Jump to content
Sign in to follow this  
Zoppa

DFI lanparty ut nf3 250gb coldboot problems

Recommended Posts

When running my system overclocked it runs fine at 8x280, vcore 1.5, vagp 1.6, chipset 1.8, vmem 2.7 and mem divider att 133 (not tried to clock the ram and 140, 150 divider seems a bit bugged), primestable (small fft) for atleast 1.5h

 

If i turn the system of for a couple of mins and then turn it on it fail to boot, usualy stops at detecting raid array (not using raid and dont see this when it boots fine) or before.

After a couple of resets it usualy boots fine again.

 

It apears that the longer the system is turned of the harder it is to boot. If i leave it over the night i have to go into bios and revert to stock settings, boot and then oc again and it works fine.

 

I first suspect my PSU, its low on 12v, only 16A so i have borrowed a brand new antec smartpower 400w with dual 12v rails, think it is 14A and 15A and this psu seems to have a bit more stable voltages. But still got the coldboot issues.

I used to run the most recent bios (the one in sig) and are now running 05/04/05 but that ain't helping much, still coldboot issues.

 

I flashed like this with tmods cd

load optimized in bios

boot and flash

load optimized in bios

clear cmos

load optimized

set everything right in bios

test boot to windows

reboot and set oc in bios

test boot windows

 

I have set everything memory related manualy in bios, but that diden't help either.

Running out of things to try, anyone got any ideas what to do??

Share this post


Link to post
Share on other sites

stops at detecting raid array= usealy means a memory problem.or that your pc is very unstable. 1.5 h of prime is not going to do anything you have to be stable for 8 hours of prime

 

Just al little Checklist of what I had to do, to get so far:

 

- Turn your cpu temp control to shut down at 60c

- DO NOT USE SATA 1+2 (Cant get over 230MHz)

- raised V-Chip set to 1.8v when it stops passing go to 1.9( it is important to use a after market sink and fan)

- LDT-Multi down to 3x, Till you hit 2550 then 2.5

- Try disabled Command per clock, so RAM running at -2T) Just to get memory out of the way (Only if your having memory problems)

- raised AGP-clock to 67 MHz (seems to be unlocked at 66MHz, sometime 68MHz, too)

- V-AGP raised to 1,7V

- disable Cool'n'Quiet

- disable Thermal Throttling

 

run memtest 5 & 8 every time you make a change in your bios

 

Caution when making changes in your bios I use the strike three your out rule. If you don’t make it into your os the second time go back were you know you can and boot the os. I have never corrupted or lost an os using this

 

Takes most people 1.696 vcore to hit 2.6xx on a E6 NO one has had any luck past 1.8

If my CPU is running 50c and above I’m not stable

I have run stable at 8x338 with stock cpu cooler

If you use any of this Please use caution and watch your temps close

Share this post


Link to post
Share on other sites

- Turn your cpu temp control to shut down at 60c

Have had this all time, my temp never go over 40c

 

- DO NOT USE SATA 1+2 (Cant get over 230MHz)

Using SATA3, 1+2 disabled in bios

 

- raised V-Chip set to 1.8v when it stops passing go to 1.9( it is important to use a after market sink and fan)

Raised to 1.8, got more stable at over 300htt, at 280 that im running now i dosen't seem to be any problem running at 1.7 but i go with 1.8 just in case. got a fan mounted on the heatsink, now it dosen't get warm

 

- LDT-Multi down to 3x, Till you hit 2550 then 2.5

Yes, im running like that at 3x now hen att 2240

 

- Try disabled Command per clock, so RAM running at -2T) Just to get memory out of the way (Only if your having memory problems)

Tried that to, dosen't seem to help

 

- raised AGP-clock to 67 MHz (seems to be unlocked at 66MHz, sometime 68MHz, too)

Always do that

 

- V-AGP raised to 1,7V

shall try that, currently at 1.6

 

- disable Cool'n'Quiet

it's disabled, dont work on my cpu anyways

 

- disable Thermal Throttling

tried with standard 50% and disabled, seems it won't help

 

Temps are not an issue on my setup (yet, stockcooler and high vcore is not that fun with stockcooler, loadtemps at around 43 with 1.6+ vcore)

Share this post


Link to post
Share on other sites

Upped chipset to 1.9 and vagp to 1.7

Turned it of for a couple of mins and now it booted fine :)

 

Edit: Turn it of for a little bit longer and got coldbot issue again, this drives me crazy

shall try a very small oc with vcore at 1.5

 

Gonna prime for a while just to see if chipset heatsink gets varm (can turn up the fan if i need, got i running at around 1800 rpm now, slowed it down to 5v or 7v, dont remember, and slowed it some more with smartguardian.

