MSI K8N Neo Platinum VDIMM Mod?

[Jasonxxx]

Anyone know anything about a Vdimm mod for this new board?

I noticed a chip on the top of the board above the dimm slots, maybe this is the chip to mod?

I don't have a close up the chip to see what it is, anyone have a good pic of this chip on the board?

This board should be out soon. It is soild all the way around, except all the reviewer report that if you change the multi in the bios and then try to reboot it will fail. You have to reset the multi in the bios everytime you reboot. I don't know which A64's the reviewers were using, but is this becuase of a locked processor?

Here is a pic of what I'm talking about.

http://www.hardocp.com/images/articles/1082350875lk8SrXrQq3_1_7_l.jpg

[YoupY]

Probally a week from now I have this board too. I don't think Vdimm is located there. I'm guessing the chip that is located almost under the [H] :(

If you can get us some numbers of that chip I think we'll be able to figure out a Vdimm mod.

Is there a Vcore mod available yet for this board ??

[Jasonxxx]

Well, I got the board up and running fine..(MSI Neo Platium)

I maxed out at 240htt X 10 with a 3200+ clawhammer CG with the memory at 200mhz 2,2,5,2 and that is stable with aircooling with 1.6vcore

Now I didn't try any other multiplers yet, except 9 X 250htt (1.65 vcore)and it didn't boot, so I suspect that it was the processor that was at fault, but since there is no chipset voltage adjustment, that could be a limiting factor too.( I have the chipset cooled with the Swifty MCX 159 cooler, so I know it's not to hot)

Now @ (1) to (1) htt/memory, I maxed out stabe with Corsair 3502 512MB X 2 @ 230MHZ with 2.5, 3, 10, 3 timming and 2.85vdimm / 1.6vcore

If a person had 3.2 - 3.4 vdimm you could run the fsb up ? (at least 240fsb I can pretty much confirm.)

Any Vdimm mods? Any ideas on the vdimm mod or even a chipset mod?

[Richdog]

I hear that A64's are a bit flakey when a large amount of voltage is pushed through the Intergrated Memory Controller, and a I believe some users have had them die with running 3.3v on the Vdimm. I'm intending to do the mod on my Gigabyte K8N PRO and also the MSI Neo Platinum i'll be getting next week. Think i'll be shooting for 3.1-3.2v max on the Vdimm though, don't want any nasty accidents. :)

[YoupY]

I haven't got any luck finding the vdimm mod. there are 2 mosfets on the right side of the proc. I did some measuring and I expect the Vdimm is controlled with the chip that is located under the CoreCell heatspreader like piece.
Vdimm can be measured at some blank spots for caps near the dimm slots.

From the two chips located near the mem I couldn't find the datasheets, and without them is hard to figure something out. Codes on the chips are:

- ICS CN320761A 0343 93737BF (centered above the mem)
- Winbond W83 310DS 403LB

The ICS chip looks like some kind of memory buffer chip. Found some datasheets of chips which numbers were close to the number I was looking for, and they didn't look like voltage regulator chips.

Vcore was easy to find. The board uses a chip that's very often used on A64 boards. (ISL6569CB). Connecting a 50k VR to pin 10 should give the posibility to regulate the Vcore.

I haven't tried it out, and there's a good chance I won't try it, since the board was crapping out on me. When I hooked up my drives my system won't boot. Bios was getting corrupted very fast also en I had to clear the bios several times before something happend. So I smell another RMA :(

[Jasonxxx]

Is it possible to heat up the two pins on the corecell block that stick through the mb and remove the block?

I kind of suspect that is the chip, but there should be a trace or resister, cap etc.. that you could do the mod on?

I will look around the board and checkings volts...

[Jasonxxx]

bump:D

[Jasonxxx]

Well I'm about 99% sure that you need to remove the core cell heatsink to be able to perform a vdimm mod.. I also noticed that out of the four blank Cap spots around the dimm slots that two showed the vdimm voltage and the other two show exactly half the vdimm voltage, which I would assume that is the Vtt..

When you mod the vdimm, does the vtt scale up with it, or do you need to also mod the vtt? I would think that the vtt must always be about half of what the vdimm is to be able to remain stable and scale correctly...

Any inputs on this? :rolleyes:

[Jasonxxx]

Any information on the those chips that "YoupY" pointed out?

I still not sure what the " Winbond W83 310DS 403LB" chip does, I can' measure anything off the chip, cause I need small tips for the fluke.

[Jasonxxx]

Well I was able to measure the voltages on the "Winbond W83 310DS 403LB" chip in the uppper right corner of the board above the dimm slots.

3.3v on legs 5 - 8

2.86v (which is the vdimm, I also change the vdimm in the bios and it followed accodingly) on leg 1

No voltage reading on leg 2 ( if this is the chip that will allow us to mod the vdimm, this would the leg that we would mod with I believe)

1.43v on legs 3 and 4, which is the same as the vtt I measured


Any comments or anyone familar with this chip and what it does?

I would try a resister between leg (2) and ground, but my smb grabbers only have a 36k resister attached to them..

I will let someone else with more experience try it..

[YoupY]

Most cases the Vtt doesn't scale with the Vdimm. They are probally both controlled by the same chip.

My board went back for an RMA so I cannot test anything out anymore, but a possibility could be to attach the 3.3V direct to an elco where Vdimm is going through.

On one of the legs of the caps will probally read the Vdimm. Give that one the 3.3V and you can influence the Vdimm. It's a risky procedure however to try it out....

