?
http://www.ioss.com.tw/web/English/RD1BIOSSavior/SelectionChart/PLCCTYPE/RD1PL.html

[cadaveca]

Why are you guys not pulling the battery? the battery is what keeps cmos settings intact over prolonged periods without power...simply placing the jumper to clear cmos is not enough!! The battery still provides power to save the cmos!! You clear the CMOS by draining the power!! This is why you unplug the power as well!! No wonder you killed your bios!! In the 15 years i have been building machines, I never heard of NOT removing the battery...it's called the CMOS battery!! PLEASE PLEASE PLEASE REMOVE THE BATTERY!!

[mrlobber]

Why are you guys not pulling the battery? the battery is what keeps cmos settings intact over prolonged periods without power...simply placing the jumper to clear cmos is not enough!! The battery still provides power to save the cmos!! You clear the CMOS by draining the power!! This is why you unplug the power as well!! No wonder you killed your bios!! In the 15 years i have been building machines, I never heard of NOT removing the battery...it's called the CMOS battery!! PLEASE PLEASE PLEASE REMOVE THE BATTERY!!

I did it. In fact, I always do it since DFI nForce2. Yet it didn't prevent the board from somehow damaging 2 closest ram slots to the cpu. And yes, I tested, the other slots work fine up to 270htt at least.

[cadaveca]

I did it. In fact, I always do it since DFI nForce2. Yet it didn't prevent the board from somehow damaging 2 closest ram slots to the cpu. And yes, I tested, the other slots work fine up to 270htt at least.


yeah , but you did not kill the bios. there must be other issues. If you kill your bios by not properly resetting your CMOS... :stick: It says to remove the battery in the manual, does it not?

[Mastakilla]

while trial and erroring with memsettings in the bios, i had about 3-4 times that the pc didnt boot anymore...

i just unplugged power, and switched the reset bios jumper...

i remember it didnt allways work from the first time (had to switch that jumper a few (2-3) times or something?)

but eventually it allways worked....

and that was without unplugging any battery

is it dangerous to do it like this then?

[situman]

Odd, I NEVER had to remove the battery to clear cmos on any of my Abits. Heck I never even had to unplug the power cable in the back. Just flip the switch and hit the power button a few times. DFIs sure are fragile.

[vlad]

What i find funny is both the Infinity Ultra & Ultra D killed bios chips on me but yet My UT Never gave me any problem what so ever. Never should have sold it. Well at least i have a few chips left over from my Infinity. ;)