The OFFICIAL MSI K8N Neo2-939 tips and tricks guide thread

[Lestat]

this is a board issue MSIMAX

this happens to alot of people when you go THAT high with the HTT.

Frankly and ive seen this MANY MANY times and even on my own board. friggin slot one is a piece of crap and has TONS of issues.

try moving your ram back to slots 3 and 4 and see if it does it.
Also try swapping the sticks around see if one of them is bad.

there is no need to sit there and yank the cpu in and out.

and frankly i dont know why you would even think about doing that. Just reset the jumper on the motherboard to reset the cmos...

fall back a little on your HTT and it wont happen. I really only see it happen when 1) your using a DDR Boostger and your putting over 3.4v into the ram
or 2) you have your HTT at or above 300 HTT.

what is the CPU Multiplier on that chip ? 10x ? 11x ?
just leave it at default and see how far you can go with your ram and cpu and then see if you start having the beeping error.
if it keeps happening then RMA the board and never get another one.
Just go get the DFI board..

[msimax]

it happens at stock settings

[Lestat]

ouch thats a bad thing
switch the sticks around see if it still happens.
then try 1 stick at a time.
then take the ram and try it in another system if you have it.
i doubt its the cpu infact i dont think the cpu can do that.
It's either the ram or the motherboard.

the fact that it works when you swap cpu's i THINK is becuase you have removed power from the motherboard totally.

try this.
next time it happens. power down and unplug the wall cord from the psu then hit the pwr button so the rest of the power is disipated from the board.
walk away for 10-15 min ok now come back plug in the psu and see what happens.
you may also need to reflash the bios.. i DOUBT that is whats causing it but a corrupted bios can do all kinds of wierd stuff

[msimax]

well u wont belive what it was. some of the dielectric grease wasnt in the middle of the cpu when i installed it. so when i examined it i noticed corrosion on the center pins just a little bit but i guess enuff so over time they didnt have good contact. so once i found that out i gave it a bath and then put a good amount of grease on it put it back in samething happened. took it out clear the grease and went over every inch, right in middle was a bent pin. guess all the grease made it do that so i did some surgery with tweezers and i got it running . going to test it

[Lestat]

hmm i forgot you were using subzero chilling

bent pin huh? sounds like you been swapping cpu's too much and got careless.
out of all my hardware days thats one of the things ive never done. and hope i never do.
good luck getting her fixed. im not sure how it locked down with a bent pin. but there's a big big chance the cpu is now toast.... its possible it fried the part of the cpu with the memory controller or it did something and damaged the electrical components of the cpu socket on the mobo.

good luck guy i hope you didnt toast that cpu.

[msimax]

all is well for now im back up

[Raillex]

Since this is the Neo2 tips and tricks thread, I wondered if anyone might be willing to share a tip on how to get my Neo2 Plat to run any faster than 285 MHz 1:1 dual channel using G.Skill PC-4400 LE. The board will do 9X285 9-3-3-2.5 3X HTT @2.65v vdimm without breaking a sweat but I start getting errors in Memtest at 286 MHz no matter what the voltage and no matter what trcd or trp values I use. Raising CL from 2.5 to 3 nets me a few more error-free Memtest MHZ to 289 but prevents me from even booting into Windows.

My 3700+ tops out at 2.69 GHz@1.6v and I was kind of hoping to make up for its shortcomings by running it at 9X299 instead of my usual 10X269. Am I smoking something here or what?

[Lestat]

Since this is the Neo2 tips and tricks thread, I wondered if anyone might be willing to share a tip on how to get my Neo2 Plat to run any faster than 285 MHz 1:1 dual channel using G.Skill PC-4400 LE. The board will do 9X285 9-3-3-2.5 3X HTT @2.65v vdimm without breaking a sweat but I start getting errors in Memtest at 286 MHz no matter what the voltage and no matter what trcd or trp values I use. Raising CL from 2.5 to 3 nets me a few more error-free Memtest MHZ to 289 but prevents me from even booting into Windows.