[Holst]

Hi all.
The 83310DS is for VBT only, not Vdimm.

This chip will track VBT with the input volage (normally Vdimm) so there should be no need for VBT mod.

I started a thread at XR on this,

If you find anything out can you post it over there as well please.
http://www.xtremeresources.com/forums/showthread.php?goto=newpost&t=32407

Those MSI corecell things might be made by winbond (ive seen MSI remark WB chips as there own before) any info you can give on the other chips on the board will help me out.

[Jasonxxx]

Well I directly connected 3.3v from psu to one of the + sides of a blank cap spot.

I noticed after measuring voltages that the blank cap spots that were 2.86v max before, were now were reading 3.3v, and the other two that looked to be exactly half the vdimm also tracked up to 1.67v from 1.43v (not sure if this is vbt or vtt)

Results, It didn't change a thing? I initially had the vdimm set to 2.5v in bios and slowly turned it up to 2.85v hoping that it would work, it seemed to be a little more stable. It would lock up windows just idling before at around 230mhz Mem w/ 2.5, 3, 3, 7 (1t) timmings, and after the mod it still would lock up, but seemed even more unstable. I changed from (1t) to (2t) and I was able to go on up to 240mhz, but when I ran sisoft mem bench it showed only 3300mb's bandwidth with 85% @ 240mhz??? @226mhz I'm getting around 3470mb's bandwidth with 95%

Maybe the vtt is not scaling with this? Any ideas?
I was able to remove the corecell heatsink and it shows MSI ICS, What should I look for?

[Jasonxxx]

By looking at the MSI Chip and comparing it to winbond ICS's I found this chip

W83304D, Advanced ACPI Controller for AMD K8, LQFP 48

Any ideas, Winbond didn't have any data sheets on them, I will do a search, maybe I can come up with one?


Anyone familiar with these chips?

[SlackeR]

measure vdimm on pin 7 in an empty Dimm-slot ;)

[Jasonxxx]

Well, I already removed the mod. I didn't want to corrupt the OS..
I haven't backed it up yet..

Pin 7 is the vdmim supply Voltage?

Since I have the Core Cell heatsink off, what can I do now?

[YoupY]

Does it show any numbers besides MSI ICS, maybe some kind of logo :confused:

[Jasonxxx]

It has an MSI Logo on it..

As far as a part number, written right below the logo the part number starts with: MS-6 3*****

I can't read it with my own eye sight, while inside the case. Until I get it back out I can't make out the entire number....

[Jasonxxx]

Here is a High res pic of the Core Cell chip: This is with the CoreCell heatsink removed...

Naked CoreCell Pic (http://www.ccountry.net/~jasonx521/IMG_0023.JPG)

[Sir Lancelot]

So, are there any modds for this motherboard or not??

[Jasonxxx]

We are working on it

[Hell-Fire]

If someone can confirm hands down that this is the W83303D IC, then I can help you work the mods out for it no problem.

[Jasonxxx]

What would you suggest in order to see if it is indeed the W83303D IC..

If you could give me a couple of tests that could send me in right direction I would willing to test it out and see if it matches up..

It is kind of odd that MSI would put there logo on it, but you never know.. Maybe they are capable of making there own IC's and just paid Winbond to use there blueprints with there name on it....

I know this is possible, but how to go about it I'm not sure.

:D :D

[Hell-Fire]

Prob same way companies rebadge Winbond ram ICs as well.

If you can manage a pic of the IC and the surrounding area...I can have you do some tests that will let us know whether its a Winbond IC or not.

In fact, that IC could be the W83303D or W83304D.

I will edit a pic I have here and ask for some readings.

[Hell-Fire]

Find out which resistors are connected to these 2 pins and then take voltage readings on each side of both resistors.

That will let me know if its a WinBond IC or not.

If you do not know how to do a connectivity test, let me know.

[Demogorg]

are there any other mods for this board? i suppose mine will arrive tomorrow

[Jasonxxx]

There is a Vcore Mod I believe, but I think they are still working on an OVP work around..
I'm hoping to figure out the vdimm mod asap..

[Cranox]

Vcore mod is same as the Epox 8KDA3+ i think ( same chip , so give it a try ;) )

Jasonxxx , Vdimm is maybe also the same ?

http://www.xtremesystems.org/forums/showthread.php?s=&threadid=36041&perpage=60&pagenumber=1

[BIG_DEADy]

Come on guys! Is there no VDimm mod for the K8N Neo? My board will arrive tomorrow and I need a high Vdimm because I'm using BH5.

So please work on a VDimm mod! :toast:

[jess1313]

well hellfire, my board is in transit to you now so we can get this thing cracked:banana:

it'll be there Tuesday:toast:

[Jasonxxx]

Sorry guys, I have been so busy lately... work and school and the 4th...

This week end I should be able to take some voltage, continuity, and ohm/resitance readings and get back to hell-fire.....

He seems to know this chip.....:D

[TARZAN]

with mine, when i set memory to 2.85v the msi utility always shows memory at 2.7v, i cant get above 210 @ 11,2,2,2 while i was 230 on nf7s @ 3.02v, v-mod is badly needed.

[Jasonxxx]

I noticed that if your not running your memory about 210mhz or above it lowers your vdimm to 2.7v, but if you run your memory at say 220mhz it will allow 2.85v vdimm..

