Low +5V rail?

[GiANT]

I noticed in my bios today that it lists my +5VDC as floating between 4.59V-4.62V.

I believe the low limit is something like 4.75V. Is this correct?

What runs off the +5V line?

I have been having some crashes, lockups, and colors smeared all the way down my screen lately. Usually when running 3D stuff.

Thought it was my new AIW X800XT cause it runs about 80C under load. But I have also recently gotten some OCZ6001024EEPE-K, so I took it (my new OCZ memory) out and put my old mem in. I couldnt get the computer to lock. I had sachs marine aquarium running as a wallpaper, opened ATI Tool over that(fuzzy cube), Started Prime95 torture test max heat, MBM5, Scanned for viruses. I let that all run for an hour or so with no problems at all. Plus, my old memory is generic DDR333, and I had it running at DDR400. No problems after more than an hour. I put the NEW OCZ memory back in and was able to get a hard lock within minutes. I opened the aquarium screensaver and had it running as wallpaper. Got ati tool open and CRASH!!

EDIT: Also I tried loosening the Timings on the OCZ from DDR400 2-2-2-5 to as loose as 3-4-4-9 and had similar results. Running the mem at 2.8V. Its the hightst my mobo will go.

So I want to know if I need to RMA my memory, get a new power supply, or both. The power supply is an ANTEC 400W. And using the power calculator mentioned in another thread I calculated my maximum power under 100% load at 326W.

[RyderOCZ]

What motherboard are we working with here?

What speeds and timings are you running on the ram?

I would not believe the 5V reading personally.

[GiANT]

MotherBoard = ASUS A7N8X-E-Delux rev2

Stock speed on memory DDR400 2-2-2-5 and tried WAY lose timings for this ram DDR400 3-4-4-9 still got lockups. And as I said, I had my generic DDR333 running at DDR400 2.5-3-3-7 without a single problem

The bois and also just looked in MBM5 and teh voltages are listed as:

12V=12.09
5V=4.52
3.3V=3.21

This power supply is not new. Its a few years old.

[RyderOCZ]

Ram needs to be in the slot "by itself" on the board and the next one over.

2.7V is just perfect.

2-2-2-5 is pretty tight for this board normally the NF2 doesn't like CAS 2..so try 2.5-3-2-10 or 2.5-2-2-10.

Ram speed is set to 100% yes?

I will have to look at this again in the morning.... 1:30am here and the day is over...headed to bed.

[GiANT]

OK I put OCZ ram back in.

Placed in slots A1 and B1. Was using A2 and B1 for some damn reason, but that is still a duel channel configuration. Anyways...moving on.

Set to 2.7V. DDR400 2.5-3-2-10 1:1 ration with CPU.

I will let these 3D apps run. Here are my thoughts:

My generic DDR333 CHEAP CHEAP ram was running stable at DDR400 2.5-3-3-6. So if the OCZ cant do better than that I thinks its a problem. I didnt spend $200 to run at DDR400 2.5-3-3-10. Not 10...

I will run some apps overnight and if its still running i the morning i'll tighten up the timings. Something like DDR400 2.5-3-2-6. If its a board issue then that is fine cause these are lifetime gaurontee. I may also take these and drop them into a nForce4 PCI-e board at my cousins and see just what these can handle.

I will post a yea or nay tomorrow. Thanks Ryder.:thumbs:

ohh yea I almost forgot. After putting the OCZ back in I looked at MBM5 and the voltages were listed at

12V = 11.55
5V = 4.44V to 4.46V
3.3V = 3.14V to 3.18V
Im getting worried...hehe

[GiANT]

I tightened up the timings to DDR400 2.5-3-2-6 and ran several apps all night without any problems last night. This morning I tightened the timings to DDR400 2.5-3-2-5 and got colors smeared down the monitor and hard lock within a couple minutes.

[RyderOCZ]

Ok....10 is best on NF2 for Tras...that is why I suggested it....in Sandra bench, etc it shows more bandwidth.....don't worry that it is not tight timings...that is not always best.

