NForce2 / KT400 questions

[Skulemate]

From what I've read over at AMDMB there's still the issue of not enough voltage to the chipset to handle those sorts of FSB speeds, and unlike some of the other nForce2 boards there's no volt mod to solve this at the current time.

[tim-may]

Originally posted by Skulemate
From what I've read over at AMDMB there's still the issue of not enough voltage to the chipset to handle those sorts of FSB speeds, and unlike some of the other nForce2 boards there's no volt mod to solve this at the current time.

http://www.xtremesystems.org/forums/...ght=a7n8x+vmod

it seems that its easy to vmod the asus board, but anything above 200mhz and it seems to be unstable/corrupts hard drives...

side note: with an xp2700, they connected one of the bridges (cant remember which exactly. if you need to know, ill try to find it again) they tricked it into thinking its a 133mhz cpu instead of 166mhz. by doing that they were hitting the 220+mhz FSBs. edit: forgot to mention they were using the abit nforce board

[Skulemate]

Yes, I'm aware of those VMods. I was referring to the voltage supplied to the chipset (i.e. northbridge) itself... the EPoX guys are VModing their chipset voltage in order to increase stability at the crazy fast FSBs, and I bet that'd help on the A7N8X as well.

[tim-may]

hmm....linkage please?
from what i found cruising on xtremesystem.org's forums some more, they dont seem to notice any difference what the epox Vio is set to...
maybe you can shed some light on this for me


edit: hot damn! nm, its a Vdd mod. seems raising it from 1.6volt (???) to 1.8-2.4 volts is getting people an extra 10 to 20mhz on their FSB and getting in the 3.2gb/s to 3.6gb/s memory bandwidth range beating out p4's with pc3500.

http://www.xtremesystems.org/forums/...5&pagenumber=1

[bigben2k]

Originally posted by tim-may
@bigben2k
i was looking around a bit to see what the max FSB or the a7n8x is....seems that the 1001.C bios do 211, but the beta bios found here for deluxe users and here for non-deluxe users boosts the bios to 250mhz. congrats, you now have your 233mhz that you wanted.

thanks should go to the guinea pigs at the nForceHQ forums in this thread for trying this out.

Ok, but I'll still thank you m8!

[tim-may]

Originally posted by bigben2k
Ok, but I'll still thank you m8!

sure no prob.

@anyone who has an nForce2 board: can you take a look at your boards and try to see if you can find an IR (or IRU) 3037 chip on it? According to the thread on xtreme, it should be near the dimm slots. If you find it, the epox Vdd mod might be possible for a lot more nForce2 boards than we think.

[Skulemate]

Sure, I'll take a look this afternoon.

[tim-may]

o hey, im gonna try to head to fry's tomorrow (birthday shopping for bro) and scope out all the boards they usually have on display. Im hoping there will be at least 3....

[tim-may]

ok so lets do this bad news first, good news second

bad news: the a7n8x doesnt have any of the chipset voltage chips that the epox board is using....(see note below though) and they didnt have the ABIT nForce2 board either, only the AT7 MAX2.

good news: the leadtek K7NCR 18-D Pro uses the exact same chips as the epox board. the difference between the 18-D Pro and the 18-G Pro is that the 18-G has the IGP intergrated graphics from nvidia. so, im assuming the vdd mod will work on this board too.


side note:
The IC i was looking for were the IR 3037. on the leadtek board it is the IR 3037A, so i assume this is the same chip, possibly a revision though.

The asus board has the following ICs near the dimm slots:

LM 358
28M

MV3581
24M

270EJ1K
7407

4425

and finally

685220
ACB
P227AAR

im pretty sure on these numbers, but they are really tiny and i might be off by one. so, if anybody is good at looking up chip model numbers/spec sheets and such, have fun. o, and btw, most of the chips are from TI (texas instruments)

[bigben2k]

FWIW...

The first two look like good candidates. The last two look like logic chips. I'll lookup tmo.

[tim-may]

Originally posted by bigben2k
FWIW...

The first two look like good candidates. The last two look like logic chips. I'll lookup tmo.

ok thx ben

[jtroutma]

This message is for airspirit specifically:

Have you been having any stability problems with your Epox board lately? I have had to reinstall 2k 5 times within the last 3 days and I finnally seem to have a stable copy on. I am also miffed because I cant run at 157FSB anymore without risking the OS.

Any suggestions?


BTW tried NUMEROUS BIOS tweaks, voltages, clock speeds, OS settings, etc...

[bigben2k]

Ok, here are the results:

The LM358 is a dual OP amp. It is part of a voltage regulating circuit.

The MV358I (and not MV3581) is a "Dual Low-Voltage Rail-to-Rail Output Operational Amplifier", also part of a voltage regulating circuit.

The "270xxxx" appears to be a memory related chip, probably a buffer.

The 4425: if it's a 4423, it's a clock chip (but that wouldn't make sense). TI doesn't have a 4425 otherwise.

