distressing experience with Iwill KK266 + Tbird 1GHz

[compay]

Hello,

I read all the great comments on the Iwill KK266 mother board here and decided to try one. Unfortunately I have had some distressing problems. I put the following system together:

board: Iwll KK266-R
cpu: Thunderbird 1 GHz
power supply: Fortron 300W
ram: PC133
os: Debian Linux
hsf: Alpha PAL6035, Sunon Fan

I ran the board at 133MHz FSB and lowerd the CPU's multiplier in BIOS so that it ran at 1 GHz. It ran quite hot; up to about 58 degrees C was the highest I saw by looking in BIOS. By lowering the core voltage I was able to get it to run stably at about 53 C.

Like another user mentioned in reference to this mboard, the core voltage settings I selected are not what I would get back: to get it to run at 1.52 I had to select 1.45. Hmmm... flaky power supply? I thought Fortrons were supposed to be good.

Right now the system will not boot at all; the motherboard will not even post. This happened after I installed and ran lm_sensors, a program that reads CPU and other motherboard sensors in Linux, and the whole system locked up. (you can read about that here http://www2.lm-sensors.nu/~lm78/readticket.cgi?ticket=510 if you want).

I guess I don't really know what I am asking here, just wondering what the heck I am doing wrong since all other reports seem to indicate that this board, once you have it booting up consistently, is stable. It seems strange that it would crap out after running OK for few days rather than crapping out right away.

 

[RagingGuardian]

Sorry to hear about that. My board works great with a Duron @150fsb. tTe boards probably dead or maybe you forgot to reset the board to 100MHZ operation before rebooting.

 

[compay]

Yeah, I've tried resetting the FSB to 100, clearing the CMOS, changing the power supply, changing the RAM, removing ALL the cards (even the video card). The only thing left to try would be swapping out the CPU or running the CPU in another board. Unfortunately I don't have another board or CPU to do this with. Is it possible I had some sort of thermal disaster with the CPU? It was running pretty hot, but well within its operating range. How good is the quality of the Thunderbird? I'm pretty sure the HSF is good, and I used the correct amount if (Arctic Silver) thermal paste on it. Geez, I wasn't even overclocking the thing that much, 1000 MHz at 133 FSB - way less than a lot of people here.

Man, the last system I put together was a pii 350 on a 440bx chipset. It seems the game has changed a bit since then!

 

[sandorski]

Does it post? I'm thinking that maybe you lowered the voltage too far or, if it posts but doesn't load the OS, perhaps that monitoring software doesn't work with the mobo.

BTW, 58c won't harm the Tbird, my Duron650@750 has been running @ 55-60c fine since last Aug.

 

[compay]

Thanks for the replies. No, unfortunately it does not post. Here's what I heard back from the support bbs for lm_sensors:


<< The only way I can think of that lm_sensors could hose the system permanently is by inadvertently writing to the EEPROM on the memory DIMM, causing the BIOS to get very confused when querying the memory. >>


I'll try using some of my old PC100 DIMMs and see if that fixes it; I had just swapped in and out the same PC133 DIMMs I had in there at the time of the &quot;incident.&quot;

At least I have the joy of discovering something for the Anandtech community: don't use lm_sensors under linux on the Iwill KK266-R!

 

[compay]

OK, after 3 days of trying I finally contacted someone at Iwill and got an RMA number for the board. Hopefully they won't take forever to send a replacement. At least I still have my trusty old PII350/440BX to use in the mean time.

 

[SerraYX]

Good luck getting it RMA'd, I hear Iwill is pretty quick with that stuff, so you should be fine.

 

[Remnant2]

I've never used that particular program, but back in the past I've had the same kind of problems.. Once, back in the DOS days, a video game hosed my system's BIOS, and a couple years ago in Win95, a misbehaving app rewrote the eeprom on my softmodem (!), causing me to have to re-flash up to the newest firmware for it to work at all after that.

 

[compay]

Hmmm... after reading this page I'm starting to think I broke either the board or the CPU myself. Maybe lm_sensors just pushed it over the edge, or the fact that it shorted at the same time was coincidence. I'm going to take my CPU to the place where I bought it and have it tested. If it's fried it's probably because I f***d up. Oh, well, CPU's are cheap, right?

 

[JohnnyTT]

Right!