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

[Hoot]

Thanks hoot going to try and find one in the uk.

something like this ? Link

With all due respect, I would not get that one. It is scaled for measuring rather high voltage. Also, be leary of DPMs where they specify the resolution, but not the accuracy percentile. This one sports a resolution of 1V at 600V. That means it can tell the difference between 600 and 601 or 599, but no reference to how accurate that actula voltage reading is. I'm sure the UK has surplus houses like Alltronics, MPJA, Surplus Sales, etc. also. Just a matter of finding the right site. If you know of a neighbor who is an Amateur Radio enthusiast, ask them. They always have the scoop on stuff like that. I know, I've been one since 1971.

You can't hurt a motherboard by simply probing points to determine where to get voltage readings as long as you don't slip and short out two adjacent conductors. Missing electrolytic capacitors are alwyas a good starting point as they are usually situated after the PWM circuitry and often close enough to the destination so as to give an accurate reading.

A lot of people make the mistake of using any point on the chassis as their meters return path (minus lead). That is not a good practice and the accuracy of the reading can be compromised. The more current the particular voltage you are looking at is drawing, the worse the error will be the further you have the probes from the load when taking a reading. You can thank Mr. Kirchoff for quantifying that phenomenon (Kirchoff's Loop Law).

Hoot

[Lestat]

Hoot what do you recommend for a good common.... common ground...... ?
i have a spot sanded on the inside of my case that i clip the ground alligator clip to.

We all know or hope we know that grounding as close to the point of power or the device its self is best and that wiggling a ground can result in varied readings. and that a ground on the mobo itself is the bes tbut to say a nice little spot , one like mine that i sanded isnt good enough ? or MAY not be good enough ? i disagree.
sure if you just go slapping a ground any 'ol where yeah thats not smart. but a nice little sanded spot (to help with conductivity) on the side of the case, would suffice wouldnt you agree ? Especially since no motherboard out there can regulate the voltages with any sort of control or accuracy.
floating voltages is a hardware issue of motherboards that cant be fixed bu grounding properly or using a different positive lead on the mobo.
maybe someday when mobo makers start using GOOD voltage hardware including Caps and other regulatory devices then maybe we can get good voltages.
which by the way it still baffles me why mobo makers still use the same crappy hardware year after year when other boards have proven to have severely flawed voltage regulators.

anywho.. your electrical expertice is so much appreciated by all of us and your expertice in PC electronics is one thats hard to find. I say you start a website or some how to's with pictures for various common needs of the overclocker. voltages readings proper testing, proper equipment, pictures of where to test for those mobo's you have used.
proper setup of temp and voltage probes.

I for one would dearly love to know how to get good temp readings with this Venice. I got to used to my CG Clawhammer that of course was a mobile chip without the heatspread so i stuck a probe right smack dab on the side of the die. you cant do this with these chips... so im a little lost on what to do. your little mod with the onboard temp probe is a peachy keen idea to me and id love to see some pictures and a "how to" on how to do it.
My super flower temp monitor is simply going to waste right now... it has 4 temp probes and i am not using a single on. Although tommorrow im going to put on on the Northbridge chipset.
and before i forget,, did i hear correctly that MSI has these northbridge HSF epoxied on ? is this true or was that another mobo. ?
I imagine Radio Shack has Epoxy for the sinks i have for various spots on the mobo but last time i went in there asking for a Solder Pen they had no bloody clue what i was talking about.

edit
i wonder how well jbweld transfers heat and if it will have a chemical reaction and hurt anything i stick it to... lol just kidding i would never jbweld anything liek that lol... although it is a thought.....hhhmmmm O_o

[harpyboy]

incase no one noticed a brand new version of Core Center Digi Cell and PC Alert were released today
and something called V Center was released on the 11th.

http://www.msi.com.tw/program/suppor...ist.php?kind=1

v-center thinggy.... tells me my mobo isn't support..

yes hoot tell us where is a good grounding point on the neo2...