TARZAN,

Here is couple of things that may be limiting you:

-Use 3 & 4 SATA ports (Next to AGP), not 1 & 2
-Set LDT Multi to 4 or 3
-turn off Cool N Quiet
-set mem to normal
-disable any overclock setting( i.e. private etc..)
-set cpu multi to highest available (default), and not on startup
-set APG to 67mhz
-Set memory to 200 instead of Auto if you are going to run 1:1

If you run the memory Async the MSI doesn't like boot with a different cpu multi other than the default....

[TARZAN]

hi, yep, im already doing those things, im running 240/200 same as yourself so ill try and change the timings and try running 1:1 to see if that works, its early days anyway, hopefully there will be a v-mod soon.

[Jasonxxx]

I got the OCZ 3700EB I ordered and it runs 250mhz 1:1 now, but I need more vdimm to tighten up the timmings...

I swear I can't get enough...:D

[jess1313]

Hell -Fire will have my Neo Plat by 3:00 tuesday evening. The mods are coming.

[Jasonxxx]

Outstanding:D My iron's already warming up....:D

[Hell-Fire]

Roger that Jess....look forward to working on it. I am 100% certain that this is a remarked IC as I cant imagine they would produce their own chip.

Jason, I hate you....but only because you have the 3700EB and I want some...lol. :hehe:

I will start on the board hopefully Wed if I get it tomorrow as I have several other things to finish up and ship out. I will take some preliminary readings to see just how hard the mod will be and if its indeed a Winbond IC.

I will keep everyone up to date on this as I move forward with the board. Also note that although I may be able to do a Vcore mod on it, I have no way of testing it as I dont own a A64 setup and therefore no CPU to test with.

Doing the Vdimm and Vtt is possible without the cpu obviously.

[Jasonxxx]

It would be nice to have more vdimm, maybe I would be able to drop the multi and run the memory@ 278mhz:D I can only hope:banana:

[TARZAN]

hi jason, yep booting at 217, 1:1, my voltate was 2.85, i was able to windows overclock up to 239 before it froze, i tried setting memory to 166 again and voltage stayed at 2.85, dont know why, i htought i d try running 1:1 at 230 see how that was but pc froze as soon as core center started, so went back to 166 again and volts have gone to 2.7, it seems a bit quirky, i dont know why, but im going to uninstall corecenter in case its something to do with that, max clock for hte board seems to be about 250, where are the frownies?

[TARZAN]

seems fine now at 230 1:1, havent had much time to test it, difference in sandrta between 200, 11,2,2,2 1t and 230 11,3,3,2 2t is about 150mb/s, which seems a bit off.

[zakelwe]

Good luck with the volt mod Hell-Fire and Jasonxxx for being persistent. I definitely need this mod as my memory at 2.85v is definitely crippled with both Corsair XL and old Corsair XMS 3500 being limited to less than 220 at tight timings.

It would be nice to see if VTT tracks VDIMM as

1) It is one less mod to do.
2) People buying the OCZ booster will not have to do any other mod :)

Regards

Andy

[Jasonxxx]

Before I got the EB3700, I was using Corsair PC3500 BH5 and my max stable was 228mhz w/ 2.5, 3,3,10 (1t)..... It gave me about 3450 or so in Sandra...

I would try not to use (2t) if possible it is a big performance hit, and it drops Sandra about 300 points....(That would explain you getting about the same Sandra score between the two settings..) Try 2.5,3,3,10 timmings and see how that works for you at around 230mhz...;) ;)

About the voltage thing, the voltage is capped at 2.7v below somewhere around 210mhz probably because they figured you would not need more than that at that frequency... That didn't bother me, as It didn't cause me any problems.. NOw if I was trying to run 220mhz + and it was still capped at 2.7v vdimm I would be pissed....:D

[TARZAN]

the corsair bh5 is what i have, (ver 1.1) on my nf7s when set to 2.9v whey would run at 3.02v, i was able to benchmark them (2x256mb duel) at 239mhz 11,2,2,2, never had a option to set 1t 2t command, i wasnt sure if they would run so well with a64, but is very nice couple of peices of memory, i cant go higher than 250fsb with my board, dont know if you have adifferent bios? i read it might help but i havent been able to change it, ive been runnning at 246 for the last 4 hours while ive been gaming and it seems happy, i though i might have problems becauseit always freezes when it hits 250, really wish i could go higher because with the divider memory is at 204mhz and id love to run them at 230 with tight timings once i we have the volt mod, theres no happy ground for me i either have to run memory or cpu below thier capablility, at the mo i have my a64 3000 at 2460mhz which is nice, the board has been faultless since installed, just a shame about the fsb, man id love to get 270-280, then i would be happy.

[zakelwe]

Originally posted by Jasonxxx
Before I got the EB3700, I was using Corsair PC3500 BH5 and my max stable was 228mhz w/ 2.5, 3,3,10 (1t)..... It gave me about 3450 or so in Sandra...

I would try not to use (2t) if possible it is a big performance hit, and it drops Sandra about 300 points....(That would explain you getting about the same Sandra score between the two settings..) Try 2.5,3,3,10 timmings and see how that works for you at around 230mhz...;) ;)

About the voltage thing, the voltage is capped at 2.7v below somewhere around 210mhz probably because they figured you would not need more than that at that frequency... That didn't bother me, as It didn't cause me any problems.. NOw if I was trying to run 220mhz + and it was still capped at 2.7v vdimm I would be pissed....:D

This EB 3700 I am using for the first time tonight is strange stuff, at default 2.5-3-2 it would do 238 maximum at 2.85v. However it does 255 at 3-3-3 and will do 233 at 2.5-2-2 !! So 2.5-2-2 it is at present.