 

When i solve the coldboot problems im gonna prime for atleast 8h and start clocking the memory, running below spec speed at the moment just to get the mem out of the way

 

Can upping the vagp damage the graphics card?

Never heard of that anyway, just curios if anyone know.

Share this post


Link to post
Share on other sites

Update:

Appears my computer don't like running memory on dividers.

 

Losend timings and ran 1:1 at 8x240, mem at 2.8 and vcore 2.75, chipset 1.7, vagp 1.6

primed 2.5h and was stable, this morning i had no coldboot problems.

 

I have upped to 8x260 now and ran some memtest 5 & 8 without errors and are now running prime again, gonna let i run for atleast 4h and then shut off to see if coldboot issues are gone.

 

Got those at 8x260 when running with divider.

 

Now i just have to keep my fingers crossed and hope for the best

Share this post


Link to post
Share on other sites

When I was running my NF3 at 10x260 with a S754 3700, I would sometimes hang at the detecting RAID screen. I was running a RAID at the time. The set-up was 12 + hours Prime95 stable, temps were under control ( I had a TT BT on the cpu) and my memory was stable. About every 5th cold boot or so, it would start up, post, get to the detecting RAID screen and then do it all over again, but this time it would load windows and all would be fine. I searched the web to figure out the problem, it seems to be a quirck of the board. Earlier on in the boards history, reading posts by Oskar Wu, the board had a problem setting up start up voltages when running a high HTT and low cpu multiplier. Later BIOS revisions seemed to correct the problem. I was running the latest official BIOS, the one before support for the Newark core cpu's memory controller. I had followed all of MSMolts's previous suggestions, then just accepted this as a quirck of the board. Some boards have it worse than others. Mine was acceptable. One thing that helps the board in setting start-up voltages correctly, is to get all fans off the motherboard headers. I only ran the TT BT on the motherboard header. I ran the board for about a year, my daughter now has it in her pc. The board was an excellent overclocker. I got what I wanted out of it.

Share this post


Link to post
Share on other sites

That is good to know.

 

Know i read about the fanheaders somewhere.

 

Im currently using all. Guess i can run the 92mm memory cooling fan on molex instead if i get an adapter or have time to solder. I added a fan to the chipset, but i like to have that on header so i can slow it down, already slowed it to 5 or 7V but need to slow it more with smartguardian.

 

5h primestable now and no coldboot problems, time to clock it some more but guess i will need dividers pretty soon.

Share this post


Link to post
Share on other sites

Haven't had a single coldboot problem since i stoped using low dividers.

Currently running 8x280 with 180 divider.

 

 

Tried change Max async latency and Read preamble. upped them booth first to 11 8 ....no luck, no boot just beep :( clearcmos and set everything right, tried upp booth by 1, same thing beep and no boot. Shall try some more when i got time, just upping one of them to see what happens

Share this post


Link to post
Share on other sites

I attached a screenshot of a64 tweaker with memtimings.

These are good for atleast 260MHz, tried 270 but got errors in memtest.

Tried losening all timings that are currently 3 to 4 exept cas latency with i leave at 3, but that diden't help. Upping to 3v diden't make any change either

 

The one thing i noticed i have forgot is trfc, i got this on 14 and that is default.

Shall try upping it later when i got time, my memsettings are not optimized, just upped every 2 to 3 and Tras from 8 to 10 so i could test to run 1:1 and figure out my coldboot problems, like i said im on 180 divider now and no problems so far.

 

Haven't primed for long with current settings, about 2h shall do that later when i got time or after upping htt some more :)

Share this post


Link to post
Share on other sites

Zoppa when you get done beating your self up trying to run your ram at cas 2 and start running it at 3 instead and have some fun let me know because your wasting your time and when you start overclocking it your going to be at cas 3 anyways

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
Sign in to follow this  

×