The 685xxx could be a voltage supervisor/buffer, but that's p/n 886850x.

[airspirit]

After I installed the BIOS updates beyond 2C13, I have been rock solid ever since. I take that back: after that AND installing a better NB fan, I could. If you haven't done something to cool that NB better, you have to. I was experiencing similar issues before I did. For $20 on a Vantec Copper Iceberg, you can't go wrong, and you'll get a spare sink to use on your SB as well in the package.

If you have already taken care of your NB cooling, I would suspect the BIOS to be the issue. If you've already put the latest/greatest BIOS on there, then I would point to your RAM.

[tim-may]

Originally posted by bigben2k
Ok, here are the results:

The LM358 is a dual OP amp. It is part of a voltage regulating circuit.

The MV358I (and not MV3581) is a "Dual Low-Voltage Rail-to-Rail Output Operational Amplifier", also part of a voltage regulating circuit.

The "270xxxx" appears to be a memory related chip, probably a buffer.

The 4425: if it's a 4423, it's a clock chip (but that wouldn't make sense). TI doesn't have a 4425 otherwise.

The 685xxx could be a voltage supervisor/buffer, but that's p/n 886850x.

ok thx again ben. any clue from the spec sheets on how to go about a possible vmod on the first two buggers?

[bigben2k]

Let me compare the 3037 to what we've pointed out here.

Do you have(or can you get) pics of the chips in question, with the surrounding area?

[tim-may]

Originally posted by bigben2k
Let me compare the 3037 to what we've pointed out here.

Do you have(or can you get) pics of the chips in question, with the surrounding area?

umm sure. do you need a in-dept pict of the area? or just me circling the area from a pict ripped from asus's site? i can probly head to fry's again with the digital camera next....friday possibly, but i can do a ripped pict in 5 mins or so

[bigben2k]

I'd need to see the chips in question, and if there are any traces that lead to the NBridge.

[bigben2k]

Ok, I just went through all 7 pages of the Xtremesys forum thread... In short, they link pin 1 to pin 4 of the 3037, with a resistor, 680 to 720 Ohm, to increase Vdd from default 1.55, to as high as 2.0.

Some nut OC'd his pc to 236FSB, dual mem modules, synch, with a Vdd at 2.01 ... You should see the memory scores...(Sandra RAM int: 3655 MB/s, RAM Float: 3447 MB/s)

Still, as an Epox tech pointed out, it'll affect the voltage to the AGP slot, so it's really not a good idea. As stated multiple times: at your own risk.

Also, although everything looks fine at first, boot, get into windows, everything ok, then run a CPU loader program and boom, crash. Now everyone is scrambling to figure out the problem... [H] thinks it's the ram, XTS points at the CPU...

[tim-may]

Originally posted by bigben2k
I'd need to see the chips in question, and if there are any traces that lead to the NBridge.

dah....ok if we can get some who has an a7n8x already and take a picture of this area more in-dept:
(red are just mosfets, green are the ICs in question)



as to the that xtremesys thread, i havent looked at it since i posted it in this thread. gonna take a look right now.

[bigben2k]

Which chip is right next to that big coil, and power connector? (top left)

Which one is on the right, closest to the Dimm slots, gold cap, and fan header?

[tim-may]

dah...sorry cant help you there ben, i cnat remember. but if skulemate is lurking around somewhere, i bet he can help us...

[Skulemate]

Ben

The chip next to the coil and ATX header has MV358I on it, along with some other writing. The chip closest to the CPU fan header (along the edge of the motherboard) has 27AX5DK on it along with the Texas Instruments logo. The LM358 chip is the IC closest to the screw hole in the corner of the board. I'll try to post a picture of the area in question and post it here in a couple of minutes. I'll do a couple of varying detail, and will first post one showing the corner in as much detail as I can (easily) manage.

[Skulemate]

Wow... it's a lot harder to photograph the writing on those ICs than I had thought it'd be. All the photos I took tonight suck, because the flash kept washing them out the writing on the ICs. I'll try again tomorrow... I just don't have time tonight to take my board out of the case. Sorry guys.

[jtroutma]

Interesting discovery that I made. Not earth shattering, but interesting....

I finally found a stable point with my setup until I can successfuly unlock my CPU.

On my way to getting my system stable, I discovered that my video card was always "shimmering" while in some intesive battles in UT2K3. I figured at first it was due to the insability of the whole system but latter after the system was fine, it was still present. I then went into the BIOS and tried lowering my AGP bus from 75 to 70. After that, my graphics are nice and clean once again. I even bumped the card up to 8X with the 70Mhz setting and the card is even faster now. (75Mhzx4 than 70x4 or 70x8 that is)

I know that this is like "DUH...I KNEW THAT!" but I was even that mentality that 75Mhz AGP was fine but apparently it wasnt. Hopefully this helps someone out.