MSI K9N problems? Post here

**ozzimark** · 07-17-2006, 12:55 PM

All right guys. I know a lot of you have been having strange issues with this board. Mine is quite finicky, but there seems to be a trick to getting it to work properly. I'm currently testing out a few things, but for me, the largest problem by far involves something called "DQS Receiver Enable Delay" (DQSRED for short) in systool. sadly, this option cannot be changed in the BIOS, but I emailed MSI asking for a few things just a few minutes ago, so we'll see how that goes.

So, for those of you who can boot into windows, I ask you to tell the following:
"DQS Receiver Enable Delay" as reported by systool in the "AMD AM2 Memory Fine Tuning" page
HTT speed
CPU Multi
Memory Ratio
Memory timings & voltage
Which memory you are using

What I have found so far is that as the HTT is raised about 10mhz, the DQSRED also rises about 0.10 to 0.15ns. However, the MSI seems to be setting the DQSRED too high in some cases.. most notably when the HTT is raised. Sometimes, the BIOS will raise the DQSRED a full ns for a 10mhz bump... when that happens, even though I can still boot into windows if the memory ratio is low enough, SuperPI will get an error for anything over 64k

i'm currently working to see if there is a relatively simple formula that you can plug in the mhz and memory divider into to get the ideal DQSRED.. so stay tuned i guess.

i also welcome anyone with questions or tips that they have found from their own board to post here, hopefully we can all figure out this board and get it working the way it should

**Eldonko** · 07-17-2006, 01:46 PM

Disabling all spread spectrum options helps w/ HTT issues and remember to drop your LDT to keep HT bus below 1k (see pic below).

Also keep an eye on MAL, it gets auto set as low as 5. Check w/ systool, not avail in BIOS. =/

**ozzimark** · 07-17-2006, 02:02 PM

Originally Posted by Eldonko

Disabling all spread spectrum options helps w/ HTT issues and remember to drop your LDT to keep HT bus below 1k (see pic below).

i haven't tested it thoroughly, but i'm not even sure if my sempron can get the ht link over ~900mhz

**faruquehabib** · 07-17-2006, 02:03 PM

Originally Posted by ozzimark

i haven't tested it thoroughly, but i'm not even sure if my sempron can get the ht link over ~900mhz

that setting SHOULD technically match what kind of ram you have. example: i have pc6400 ram, so i run that setting at 800mhz. works quite well this way.

**ozzimark** · 07-17-2006, 02:28 PM

Originally Posted by faruquehabib

that setting SHOULD technically match what kind of ram you have. example: i have pc6400 ram, so i run that setting at 800mhz. works quite well this way.

are you talking about the option that eldonko posted a picture of? it doesn't have anything to do with ram though.. just keep htt * ht link multi under 1000mhz

edit: hrm. cpu temps can't be right. cpu is cooled by a waterblock, with a radiator semi-far away from the cpu. moving from ~40cfm to ~100cfm over the ram/cpu area drops cpu temps in the bios ~6-7c.

**Zebbo** · 07-17-2006, 02:54 PM

Already posted this into another thread but here it is again:

VDIMM seems to overvolt about ~0.13V, Vcore is steady and very stable but somehow more volts than 1.5V seems to active sort of an "overvoltage protection". Board just cuts the power off when rebooting, sometimes when saving CMOS & Exiting BIOS gives same symptons too.

**ozzimark** · 07-17-2006, 02:57 PM

just measured vdimm on my board. closer to ~.03v high at 2.2v

where are you measuring vcore?

edit: here we go.

first column is the actual ram speed. 2nd is the DQSRED that it boots with, 3rd has what it takes to be 1M stable. if it works with what the bios set, then it's blank

now to just reinterpret into something useful

**Zebbo** · 07-17-2006, 03:38 PM

You can measure Vcore from the coil legs that are located left side of the CPU socket.

Edited both posts to clarify this a little, only time board sets vcore down to 1.1V at POST happens if it has cut the power off for no reason but it looks very familiar to what OVP is.

Ozzi, do you have problems of using higher vcore than 1.5V ?

**ozzimark** · 07-17-2006, 03:56 PM

i sometimes have issues with voltages, but not always. i just booted at 2.8ghz and 1.55v before without any problem, and rebooted a few times in the bios, so.. i'll look into it closely to see if i can find the exact issue.

edit: 1.4 + .2v = ~1.53v for me

**Zebbo** · 07-17-2006, 04:24 PM

Do you have single core CPU or X2?

**ozzimark** · 07-17-2006, 04:26 PM

Originally Posted by Zebbo

Do you have single core CPU or X2?

3400+ sempron

edit: can't reboot at or over 1.6v. 1.55v is fine. rebooting from 1.6 down to 1.55v is a no go, but 1.55 up to 1.6 is fine. very wierd

**Zebbo** · 07-17-2006, 04:40 PM

Ok, you are first person I've heard is able to power up system and boot with more than 1.5Vcore. Though you make exception for that also since all the problems I've heard about has been reported with system using X2 CPUs, and you are first guy who's experience I heard with single core CPU.