Run 2.5-3-3-10 and up the FSB some....do you have an unlocked CPU?

I used to be able to run 250 FSB with my Platinum Rev 2 sticks on this board with 2.5-3-3-10 timings....much better than 200 2-2-2

[GiANT]

I ran DDR400 @ 2.5-3-2-6 for a few hours today and got another lockup.

I do not have an unlocked CPU. Its an AMD XP2500+ Barton. From like week 3 of 2004. I wanted to unlock but as far as I know it is not possible to unlock this CPU. If you know different for barton cores after week 39 of 2003 you let me know.

I will reset the timings to 2.5-3-3-10. Not sure how hight I can go with the FSB. Only about 230 or less if I want to keep the ratio 1:1 with CPU. The CPU multiplier is locked at 11 x 230 = 2.53Ghz and thats pretty high for a 1.8Ghz CPU. Should I try running memory at more than 100% of CPU FSB?:shhh:

EDIT: if your interested I rand sisoft and got mem bandwidth of:
Int Buffer = 3362 MB/s
Float = 3152 MB/s

[RyderOCZ]

2500+ Barton....Mobile chip? should be unlocked.

Do you have a good heatsink? I run 1.75-1.78V through my 2500+ Mobile and run it at 2.4GHz 24/7 as my workstation and it folds Proteins too.

So that has a FSB of 166 normally then? so at 200 you are already at 2.2....hmmmm

Running anything other than 100% for ram speed is highly NOT recommended with NF2....leads to all kinds of crazy abnormal behavior and errors :(

Have you tried changing the multi and it still boots at 2.2 when your at 200?

[GiANT]

it is not a mobile and wont boot at all if I try to change the multi. up or down. Wont boot at all.

[GiANT]

Ok comp just locked again. I had just posted the last message and comp was just sitting...I saw the screen saver come on for a couple seconds and it locked and all the colors smeared down he screen.

I backed the FSB back to 200 and the timings stay at 2.5-3-3-10. When it locks again Im pulling the memory.

I will run my old memory for at least a day to be sure Im not just talking out of my you know what. But I had my old memory back in twice and was not able to make the computer crash. And believe me I tried. I had 5 or 6 apps running, Sachs aquarium running as a wallpaper, ATI Tool, had the TV open, prime95, MBM5, Scanning for virus. The computer never locked like that with my old mem. and the old stuff was DDR333 running at DR400 2.5-3-3-6

[RyderOCZ]

Are your old sticks 256 or 512?

Set the FSB to 200....timings to 2-2-2-10 and run memtest from www.memtest.org (v1.65) Voltage at 2.7.

Let me know if there are any errors.

[GiANT]

Old sticks are 2 x 256. OK ryder will do on the memtest, will post again when finished. I appreciate your patients with this issue. This is the stuff that made me decide on OCZ int he first place. :coocoo:

[RyderOCZ]

The NF2 board has issues with 512 sticks compared to 256...You have Rev 2 of the board....google for a modded bios with CPC off....you should be able to clock those a lot better then.

[GiANT]

OK. First off, I ran Memtest86 1.65 for 3 passes without any errors. I will try to find some modded bios with CPC off..whatever that means. I have no clue what to put in google so give me a hint. In the mean time I'll see what I can come up with.

But my kneejerk reaction to "modded Bios w/ cpc off" is not comfortable at all with it. How do I know I am getting something legit and not some crap what will auto destruct my board. :hide:

[RyderOCZ]

Send me an email....I think I have 2-3 of them still on my server.

You do have the A7N8X Deluxe Ver 2.00 right? Not the -E version?

[GiANT]

I have the "E" version. A7N8X-E-Deluxe

[GiANT]

I found this thread/link but at the end he says he didnt touch any CPC??
http://www.nforcershq.com/forum/viewtopic.php?t=52662

Ifound this thread containing reddwarf modded bois but never any mention of working in the "E" deluxe boards.
http://www.nforcershq.com/forum/1-vt52555.html?postdays=0&postorder=asc&highlight=1008mod3sataboot4250&&start=0

And this trats modded bios thread but no "E" delux board
http://forums.pcper.com/showthread.php?p=3095488&postcount=62

do yo know of any that were made for my board?