[Lestat]

yeah if you look at the supported mobo list i believe its nforce4 and 775 boards.

nothing new in corecenter for this board must have been stuff for nf4 boards.

but its still the only tool that will allow me to go up to 1.7v and higher. the bios no matter what i choose stops all voltages at 1.65v *yawn* oh well...


oh yeah has anyone had any issues booting with 2x HTT multi ? i cant i have no idea why but i cant... it just wont boot.

[Hoot]

Hoot what do you recommend for a good common.... common ground...... ?
i have a spot sanded on the inside of my case that i clip the ground alligator clip to.

We all know or hope we know that grounding as close to the point of power or the device its self is best and that wiggling a ground can result in varied readings. and that a ground on the mobo itself is the bes tbut to say a nice little spot , one like mine that i sanded isnt good enough ? or MAY not be good enough ? i disagree.
sure if you just go slapping a ground any 'ol where yeah thats not smart. but a nice little sanded spot (to help with conductivity) on the side of the case, would suffice wouldnt you agree ?...snip

Regarding the best place to get your return. As you said, close as possible to the load. Even though the chassis is grounded through 9 or 10 standoffs to the board ground plane, minute amounts of resistance (milliohms), in the presence of large current flow, can cause voltage drops. I'll give you an example. Running Prime95 Torture Large FFTs, I took Vcore reading from different ground points and this is what I got:

DMM (Fluke 189) plus lead attached right next to the CPU socket on the missing capacitor + solder pad. Minus lead placed at the following locations:

Bare spot on chassis floor = 1.582V
Black wire at the ATX connector = 1.575V
Black wire right where it first emerges from the PSU: 1.567V
Minus pad at the capacitor where plus lead is attached = 1.570
Minus pad on a missing capacitor at the other end of the motherboard = 1.564

As you can see, the subtle variances in ground current effects DMM readings. Admittedly, the difference in readings can be small as long as the point is somewhere on the motherboard. The closer to the load, the better.

When I used to glue strip thermistors onto my CPU cores, in the days before the thermal diode, I used Arctic Silver epoxy and as little between the thermistor and the core as possible. Naturally, with an IHS over the core, external probes are worse than the so-so accuracy that the Winbond 83627 in my honest opinion.

The MCP heatsink is not epoxied to the chip. The TIM between them is just ornery. Heat up the heatsink with a heatgun and workit slowly in a gentle twisting back-and-forth pattern until you feel it start to give.

Off to bed now. 5AM comes early.

Hoot

[Lestat]

thanks hoot.

as for your various voltages i think your aware that those readings are almost identicle to the the crapy floating voltages of this mobo. and other mobo's for that matter. I just want a product that matches its pricetag. one that actually regulates voltages. one that actually reads temps properly, which friends is a VERY VERY rudamentary electrical thing to do and yet we still pay up to 200 dollars for a mobo that can not do either very well.
anyways we all agree that grounding closest to the positive testing point is best.

and sorry for my lack of knowledge of terms but TIM by that do you mean the thermal tape they use on the back of northbridge and south bridge chips ?

as for epoxy on the cpu itself, no way i want to be able to RMA this chip if anything ever happens lol.
personally i have 2 types of probes. one is thing and has the gold colored tape surrounding the probe. the other has a bubble of ceramic or epoxy covering the tip of the probe. the 2nd type dont use. the first type is my probe of choice as its thin and more accurate in my eyes.

[Sleepy Samurai]

i do believe he means OCZ VX there is no GX and the VX stuff is all Winbond UTT

its gunna take about 3.6v to get that memory up to 260mhz and it will run 2-2-2-5 but even at 3.6v the VX stuff has typically shown that it doesnt go much higher. some have gone slightly higher but its UTT not BH-5

you wanna spend a few extra bucks for some seriously high performance high voltage low latency ram go get some of the Mushkin Redline PC4000 its 196 at newegg and that stuff will pump 270-275 2-2-2-5 at around 3.6v and 300+ on a 166mhz Divider

The ram is GX (gamer extreme) although it is almost the same as VX. It is garrented to 3.1V ± 5% instead of 3.5V ± 5%.

I could also get TWINX1024-3200XL / DDR-400 (XMS-3200XL) / 2-2-2-5 but the LCD display maybe in the way of my HSF, an XP-90.