I do not have the 1T / 2T option in the BIOS, so that is strange, it is in the manual ...

Regards

Andy

[jess1313]

Hell-fire

Did you recieve the mobo, it should of been there yesterday.

[Hell-Fire]

Yepperz...got it yesterday and havent looked her over as of yet tho bud.

I am gonna fire it up later tonight just to see if I can figure out what Vreg is controlling everything.

[jess1313]

:cool:

just glad 4 the moment that you got it.

Good luck:rolleyes:

[Hell-Fire]

Ok, identified the Vtt controller, working on Vdimm action now.

[Hell-Fire]

**UPDATE**

Ok, I am tired and will recheck this tomorrow, but I am pretty certain that MSI remarked the WinBond W83304D Vreg and stock that CoreCell sink on it.

I did some checks and am pretty certain of this. Connectivity between the FET reading Vdimm, and a missing (+) solder pad of a cap all point towards this conclusion.

Did I mention that I hate black pcb?

[jess1313]

makes it hard to trace anything.

[Demogorg]

how far are we with the mod ?? anything better ?

[Hell-Fire]

Its coming along, but I am working mainly on other components I have here at the moment so I can get them out by Monday.

I have identified the chip that is under the CoreCell sink as a rebadged WinBond IC, so mods arent far away.

[jess1313]

Rock on dude, I bet you have four soldering Irons going at once.

[Jasonxxx]

Steady.... Lmao...
I hope the best...
:D :toast:

[Jasonxxx]

Hell fire, Any luck?

[TARZAN]

Did someone say there is already a v-core mod? let me know someone,

thanks,

TARZAN

[Demogorg]

whats up now ?? i got my new adatas but there is still no oc 1:1 :( . If there are any mods please tell me maybe someone can help me with my probs http://forum.msi.com.tw/thread.php?threadid=55561&sid= i am so angry right now :(

[TARZAN]

i want to do the 3.3v mod for memory, does this board have ovp?

[Demogorg]

will the project have succes ?? i only want more vcore and vdimm :D (2 volt vcore and 3 volt vdimm would be enough ;))

[jess1313]

It is a strange board & will be hard to figure, without the chipset & LDT mods I dont think the VDIMM & Vcore will be worth having.

But thats me if it want do High HTT with 1:1 it is of no use to me & in IMO just plain junk, which is where this board stands right now.

But maybe with some off Hells touch he can turn it in to something.

[DuPu]

Hell-Fire, is that the same Winbond IC what was used on IC7 mobos? Maybe finding that 680ohm resistor and solder 10ohm vr between ground and other side of 680 resistor.

Or remove the resistor, and solder ~1kohm fixed in there...

Maybe, just my 20 cents ;)

[TARZAN]

I did the 3.3v mod, its good, ive got 3.36v on pin 7 on the dim slot, i was able to run memory up to 250mhz 11,2,2,2.5, my board always freezes as soon as it hits 251 so i dont know how well it could do, didnt seem to like cas 2.0 though for some reason, but there it is, a succesful mod, 3,900 odd points in sandra :moon:

[forumer]

Originally posted by TARZAN
I did the 3.3v mod, its good, ive got 3.36v on pin 7 on the dim slot, i was able to run memory up to 250mhz 11,2,2,2.5, my board always freezes as soon as it hits 251 so i dont know how well it could do, didnt seem to like cas 2.0 though for some reason, but there it is, a succesful mod, 3,900 odd points in sandra :moon:

pics ;) !

[TARZAN]

no can do, sold my camera to buy x800 xt (still waiting) i connected 3.3v to 2 of dimm slots, guess you only need one though, dont think itll make any difference, i connected to the pin shown on the very outside of the short bit shown in the thread here for the 3.3v memroy mod (on the next page i think)

[Demogorg]

and is it possible to get only 3 volt ? 3.3 is kinda much ;)

[TARZAN]

i dont know, i was just following what i read here http://www.xtremesystems.org/forums/showthread.php?s=&threadid=29178
im a monkey see monkey do kind of guy, maybe Rio71 could help you out there.

~thanks btw Rio71 :toast:

[CaTalyst.X]

I hope this helps speed things up a bit.

http://www.savepic.com/freepicturehosting/is.php?i=218160&img=neo1E.jpg

-CaT

[Hell-Fire]

Originally posted by CaTalyst.X
I hope this helps speed things up a bit.

http://www.savepic.com/freepicturehosting/is.php?i=218160&img=neo1E.jpg

-CaT


We are aware what IC this is, so no, this will not speed things up a bit for us. ;)

I would have long since been done with this mod had I not been working on other components.

I already found the trace on the back of the board where pin #19 bunches thru. I have located everything resistor, capacitor, and FET in the Vdimm circuit loop.

I will work on the board first thing in the morning and hopefully find a way to do this mod without having to cut the trace, but seeing as there is zero resistance in the loop, a trace cut is pretty much a given.

[Demogorg]

2bad but at least i suppose we will have no choice you are the professional ;)

[Hell-Fire]

Well, the main problem is this damn black pcb. I would love to find a top side trace connecting 2 caps or something, but no luck yet.