**nn_step** · 07-17-2006, 04:44 PM

This could be helpful

**ozzimark** · 07-17-2006, 04:53 PM

Originally Posted by Zebbo

Ok, you are first person I've heard is able to power up system and boot with than 1.5Vcore. Though you make exception for that also since all the problems I've heard about has been reported with system using X2 CPUs, and you are first guy who's experience I heard with single core CPU.

hm, well, try this:

boot at like.. 1.4v, and in the bios, change it to 1.6v. i'll be surprised if it doesn't boot. it just doesn't let you reboot while 1.6v is applied, even if it's to lower it back down to 1.4v

why i say this:
with 1.55v set, i was able to reboot after applying 1.65v
then priming for a while, i reboot. shuts down, as is expected. then when i boot up again and lower the voltage back to 1.55v, which definitly boots, it won't reboot. but after i get it to boot again, no problems

kinda awkward wording, my apologies

**Zebbo** · 07-17-2006, 05:15 PM

For what I've tried do this far there doesnt seem to be matter which voltage + extra voltage combination I use if the Vcore goes above 1.5V, after I'm done in BIOS and save changes to CMOS the board just simply turns power off. If I hit power button after this it starts and goes through POST without warnings and makes it way to Windows. However the case gets harder if I go with more Vcore than 1.5V, lets say 1.52-1.6V and when I'm done in BIOS and quitting it will shut power off like with 1.5V as well but it goes more complex for this time since nothing happens when I'm trying to power it back on. I have to turn PSU off and let it be there for 10-30secs (more the core voltage, longer the wait). After the while of pending I can power system back up and again I see no errors on POST screen and it will boot without problems to Windows.

**Jaco** · 07-18-2006, 02:55 AM

1,45vcore + 0,1 is the maximum my board allows.
Chipset gets freaking hot (55°C sitting idle) , although I don't think this is causing instability. It dumps a lot of heat in the case

**speed bump** · 07-18-2006, 07:37 AM

My first board was okay to 1.525v as far as not having the reboot bug. However it didn't like vdimm beyond 2.20v without having the reboot bug either. It did seem to work pretty well though other than the stupid voltage problems.

My second would usually lock up at post if I tried to boot dual channel and couldn't install windows at anything other than the 400mhz FSB rating so it went byebye.

My third one should be here today and I hope it comes out alright.

the first board I pulled this off on
3000+ orleans that I have taken up to 1.65v and 3ghz
2x512mb of crucial PC-5300 ballistix with BQM chips.
533mhz memory ratio
HTT of usually 330mhz
memory timings of
4,5,4,4,2t,auto,2t. if I remember correctly.

**drkCrix** · 07-18-2006, 02:57 PM

I am also having issues, ~1.4v I start having reboot and shutdown issues, happens at stock speeds and overclocked, this is with a X2 3800+

**ozzimark** · 07-18-2006, 03:05 PM

guys, i just got a beta bios from MSI, labeled as version 3.12.. quite a jump from the 1.21 that i was previously using. while it has no more options than the older bios does, it seems to have raised my max stable ram speed about 10mhz, and max bootable htt from 360mhz to 375mhz. i'll see if there is any effect on the voltage protection issues.

**Zebbo** · 07-18-2006, 03:14 PM

Feel free to share it

**Jaco** · 07-19-2006, 01:25 AM

Anyone tried this bios already? version 1.2 (dated 4th of july )
http://www.msi.com.tw/program/suppor...UID=730&kind=1

Ozzi,
I'm also interested to test your new bios.

**Jaco** · 07-20-2006, 09:05 AM

Originally Posted by ozzimark

guys, i just got a beta bios from MSI, labeled as version 3.12.. quite a jump from the 1.21 that i was previously using. while it has no more options than the older bios does, it seems to have raised my max stable ram speed about 10mhz, and max bootable htt from 360mhz to 375mhz. i'll see if there is any effect on the voltage protection issues.

hi Mark,
any news so far?

**Zebbo** · 07-20-2006, 12:23 PM

I'm interested to know as well. The latest official version (1.20) didnt do any better than 1.31 beta already did.

**ozzimark** · 07-20-2006, 06:38 PM

so far, i'm loving this bios. working on 1T stable speeds.. i'm up above 400mhz with 2x1gb in, while before i could barely get 300mhz

also seems to have *sorta* fixed the DQSRED issues. while the values are the same as before, it doesn't seem to be effecting stability quite as much, must likely due to other improvements.

obvious disclaimer: 100% beta bios, use at your own risk, for k9n plat/sli only
http://www.eclipseoc.com/image/A7250NMS.rar

btw, it's version number is "3.12"

edit: oh yeah, sorry about the long delay. thunderstorm the other day took out my cable

edit 2 for 8m picture. it would appear that the new bios breaks clockgen.. moving even just 1mhz locks it up

i know the timings aren't all that great, but it's at 1T. the question is which is better... 3-3-3, 2T or 4-4-3, 1T at the same mhz. this might require more testing.