I have not tested the OCZ GX as I am testing some Hynix BT-D43 2x 512mb (selling them if any one is interested.)

My computer topped out at 285 fsb.

With the hynix at 240fsb 3-4-4-8 2T@2.75v. My CPU can do 2.25GHZ no problem and my ram can do 250fsb but put them together and it crashes.

[Lestat]

yeah i looked up the GX stuff some get 250mhz 2-2-2-5 with 3.1v but not much more than that. its shown as a gamers ram since it has low timings which does indeed help but some reviews said its not the overclockers ram.. although 250mhz 2-2-2-5 at 3.1v seems pretty darn good to me. not sure how much higher it goes or what ram chips it has but with those volts its either UTT or the new BH-5 (which is not the UTT chips btw)

[Hoot]

thanks hoot.

as for your various voltages i think your aware that those readings are almost identicle to the the crapy floating voltages of this mobo. and other mobo's for that matter. I just want a product that matches its pricetag. one that actually regulates voltages. one that actually reads temps properly, which friends is a VERY VERY rudamentary electrical thing to do and yet we still pay up to 200 dollars for a mobo that can not do either very well.
anyways we all agree that grounding closest to the positive testing point is best.

and sorry for my lack of knowledge of terms but TIM by that do you mean the thermal tape they use on the back of northbridge and south bridge chips ?

as for epoxy on the cpu itself, no way i want to be able to RMA this chip if anything ever happens lol.
personally i have 2 types of probes. one is thing and has the gold colored tape surrounding the probe. the other has a bubble of ceramic or epoxy covering the tip of the probe. the 2nd type dont use. the first type is my probe of choice as its thin and more accurate in my eyes.

1) Yeah, the voltages read all fall within the likelihood of what you might see using the onboard chip. The difference is you can trust what you read with the DMM, though often they only miss one another by a few hundredths of a volt.

2) Tthermal Interface Material. Which can mean Tape, Arctic Silver, Zinc Oxide, whatever. In the case of most motherboards, they use an easy to apply tape type material, but some use zinc oxide. Whatever they use, some are more prone to becoming adhesive in nature even when they are not old. Tape types fall into that category since they rely upon the device heating enough to melt them into place. Kind of like casein based "hot glue" but more thermally conductive.

3) I did some extensive study of the strip thermistors some time ago and published a good article at Overclockers.com. It may still be in their archives of older articles. The most important part of implementing them is removing the capton (yellow plastic) at the tip so that the actual thermistor is exposed. Otherwise the tape acts as a thermal insulator to a certain extent. I found three pictures in at my ftp site form that article, detailing the preparation and mounting of the strip thermistor. Here they are:




Hoot

[sideeffect]

Hoot what was that new bios like?

And to do that thermistors mod you would have to remove the heatspreader which im not so keen on doing. But i do have spare inputs for temperature reading for a thermistor on my vapochill so maybe in the future ill consider it. Thanks for the pics.

[Lestat]

hoot -

dear god dont tell me that that thermistor is soldered to that chip cuz it looks exactly like a drop of solder was dropped on that and then its resting against the die. i mean how utterly foolish can one be ? hot solder sat down on the tip of a thermistor resting right against the silicon ? so much for the resin that covers the cpu die. i hope thats just silver epoxy.

and my super flower will not function with a thermistor that is exposed there is such a delicate amount of stuff that joins the tip of that thing. i already have one that i had to throw away as it got caught on something and tore just a tiny piece of the head of the thermistor tape open and exposed the metal tip and now the alarm goes off everytime i plug it in..
same thing for thermistors that im not using they will randomly throw up an alarm. and especially if they touch the side of the case or something another alarm..
there are times when i touch the tip and the alarm goes off and its not a temp alarm its something like a short alarm or sensor alarm thats not heat related.

again i paid for a warrenty that i dont intend to void by epoxying things on the chip no way no how...
temps on the venice are the least of my concerns they run nice and cool on water even at 1.7v its still 30-33 idle and 44-47 load.

[toledo]

Hi Guys . I new here , first post

Does anyone know what has been changed in those last modded bios (v.1.a3mod and v.1.a) ?