There is a single resistor that is in the loop, but to tired to mess with anything else tonight....when your eyes start bugging out from looking at tiny things for 2-3 hours...and 6 mods later on other components, its time to take a break before a mistake is made.

[HKPolice]

Hell-fire: does it work? My decision on buying a K8N Neo depends on the results of this volt mod :)

[jess1313]

HKPolice , you know one way or another its going to work. Even if he has got to let the smoke out a time or two.

By the way how does that AN50r run. Do you think it would do 275 bus if it had the mem.

[HKPolice]

Originally posted by jess1313
HKPolice , you know one way or another its going to work. Even if he has got to let the smoke out a time or two.

By the way how does that AN50r run. Do you think it would do 275 bus if it had the mem.

Ya, but I think filter tried the OCZ booster, and it worked, but he couldn't OC the ram a single Mhz higher even with 3.7v on BH5s.... That doesn't make sense.

I think my board is dead, it freezes after posting. It wouldn't boot at CL3, then i cleared the CMOS, then it started hanging.

I've seen ppl do 300+ FSB, but I couldn't get that high. Didn't try very hard tho, since I liked using 240FSB 1:1.

[jess1313]

I can run 275 1:1 with mine, I have run 300+ but with the clock down. The 275 is with 2450 - 2550mhz. I dont see how this is so when the PCI is not supost to lock. But it will do it & it loves it.

That was with my OCZ, now I am running 10x250 with Mushkin 222 maxed out because it is faster than the OCZ was at 275(OCZ is not EB). Thats why I ask.

Mine is not really a AN50r, it is a SFF SN85G4, the board is a minnie AN50r & when I check the PCI bus it is locked(not whith clockgem) PC Wizard 2004. I really dont understand this rig, it is faster & clocks better & is more stable than anything I have built a64 wise.

If the AN50r is this good I want to try one.

Oh back to the MSI, it makes sense to me I have messed with this board alot( it is my board Hell has). It does some stupid stuff.

One you dont have chipset voltage ajustment.
Two you cant ajust the LDT voltage.
Three there is some timings that must be crazy.

When you try & clock the board it acts like the voltage is AUTO AJUST but it dont give enough. This is from the fact that it will only work with the 3xLDT multiplier. 2x dont have enough voltage 3x seems to up it alittle & 4x is still trying to run on what 3x gets, so 3x is the only one that has enough volts to work.
(I may be wrong on this).

Then you have the VTT, I dont know how it works & if it want track the OCZ booster may not help.

I tried many sticks of ram in this board & I couldnt get it to do much.

I finaly got it to 260 FSB but not stable. 250 I could get stable but this was not with BH series ram so it was no good.

Give Hell time to work this thing, its not going to be cut & dry. The Vdimm mod itself is trying to be a pain & the rest may be harder. Without all the mods I dont think the board will be worth having. But we will see.

To voltmod something that is not even ajustable could be real tricky(chipset/LDT). These are the bigger hold back than the Vdimm.

Personally I would like to stick this board right up a MSI REPs you know what. Cost me alot for somethink this bad.

[Minnyboy]

Bump for this mod.

I'm about to get this board & wanna see if there's a Vdimm mod for the board or not.

Couldn't live with anything less than 3.3Vdimm as I'm poor & don't have high OCing memory other than my beloved BH5's...

Waiting anxiously for your results Hell-Fire...

Cheers,
Minnyboy

[Hell-Fire]

Ok...think I got her figured out.

Will tidy it up and post first thing tomorrow. ;)

[Hell-Fire]

Ok then.

I dont have a cpu to test with, so not sure if that is the cause of what I am experiencing, but I can now increase/decrease the Vdimm. The odd thing is that it will jump around quite a bit per adjustment. Still turning in the same direction the Vdimm can increase and decrease as well. It still steadily increases, but it may decrease by some millivolts while adjusting.

If anyone has any ideas about why that is, I am all ears.

Also, is 2.5v the default Vdimm for this board?


**EDIT**

So far the best way to do this mod, imho, is with the 3.3v = Vdimm method. The pots are getting pretty damn toasty during adjustment. I had the Vdimm up to 2.95v and the thing was almost to hot to hold.

On a better side of things, this board does not require a Vtt mod as it tracks Vdimm. I ran the Vdimm to 3.318v, and the Vtt tracked at 1.659v. ;)

[HKPolice]

Originally posted by Hell-Fire
Ok then.

I dont have a cpu to test with, so not sure if that is the cause of what I am experiencing, but I can now increase/decrease the Vdimm. The odd thing is that it will jump around quite a bit per adjustment. Still turning in the same direction the Vdimm can increase and decrease as well. It still steadily increases, but it may decrease by some millivolts while adjusting.

If anyone has any ideas about why that is, I am all ears.

Also, is 2.5v the default Vdimm for this board?


**EDIT**

So far the best way to do this mod, imho, is with the 3.3v = Vdimm method. The pots are getting pretty damn toasty during adjustment. I had the Vdimm up to 2.95v and the thing was almost to hot to hold.

On a better side of things, this board does not require a Vtt mod as it tracks Vdimm. I ran the Vdimm to 3.318v, and the Vtt tracked at 1.659v. ;)

Can you please post pics of the 3.3v=Vdimm mod?

[Hell-Fire]

Sure....I will post them in a bit.