My 3700+ tops out at 2.69 GHz@1.6v and I was kind of hoping to make up for its shortcomings by running it at 9X299 instead of my usual 10X269. Am I smoking something here or what?

ok well what happens when you lower the cpu multi to 7x ?
I would consider myself VERY lucky to have tccd ram that runs that high on such low volts, its not very often you see ram like that.

overall it's a mix of the cpu and the ram and the mobo.

you just gotta mess around a bit.
1)
raise your vdimm
2)
lower your cpu multiplier.
3)
lower the HTT Multi to 2.5x
4)
raise your vcore slightly to help stabalize the cpu/memcontroller
5)
raise your AGP voltages to 1.6 or 1.65
6)
lower the ram to 183 or 166 depending on whaqt bios your using, then lower the cpu multi far enough that you can raise the HTT high enough to go higher then where you are right now.

over all if none of that works then

WELCOME TO THE MSI NEO2 PLATINUM OWNERS CLUB!!! and get used to issues like this. lol. your not alone my friend. mine dies at 270-ish actual memory speed.

most of the neo2's that i have seen or used have this issue and there is no fix. you can TRY RMA'ing it but i doubt MSI will RMA it. and you probably would get a motherboard that's worse.

1:1 NEO2's start getting realllllyyyyy flakey at around 275-ish. not all of them but a good many of them.

Lastly... and probably even just as important and forgotten about my AMD64 owners... alot of AMD64's have pretty flakey Memory controllers and there is no fix.. no matter what you do, it wont get higher but running the memory on a divider can sometimes help but usually not.

Sideeffect also has a TCCD bios that has memory settings made up for TCCD ram. look for one of his posts on this page or the last page and he has a link in his sig for his bios's. Grab the 1.B3 TCCD version and give it a try. it might help.

[msimax]

[IMG][/IMG]

[IMG][/IMG]

[IMG][/IMG]

[IMG][/IMG]

[IMG][/IMG]

[IMG][/IMG]

i was just bored
i think its time for dfi but i want to wait and see what ATI does going to start tweaking timmings

[Lestat]

holy MSIMAX

i mean seriously holy mother of god

im gunna email these pictures of Sean at OCZ and tell him to stick this OCZ PC4800 Platinum up his arse.
i cant even get past 270 without issues.


what voltages you running ?
and does that boot into windows ? if so get some cpu-z verified screens you may just have some record breaking ram there dude. serious that ram is unheard of.

[msimax]

i just modded my booster with a 15 turn pot so thats with 3.0 volts even

[Lestat]

why would you mod it ?

[msimax]

because the stock trimmer sucks and if u keep adjusting it wears out quick. with the mod
the voltage is more precise and i check with a dmm ex.. if i want to go from 2.9v to 2.93or to2.95 u will NEVER get that stock also i put mosfets on the back of the board to keep all that heat away.and it also helps air flow it works well,and i will say a must for runnig bh5 ch5 or any utt and trying to find the upper limits

[msimax]

[IMG][/IMG]

[IMG][/IMG]

[msimax]

hey sideeffect can make a tccd ver of 1.b3

[Lestat]

yeah the one thing that pisses me off the most about my booster is that if i power off or reboot the voltages change. mostly when i power off then power back on the voltages are never the same.

thats one brutal looking mod dude lol... but if it works...
i sent you a pm msimax read it...

[Raillex]

ok well what happens when you lower the cpu multi to 7x ?

I didn't try 7X mainly because I'd never be able to run anywhere near the 330 MHz I'd need to reach my 2.69 GHz goal on the CPU.

I would consider myself VERY lucky to have tccd ram that runs that high on such low volts, its not very often you see ram like that.

Yeah, I was pleasantly surprised to see it run so well with so little juice, which is exactly the opposite of my SD. I have a feeling this RAM would absolutely fly if the mobo would only let it.

1)
raise your vdimm.

I did, all the way to 2.85v. The more volts I pump into it, the more errors I get in Memtest. Like I say, the RAM works just fine with 2.65v @ 9X285, so I don't think it's a voltage issue.

2)
lower your cpu multiplier.

I'll check it out and see what happens, but I think it will defeat the purpose since I just can't see this RAM running the 330 MHz it will take to get my SD back to my 2.69 GHz goal.

3)
lower the HTT Multi to 2.5x.

Tried that, no help.

4)
raise your vcore slightly to help stabalize the cpu/memcontroller.

Tried that too, no help.

5)
raise your AGP voltages to 1.6 or 1.65.

Yep, that too.

6)
lower the ram to 183 or 166 depending on whaqt bios your using, then lower the cpu multi far enough that you can raise the HTT high enough to go higher then where you are right now.

I thought about that, but decided against it since I would never be able to run the RAM anywhere near its potential due to my low CPU ceiling.