The link :http://www.lejabeach.com/MSIK8N/k8nneo2.html

Thanks

[Lestat]

im not even gunna go into that bios,,, its junk just like 1.9

USB issues,,, Temp issues,, RAID issues.... need i go on. also the voltages are limited to 1.6v as was stated here earlier unless they were raised in this release....
They need to go back to 1.8 and rework 1.8 instead of taking 1.9 and messing things up even worse. 1.8 kicks butt... and could do more if they really wanted it to.... but MSI wont.. they will just keep making things worse.

and its all in german lol... go use Babelfish and translate things and see if you can understand whats going on.

[toledo]

im not even gunna go into that bios,,, its junk just like 1.9

USB issues,,, Temp issues,, RAID issues.... need i go on. also the voltages are limited to 1.6v as was stated here earlier unless they were raised in this release....
They need to go back to 1.8 and rework 1.8 instead of taking 1.9 and messing things up even worse. 1.8 kicks butt... and could do more if they really wanted it to.... but MSI wont.. they will just keep making things worse.

and its all in german lol... go use Babelfish and translate things and see if you can understand whats going on.

Ok . Seems to me that 1.8 1T moded is the best option so far . Witch one are you using ? I didn`t know that 1.9x is crap ....

[Hoot]

hoot -

dear god dont tell me that that thermistor is soldered to that chip cuz it looks exactly like a drop of solder was dropped on that and then its resting against the die. i mean how utterly foolish can one be ? hot solder sat down on the tip of a thermistor resting right against the silicon ? so much for the resin that covers the cpu die. i hope thats just silver epoxy.

and my super flower will not function with a thermistor that is exposed there is such a delicate amount of stuff that joins the tip of that thing. i already have one that i had to throw away as it got caught on something and tore just a tiny piece of the head of the thermistor tape open and exposed the metal tip and now the alarm goes off everytime i plug it in..
same thing for thermistors that im not using they will randomly throw up an alarm. and especially if they touch the side of the case or something another alarm..
there are times when i touch the tip and the alarm goes off and its not a temp alarm its something like a short alarm or sensor alarm thats not heat related.

again i paid for a warrenty that i dont intend to void by epoxying things on the chip no way no how...
temps on the venice are the least of my concerns they run nice and cool on water even at 1.7v its still 30-33 idle and 44-47 load.

Naw, that's not solder, it Arctic Silver epoxy. The thermistor is mounted so that the active thermistor component on the end barely touches the edge of the core, but flatness is maintained such that it is not higher than the core. That way, it does not touch the face of the heatsink. If you look closely, you'll see the CPU in the picture is a 1.0 Mhz Tbird. That should give you an idea of the age of the pictures. If you look at one of those thermistors under a stereo microscope, you'll see it consists of two plates with a resistive material between them, just like an SMD resistor or capacitor, except it has leads attached to each plate. As long as you don't short the two plates together by having them touch a conductive material, it'll work fine. That's the beauty of AS epoxy. It's non-conductive. An issue that arises from exposing the active component inside the capton tape is the fact that is is a fairly fragile component. Without the tape to protect it, you have to be very mindful not to crush it between two things. It crumbles almost as easy as a piece of Grape Nuts cereal. Having the device actually touch the part you are interested in seeing the temperature of insures as efficient a transfer of heat to it as possible. As for IHS' (Integral Heat Spreaders) are concerned. They exist for one reason. To keep inept owners from damaging the core. No other reason. They add a layer of thermal resistance between the core and the heatsink that only serves to decrease the thermal transfer efficiency and as such, I remove them from any CPU I buy that has one. Admittedly, not until I'm satisfied that the CPU is not defective and meets the manufacturers specified standards. I do not RMA parts just because they don't overclock well. That is dishonest. I do RMA parts that fail under warranty or do not meet the manufacturers specified standards. Bogus RMA costs are passed along to us the consumers one way or another. Just like insurance fraud. I had my SD for about a week before I removed the IHS. There is a thread at ocforums about the endeavor. My temps dropped substantially after I did so because AMD pooched the IHS mounting job. Normally, it only buys you about 1-2C at idle and ~5C under load, but that 5C could be what's holding you from the next step up in an overclock.