Be aware that I dont have a cpu to test with, so I dont know if it will boot at that voltage. I hear some boards will read a high Vdimm, but OVP keeps it from posting until Vdimm falls to around 3.1v.

http://www4.ncsu.edu/~rwbradl2/Vmods/K8N_NEO_3.3=Vdimm.jpg

http://www4.ncsu.edu/~rwbradl2/Vmods/K8N_NEO_Vdimm_Trace.jpg

http://www4.ncsu.edu/~rwbradl2/Vmods/K8N_NEO_Vdimm_Circuit.jpg

[TARZAN]

Ive already done this mod, i connected 3.3v straight to the dimm slot, ive had no problem booting, vdimm is 3.35v. I cant test my ram because its such a crap overclocker, ive booted at 260mhz with 11,2,2,2 1t settings, i dont know what the aggresive timing is in bios, enabled i can only run cas2.5, but i get better results with it enabled and cas at 2.5 then disabled cas2.0, im going to sell it and get an abit kv8 pro, i really liked my last abit board and a lot of people seem to like kv8 pro.

[Hell-Fire]

Good to know that the board atleast posts with high Vdimm.

[Demogorg]

yeah but 3.3 is to much 3 volt would be fine

[DuPu]

Put a resistor between 3.3v rail and the wire what is connected the memory slot... Use a trimmer and You can change the voltage.. Am I right? I guess so..

[Demogorg]

ahh ok i think i will try that

[soul]

Excuse my vast levels of newbieness, but I have a few questions about this mod. First, is there any way to do this with it being a little less permanent than soldering the board? Also, does it feed 3.3v to the memory regardless of the BIOS setting?

[jess1313]

Well bud it must be giving you a fit. You ask if vdimm was 2.5 default I believe this is right if the cmos has been cleared. I am not 100% shure on that.

To tell the Truth I have been waiting for DFI for awail to get there boards out for the 64s. It looks like this may be soon.

After the the ABIT dyed on me for the second time I give up on the KV8. Two times in a row with no reason is enough for me. ABIT sayes they are investigating the problem right now but if they dont give me a board soon & it hold up I will build with the DFIs for shure. Although the board ran very good till it died & I would have built with it.

It dont look like the MSI is going to make the cut either. DFI has surved me best anyway, I stop building with the ABIT boards after I bought a Lanparty with the socket A boards before,maybe they will get there boards avalible soon and I want have to worry with these ABITS or MSIs.

Just do your best & take what time it need. If it dont want to play ball I will make a FOLDING machine out of it & leave it at 200 bus, so it will be able to be stable.

[Jasonxxx]

Okay, I have used the 1 wire from the 3.3v line to the (+) side of the cap shown and I have 3.31v on the vdimm, and exactly half on for vtt, but it is not very stable.... What would be causing the instability?

It 3.3v to much for EB3700? Is that causing the instability?

If I put a variable resistor in series in between the 3.3v line and where I connected to the cap leg, would this controll the amount of voltage that it would recieve?

Maybe 3.0v - 3.1v is the magic number for EB3700? or is there is something else causing the instability..

It is weird I can use 3,2,2,8 timmings at 245mhz where before I could not even post with those timmings at 240mhz +, but it still unstable.. It actually locked up in windows and when I reset the system, the bios was also reset.. (that is the first time I have ever seen the bios reset) Is this an OVP circiut causing this?

[HKPolice]

Uhh, are you sure you have proper cooling on the chips?? Take off the heatspreaders cuz they are insulators.

[Jasonxxx]

Well, I had some 1/2 watt resistors laying around and will try to drop the voltage from the 3.3v line to the (+) side of a vdimm circuit cap.. I first just used straight wire, and was getting exactly what my 3.3v line was reading.. I then put some resistors between the 3.3v line and vdimm circuit and it shows only what I set the bios?

Is the resistors not able to handle the wattage?

What size of resistor should I be using to drop the 3.3v to 3.0 - 3.1v? 200K? or 500K? or 50K? or 20K?

Is 1/2 watt resistor enough?

Anyone have some ideas here?:toast:

Update: I do need a much bigger wattage variable resistor. Like 25watts or so..

Anyone know where to get a 25watt or bigger variable resistor?:D

[Jasonxxx]

Hell-Fire,

So pin # 19 controlls the vdimm voltage? or is this just output line?

What do you suspect should be done with Pin #19? Are you looking for a resistor etc.. in the circuit?

After doing the direct 3.3v mod, my system is unstable, so I'm guessing that it's too much vdimm for my EB3700... I tried a fan blowing directly over them, but no change... I hear that the OCZ Vdimm Booster works, so the vdimm mod should work just fine.

The black pcb is just a pain uhh....:D

[jess1313]

I can see it now.

Pic of Intel MSI neo. ( no mods)

http://www.orifice.pwp.blueyonder.co.uk/burn6.JPG

[Jasonxxx]

Ouch..... I seen that earlier.. I hope that is not what happens....I will be sure not to leave my puter unattended..:D

[CaTalyst.X]

Hellfire, use the 3rd leg of the pot to ground it, that should help stabilize vdimm, the part about it getting toasty you will have to figure out, looks like mobo manufacturers are getting smarter, all the new boards either have a limited (sub 3.3V vdimm) after the mods, the mods dont exist, or the mods will work but only with a super heavy duty pot. Damn there ruining our fun! :'(

-CaT

[TARZAN]

maybe connecting straight to the dimm slot will work better? my voltages are stable, 3.36v, drops to 3.34v when running memtest, no fluctuations.