[RyderOCZ]

Everything that I know of was made before the E board came out :(

Everything I have is for the Non E also because that is what I had.

Try raising the AGP voltage to 1.6 as that will put some more juice through the northbridge chip where the memory controller is. Up to 1.7V is ok...most Vid Cards are ok up to that.

[GiANT]

Will do it when I get off work. Since the ram runs 2.7V I assume it runs off the 3.3V rail. Is that correct?

Cause Im wondering if I shouldnt just go get a "good" PSU. Im back to wondering if Im possible taxing my PSU. Its supposed to be 400W but you know how that goes with low to mid range PSU's. Could really only be 320W effective?!?! After all I bought it several years at best buy...hehe

[GiANT]

Ok ryder any other suggestions. I upped the V. to 1.6 then to 1.7V, got lockups on both setting. Even set the timings to DDR400 2.5-3-3-10...locked up.

Im not sure what to think since memtest86 says no errors after 3 cycles. Is that enough to say its not the memory or do I need to let it run for 6-8 hours?

Im pretty much willing to try anything short of voiding my warranties but if I cant go more than a couple hours let alone a few minutes without this thing locking up Im gonna put my old memory back in for a couple days.:beat:

[RyderOCZ]

PSU can be a problem also yes.

Let memtest run overnight or while at work...something like that.....that would be a good idea....but that still may or may not prove bad ram...2 x 512 and these memory controllers can be fickle.

What voltage are you running on your CPU while you have it OC'd to 2.2?

[BGP Spook]

Sorry for butting in but....

What about S&M (http://testmem.nm.ru/snm.htm)?(click link scroll all the way down and look for the gray link "версия 1.7.6(160)," I think it is in Russian but not sure)

I use it as a proof for cooling/overclocking/general stability. It is excelent at making heat, and at testing memory/CPU stability.

Run it with all safe modes turned off and "average" CPU and "long" memory test(I also run variable pattern).
(takes mine just over 8 hours to complete with these settings)
I hope you have good cooling because the CPU test can kill your computer if you have sub-par cooling.

[GiANT]

PSU can be a problem also yes.

Let memtest run overnight or while at work...something like that.....that would be a good idea....but that still may or may not prove bad ram...2 x 512 and these memory controllers can be fickle.

What voltage are you running on your CPU while you have it OC'd to 2.2?

I put my DVM to the power connector and all power rails read good. The oscilloscope says my ripple is within spec also. So unless this PSU's wattage rating is grossly overstated, I dont think the PSU is the problem. I entered my hardware into one of those wattage calculators and at 100% power it said I only need like 300Watts. My PSU is an Antec400W model#SL400.

I will let memtest run tonight.

As for voltage I have tried every setting from 1.65V to 1.80V at least a couple times. You want me to go higher?:fire: CPU idles at about 46C and full load at about 53-55C:fire:

[GiANT]

I ran memtest all night at DDR400 2-2-2-5 and some 13 passes later I had zero errors. I also said earlier that all V rails were good. My DVM said the 5V rails on the atx power connector were all like 5.09V except one. The 3rd pin from the bottom outside edge of the board. The PWR_OK pin read 4.78V and when I looked in the BIOS it exactly matched the readout for the 5V line??? Which is strange cause before, as I stated earlier, it was telling me it was only 4.5V. So maybe it was??? Anyways I ran all night last night at DDR400 2.5-3-3-10 without any lockups. thats when I noticed the 5V rail was higher (within spec) than before. Coincidence?? Maybe. Anyways. Im not sure what Im going to do about this memory. Do you think that the memory controller is just being "fickel"?

It just seems to me that if it was a compatability issue that you would have come across it alot more. Since we had to "try" soo many things to see if it would work I am assuming that this isnt a "incompatability" problem.