Hoot

Edit: I almost forgot. The 1A3 bios still has flaws. One of them is the fact that as you work up an overclock, only restarting between steps, you'll find a stable overclock, only to discover that if you turn the PC totally off for a while, it won't pass POST. This does not happen if you don't overclock too far apparently.

[sideeffect]

Yes i would have considered removing the heatspreader if they had messed up the contact but mine has pretty good contact as the temperatures are consistent with what the vapo shows.

Also mounting a vapochill on a chip with no heatspreader is an accident waiting to happen

Worst thing i ever had to do was mount a vapo on a socket A cpu and was so pleased that these new chips have some kind of protection.

Yes I agree they should have continued with the 1.8 bios for the venice and san diego cpus. Because they work best with it. They only needed the new bios for the dual core cpus. They made us poor san diego uses wait for ages for a bios just to fix the 1T issue that could surely have been fixed quicker.

I havnt tried the v.1.a3. But like Lestat i found the 1.9 bios versions to be not as good as the 1.8. The best version of the 1.8 for a san diego is the NFM rev 3 1t mod version.

[Lestat]

i cant recall if im using the 1T mod or the regular Rev3 ,,, hhmm i dont run 2T anyways no matter what,, its way way too large of a loss in memory performance. you ahve to overclock too high to make up for the loss.

and yes my CG Clawhammer had the thermistor smack up against the side of the die so i loved the accuracy but ugh dang heat spreaders id remove mine but i cant afford to replace this chip if it goes bad down the road, nor replace it if the heatspreader removal goes bad.

your mod with hooking a thermistor to the onboard temp gague is something id like to know more about is there any reference for this mod ?
although id have to start using the no temp modded bios's ...

[h4nc0]

Edit: I almost forgot. The 1A3 bios still has flaws. One of them is the fact that as you work up an overclock, only restarting between steps, you'll find a stable overclock, only to discover that if you turn the PC totally off for a while, it won't pass POST. This does not happen if you don't overclock too far apparently.

Yeah, this BIOS also seems to be a bit wacky when overclocking. Clockgen set 2.6 and 2.7ghz runs prime-stable, but once BIOS set, it won't boot Windows.

I've been following the German MSI forum, but I feel like I am completely ignored.

[sideeffect]

If you post in english they will ignore you. Even though they understand what your saying. If you post in bad german they'll tell you to post in English

[Lestat]

well murdok will reply in english but not very often. only if enough people talk in english hoot and i have asked in english and they have replied in english just not very much.


and yes if you set the bios to lets say 250mhz it wont boot into windows but you can set it to like 260 270 with clockgen and its fine.
the bios really has some issues...

[SNIPoR]

yo guys . from where i can measure the 3.3v in my mobo??

[Sleepy Samurai]

and yes if you set the bios to lets say 250mhz it wont boot into windows but you can set it to like 260 270 with clockgen and its fine.
the bios really has some issues...

Don't know if it makes a difference but I have the same problem with BIOS 1.8 250 is unstable.

Would 255 work?

[Hoot]

yo guys . from where i can measure the 3.3v in my mobo??

There's two convenient places. I'm at work, so I can't give you a picture, but look right next to the ATX connector. There is a missing electrolytic capacitor, though its location is silk-screened on the board as a circle that is half-shaded in white. I believe you can read it there. Also, right next to the SATA3/4 connectors is another missing electrolytic capacitor. I also believe you can measure it there.

Hoot

[Hoot]

Don't know if it makes a difference but I have the same problem with BIOS 1.8 250 is unstable.

Would 255 work?

that's interesting. I only see that issue using 1.9x and 1A3 bioses. Either 9,10,or 11X250 1:1 runs stable for me using 1.80 bios. These NF3 Ultra boards are a walking contradiction.

Hoot

[Lestat]

just look in the bios under pc health thats a pretty accurate reading from what i can tell.

and thats hynix ram so 250mhz might not work for that ram are you also upping the voltage and cas timings on the ram ?

what brand and speed of ram is it ?