[Demogorg]

but how can u reduce it to 3.0 volt and will it destry my cpu?

[Jasonxxx]

Originally posted by Demogorg
but how can u reduce it to 3.0 volt and will it destry my cpu?

You can use the direct 3.3v line method and get 3.0v with the correct resistor... I don't know the amperage on the vdimm circuit?

I think it's about 7amps if I remember correctly... So, then if you want .3 or .4 voltage drop, you need a resistor that can handle that wattage

Can anyone test the amperage on the vdimm circuit to verify the this? at idle and full load?

Then we can get the correct resistor..:D (By my calculations you should be using at least a 16guage wire for the mod..

the resistor needs to be very small..

If my calculations are correct: Anyone remeber how to calculate this correclty..

I think it is: R = E / I , which would be .3v / 7a = about .042 ohms

or R * I = E, so .042ohms x 7amps = about .3volts

[CaTalyst.X]

lol a 1/3ohm 25W resistor should do then eh? ROFL.

-CaT

[tennvols_69]

the vdimm also goes to the cpu memory controller right? is thats whats causing some a64's to die? if so maybe we can find a work around for this by creating to circuits one using the 3.3 rail for just the memory and the other for the controller. as im sure just cutting the trace will not work cause im sure it has to be in a loop. whatcha think hell-fire? something similar has been mention but it looks like the idea quickly died.

[jess1313]

Hell-Fire been awail sense you last post. Any progress.

[Jasonxxx]

I used a very small wire(small fan wire about 24 - 26guage) from 3.3v line to vdimm circuit and came out with 3.16v for voltage..:D

I need to know the correct amperage of the vdimm circuit to be able to know the correct resistance... (

A 3 - 5 watt resistor is good enough if you plan on using the 3.3v line... If you wanted more vdimm than 3.3v you could use the 5v line, and a bigger wattage resistor... I will figure out the amperage and do the math for 5v line and 3.3v line..

Anyone know where to find variable resistors from 3 to 10watts very low ohms?

Anyone able to add to any of this?:D

[CaTalyst.X]

Is your board having stability problems at 3.16??

-CaT

[Jasonxxx]

CaTalyst.X,

NO, no stability problems... I can run 250mhz w/ 2.5, 2, 2, 5 (1t) timming stable...

The one problem with this is when the system is under load the current goes up and the voltage goes down because the resistance stays the same with just a resistor in series..

As of right now my vdimm voltage is 3.17v at idle and if I run prime it drops to 3.13v, but this is with a small wire.. I believe if I just used a small resistor it would drop even farther..

I need a clamp style amp meter.. to measure the amperage.. I tried to measure the amperage with my meter, but it poped a 10amp fuse?.. I then replaced it with a 30amp fuse and checked it with one lead grounded on a motherboard hold down screw and the other on the vdimm circuit and it was reading 24amps?

It must be reading more than the vdimm circuit?

I have extended the vdimm mod with 2 ft of 16guage wire and I believe if I had a clamp style amp meter I could measure the amperage with this wire...:D

Are you good with electrical / electronic theory? I.E. Trimpots and regulatory circuits?
If I understand it correctly, with trimpot it would be a series parallel circuit if one end is connected to the 3.3v supply and the middle is the output and the far end is grounded?
Wouldn't this created a potitental difference and divide voltage? But will this stop the voltage from dropping when the amperage goes up under load?

I really want to be able to use dimm slot 1 with the neo plat...
I would just use the OCZ booster..
"idea, maybe I could just get the the booster and solder a wire to the output to the vdimm? instead of actually placing it in the dimm slot?":D

[Jasonxxx]

What about chipset voltage? Since it is not adjustable... What is it set at? I noticed that the dfi has chipset voltage adjustment up to 1.9v... Can we get the neo plat chipset voltage up?:D

[TARZAN]

Id second that, but dont hold your breath. Im getting the epox version, epox mods anyone?

[Jasonxxx]

trader:D

Just kidding... I will continue to work on the Neo Plat, but when the DFI is available I plan on getting it..

The Neo plat has been good to me.. I have faith in it.. It just needs to be modded correctly...

I will start testing different bios soon.. I have used a couple of the bios, but I went back to 1.1

I have heard that either 1.31 or 1.33 bios is OCing better and that there will be a official 1.3 bios very soon.. The Neo still has hope...:D

[Jasonxxx]

Hell-Fire any word? You have any ideas on the vdimm or even chipset voltage?

[Hell-Fire]

I need to know the stock Vdd on this thing.

I think I have it narrowed down to what I need to mod for chipset, but I need a chipset voltage to reference as I dont have a cpu to get into to bios for comparisons.

Also, is VLdt something that we wont modded, and if so, what is the stock voltage on that as well.

Is this one of those boards where chipset can not be read without a cpu running?

[Jasonxxx]

Well in the bios, there is no chipset voltage reading or adjustability...

I was poking around the chipset, and the mosfet above and a little to the right of the chipset was showing:

Left leg = 3.75, Middle leg = 3.34(same as my 3.3v line), Right Leg = 1.62v..

I think that the right leg is the chipset voltage..

Above to the right of the mosfet is an IC chip that has 3.75v showing on it, so I would bet that it controls the chipset voltage..

Any ideas?:D

[Hell-Fire]

Yeh, on the FET that is just above and a little right of the actual NB I am reading 1.629v on the right FET leg.

If this is indeed the chipset voltage, then I think I have it figured out and will post pics once I do testing.