... alot of AMD64's have pretty flakey Memory controllers and there is no fix.. no matter what you do, it wont get higher but running the memory on a divider can sometimes help but usually not.

It could be the memory controller, but I have a sneaking suspicion it's the Neo2 itself. When I first put this rig together, I had an awful time cold booting at anything over 250 MHz as described in this thread. NFM's 1.36 and 1.8 SD bioses fixed that problem and I haven't had an issue since. My point is that this problem appeared after reaching a nice, round 250 MHz figure, not 249, not 247, not 253, but 250 right on the nose. This, plus the fact that MSI's core center overclocking software also just happened to max out at 250 MHz, made me suspect that a this was an artificial ceiling set by MSI's bios writers when they wrote the 1.9 series bios. I'm seeing the same pattern here with this nice, round 285 MHz RAM limit and since other NF3 boards don't seem to have this "feature" at 285 MHz, something tells me I'm dealing with yet another arbitrary MSI-imposed ceiling.

Sideeffect also has a TCCD bios that has memory settings made up for TCCD ram. look for one of his posts on this page or the last page and he has a link in his sig for his bios's. Grab the 1.B3 TCCD version and give it a try. it might help.

Tried that one too, no help.

[sideeffect]

Msi has released another beta of the 1.b bios.

Only change they listed was USB 2.0 works.

Well weve been enabling the usb 2.0 since the 1.9 bios and its always worked

Well maybe MSI did some other things too. I made a mod bios of the 1.B4 with the extra dividers and some more relaxed memory timings.

you can get that from my site.

Ill make a tccd version of the 1.B final when that comes out. It will be soon im sure.

[Raillex]

Well maybe MSI did some other things too. I made a mod bios of the 1.B4 with the extra dividers and some more relaxed memory timings.

you can get that from my site.

Hmmm... I looked there but can't seem to find a 1.B4 version. Which one is it?

I was thinking that one thing that could help me if I can't bust out of this 285 MHz RAM ceiling is a working 216 MHz divider. Any chance of this happening?

[sideeffect]

I changed my link so it goes to the homepage and not the 1.A official bios now.

[Lestat]

Raillex -

you have one of the infamous NEO2 boards with the mhz bug. My board tops out around 270mhz no matter if i run a divder or not.
My HTT tops out at 370-ish

im not sure what you mean by "wanting to run the ram at 330mhz"

MSIMAX is the first guy i have ever seen who has run a NEO2 board that high. and he was running 2T i believe.

Over all in general you picked the wrong board to go crazy with high memory settings at a 1:1 ratio.
If your running it at 269 right now then you may have just reached the max of the board.
and the max of the cpu/memcontroller.

honestly run the cpu at 7x and then raise the ram some more see what happens.
with TCCD based ram a Nforce 4 DFI board or ASUS board is the only thing thats gunna get you close to 300mhz 1:1 1T its just not gunna do it on a NOE2. and like i say if it does then you have one of the rare neo2's that does.

You have pretty much done everything already except lower the cpu to 7x and/or run the ram on a divider.

try those and if it dont work. , once again i say "Welcome to the Neo2 and Hating It" owners club. This board has been nothing but issue after issue since its release. but more so than any other has been its horrible FSB stability. its always been called the "250mhz bug"

[Raillex]

What I said was I just can't see this RAM making it to 330, which is what I would I would have to do in order to get the CPU up to respectable levels with an 8X or lower multi. My goal is 9X300 1:1 1T, which is, as you point out, doable with a few of the other NF3 boards.

I made a conscious decision not to buy another DFI board because my last one, an NFII Ultra Infinity, fried my Geforce ti4600 out of the blue one day after only a year's worth of use as well as the brand new Radeon 9800 Pro I bought to replace it. I also got sick and tired of that board corrupting my bios with every failed overclock and I'm not convinced from reading people's experiences at various DFI forums that the newer DFI boards are any more reliable.

At least it's nice to know that I've pretty much maxxed this thing out so I can stop banging my head against the wall.

[msimax]

quick ? iwas running at ddr620 and at 3.0 - 3.1 volt the errors are starting to go down is this burning in. i remember reading that some tccd likes volt if so what should be my max


bbl

[msimax]

update just now i got no errors in test 5 for the past 3 test i have to leave so bbl :-)

[uniacid]

I wonder if I had a ocz booster if my problems would be solved, but its odd how I can't use settings I've used before now, and raillex I understand how you'd feel against DFI but they are a lot better now I must say but you can always for for the MSI NF4 board