This is a large FET, so I would wager it is indeed for the chipset voltage. I cant think of anything else in that area that would require these types of voltages.

[Hell-Fire]

Okey dokey.

Got the Vdimm mod figured for sure. I am working on a way that may be easier for those who dont want to cut traces.

I will post pics tomorrow once I am done testing.

[CaTalyst.X]

Nice job HF way to go! Jason will be happy :D

-CaT

[Jasonxxx]

Outstanding......:D :banana: :banana: :banana:

I got my soldering iron warming up..;)

Thank you, Hell-Fire...

hmmm..."waiting patiently for pics;) "

[Hell-Fire]

They are a coming bud.

I am finishing up the Chipset mod, and adding voltage monitoring atm. ;)

Wondering if I should bother worrying about a VLdt mod or not. :confused:

[Jasonxxx]

I would wait and see if it is needed... I think the chipset and vdimm mod will most likely do the trick...:D

[Hell-Fire]

Ok. I made a new thread to cover the Vdd and Vdimm mods on this board just to make things easier for all.

MSI K8N Neo Plat Mods (http://www.xtremesystems.org/forums/showthread.php?s=&threadid=40136)

Someone please test the Vdimm mod for me before I have to ship this board back to the owner.

::cough:: Jason ::cough:: ;)

[jess1313]

Hell dont sweat it if it dont work will figure it from here, I think you have had enough fun with it & when I put a CPU in it if its not just right you can send me in the right direction to get it going from the reading I get. I dont have a camera worth a hoot so keep you some good picks of it. Your work looks very good there I posted some things in the other tread about the connectors you need.

[Hell-Fire]

I am pretty sure the Vdimm is working fine and gonna test it tomorrow at a local computer shop just so I know....may need to adjust the thread about it if not working.

[Jasonxxx]

I will get on that vdimm asap.. I will also try the chipset mod..:D

Thank youuuuuuuuuuuuuuuuuuuuu:D :D :D :D :D :D :D

[zakelwe]

Originally posted by Hell-Fire


On a better side of things, this board does not require a Vtt mod as it tracks Vdimm. I ran the Vdimm to 3.318v, and the Vtt tracked at 1.659v. ;)

Nice thread from jasonXXX and nice work Hell-fire in doing the mod. My hands a bit shakey for cuttin traces but knowing the VTT tracks the VDIMM mod means I can try one of the OCZ DIMMBlaster things, or whatever it is called, and hopefully get a good result. :)

Cheers to all involved with this thread.

Regards

Andy

[zakelwe]

Just as an aside, people have always mentioned IT and 2T settings for command rate for the memory and it even has it in my motherboard manual, but i do not get this option when I go to CELL or whatever it is called.

I downloaded the BIOS off the MSI web site but there is only one and that did not give it to me either. Very strange.

Any thoughts ?

Regards

Andy

[Demogorg]

when u have a C0 cpu u cant decide between 1t and 2t

[zakelwe]

Originally posted by Demogorg
when u have a C0 cpu u cant decide between 1t and 2t

thanks, I guess that is only CG then. I suppose C0 sticks to 1T so less flexibility / options when overclocking. Ah well ...

Regards

Andy

[Jasonxxx]

Yes, (1t) only for the CO versions. Not bad though as 2T is big performance hit..

[jess1313]

OK Jasonxxx, whats the story. Lets see some cool screen shoots 300 htt with mem 1:1. lol

[Jasonxxx]

LOL,

I won't be able to tweak it until next weekend with work and finals, so until then I will get things together...

I don't think I will reach 300htt w/ 2 X 512mb's, but you never know...:D I really think once I get everything in place a reasonable goal will be 280mhz range 1:1.. I plan on buying the DFI MB when it comes out.. I will see then if the motherboard is the limiting factor or not.... If the dfi can' t do anything significantly better, then I will just stay with the Neo Plat... I will selling either the Neo or the DFI depending on how things turn out... As soon as I can get my hands a 6800 Ultra (hope Evga since they seem the most reliable so far) I be able to see if any problems arise with either board also.... Right now I'm using a 9800pro..

When do you get you MB back jess1313 ?

[jess1313]

Ah not shure, it was going to be tested yesterday. I have heard no work from this so who noes. I am afraid I dont have much faith in this board but its the only one I have left untill I can get a rig sold.

I agree about the DFI, I have always run & built DFI even my old K6 rig is DFI. I hope they will get there boards out soon, they have took the time to try & get this chipset working. I really would like to see them build a board with the new K8T800 pro VIA chipset. It has been the best I have built with in the 64s.

But maybe this MSI board will run good, I was very disapointed in it when I bought it. I cant afford to make mistakes like this, it looked good on paper & in the reviews. But still iff the mods pan out it just might be a good one, we will have to wait and see.

[Barr3l Rid3r]

there still no findings about wich 5-25W resistor should I use for the straight 3.3V rail mod? I thinking about something that could give me around 3.1V

[WannaFly]

just a quick question

Should i set Vdimm to 2.85V or auto or whatever after i modded the Vdimm with 3.3V directly??

I mean i want to get a real 3.3V Vdimm and what should i set in Bios setup??

Thx

Edited:

Thanks, i know the answer now :toast:

[vh14a]

can someone please post the pictures of the vdimm volt mod for the K8N Neo Platnium agen? It seams all the picts. have been removed from this forum. I just want to make sure I do this right so as not to "fry" anything.