D900K shutdowns

Its a bit long i know but it should help people answer my question i hope. if you want to skip to the question just find the ******.


I have what seems to be a fairly popular laptop on this forum, the Clevo Co. / sager D900K laptop monster. I love this laptop dearly and have used it with complete success through one tour in iraq and had NO problems but after i got home i gave it to my wife to use and she loved it too, then the shutdowns came. At first it would only shutdown at what ever time if felt like after giving what seemed to be an error code of, the scroll lock, caps lock, and Num lock lights would flash together about 7 times and then OFF,. everything. But it started to happen more and more often so after much research it was decided that it was either a power problem or a thermal problem but seeing as how i cleaned the laptop all too thoroughly and very well i knew that thermal was close to impossible so i began the quest to find a computer repair place to get this fixed and it became the general consensus that i would need a new Mobo and since they were hard to find and when i found them they ran about $500 at least with no return policy we just went ahead and built a new desktop instead. I have since ruled out the mobo problem and decided to replace all thermal "grease" on the mobo and TADA . it came on like nothing ever happened. But to my dismay it imediately started to give the same error code, lights, and do the same thing, I have yet to use any outside programs to determine the actual temp of the components but***** i was wondering if any of you who have this laptop have had the same issue and if so how you fixed it if at all??**********
I wish very much to save my 28lbs laptop dominator.

And im waiting for Gophn to answer cause i know he will as it seems he is all over this forum.

 

welcome to notebookreview,

The D900K is one problematic machine. I own one so I will speak from my experience.

When you get the machine on, you should install software like coretemp to read the cpu temperature.
You can install rivatuner in order to read the graphics temperatures.

You should run your fans at full speed all of the time by pressing FN+F2

As far as the problem goes, it sounds like a hardware failure.
The problem sounds like it could be the BIOS, memory, connection issue or a motherboard failure.

You should check the memory first as it is the easiest to do. The memory is located in the center bay of the laptop. It takes 4 screws to remove the actively cooled plastic cover.
Try powering the system up with one card at a time. Try each memory card in each slot. The D900K is very picky on memory. If the memory is not matched all of the way down to the spec code on the memory chips, problems will occur. This is because the D900K will always force dual channel memory, so if their isnt 100% compatibility the system will crash and be a real piece of work

If the problems still exist, the next step is to reflash the BIOS. You should flash to BIOS version 1.00.09. Gophn has a copy of them, just do not install version 1.00.10b, it will kill your system.
You should flash the BIOS via a USB memory card, CD, or floppy disc.

If that doesnt help, you should check for connectivity issues, check that all of the antennas are connected, if there is a loose one tape the end, so it doesnt short the motherboard. Check that the keyboard, touchpad, and indicator ribbon cables are connected.


Please give the specs of your system
Are you running a AMD Athlon, or AMD Opteron series processor?


You can get a replacement motherboard for less than $500. You can get one from Z-TRONICS for $170.
http://www.ztronics.com/en/laptop-parts-by-model-number/alienware-area-51-models/alienware-aurora-m7700-amd-motherboard-d900k/flypage.tpl.html
Order over the phone, their website isnt the best.
Talk to Jerry, in my opinion he is the most knowledgeable of the people there.

Please note everything they sell is repaired/used hence the lower price. You can buy a new board from RJTECH, but it will cost about $480

Where are you located?

Lets see what Gophn has to say

K-TRON

 

wow, talk about a great walk-through first off
and as far as specs:
VGA: Nvidia 7900 GTX (512mb)
Mem: sts mem 624 kb / extended mem 1023mb *!! but i have 2x 1Gb sticks in there!!*
CPU: AMD Athalon 64 X2 4600+
OS: XP pro (not fully updated)
BIOS: v1.0.08 (i think) i know it ###.08 though, scrolls fast

I currently trying to install those programs, but im not very successful considering it keeps shuting it self off faster and faster and it is getting pretty hot, too hot, so i have elevated it and placed 8 40cm fans under to try to cool it a little better but they are only run off of 8 AA's so im not how long i can keep them on before i find myself looking for more AA's.
the only things i have changed to it is it has new CD (optical) drives but i only have one in now for better airflow. And i may have replaced one or both of the hard drives but im not sure at this point. The mobile desktop was made in 2006 after all and a lot has happened since then.

I will go through your steps but they will take time due to the "frequent" shutdowns and just for the record it will shutdown for the first time after about 5-15 mins depending on whatever it feels like apparently. And after that its lucky if it can get into windows.

 

If none of this works do you know of anything i can do with what i have left over like make a new laptop, possibly with new better hardware? i just thought of it during my wonderful cooling break and thought it was pretty interesting.

 

And another update, i think i might have found something. I took the ram out and put each stick through each slot and with the bottom Ram slot the comp would beep for POST but nothing would come on the screen. If i used the top slot then it would POST and the screen would come on and both ways it would shutdown before OS could load. Same error code lights as described before.

What ever could it mean?

 

Well the system components are specific to the D900K, you can use the processor and harddrives/optical drives in a homemade desktop, but thats all.

Back to the problem.
Test the memory again, try only one card at a time.
I had a similar problem with my first replacement motherboard, where it would only see 1Gb of memory regardless of what slot it was in. Both memory cards worked, but upon inspection the memory card slot was indeed defective, as in a pin had snapped off inside the socket.

Take a flashlight and look at your slots and see that everything looks normal.
Than test your memory again.

I have a doubt in my mind that you didnt clean your system well.
You need to physically clean out the heatsink, and the fans with q-tips or so.
You can remove the heatsink from the underside of the system. You need to do this to clean out the pad of dust which will generate inside.
Honestly with the fans going at full, cpu load peaks at 136F, and my CPU gets alot hotter than yours. 125Watts, versus 89Watts.
You should also check the graphics card. Their are 4 screws or so which hold that black metal plate to the top of the graphics card. Clean that out.
Use Q-tips and rubbing alcohol.
When you turn the system on, as soon as the BIOS loads, press FN+F2. It will force all fans to run at full speed. Your system should not get hot when the fans are on full.

About your fan cooler, why use batteries?
You can buy a plug power supply at any voltage you want for like $5 at MPJA, MOUSER, DIGIKEY, NORTHERN, or any electronics place.

I have a feeling your problem is completely memory related, since I had similar annoying activity when my memory slot became defective.

You can also try upgrading your BIOS. Version 1.00.08 is what you have, you should upgrade to 1.00.09

try powering your system on without the wireless cards, tv tuners and optical drives, it makes life easier when finding the problem


K-TRON

 

ok ill try removing all unnecessary hardware and the heat sinks for the graphics card and CPU, but both have been cleaned quite thoroughly and had new Zalman super thermal grease applied to the components. But a memory issue would make more sense than thermal at this point. And i used batteries cause i made it very late at night and nothing was open, and im a electronics engineering student so i like making weird stuff. The fans are by no means a permanent thing, yet.

 

Ok i checked out the RAM to see if there any missing/messed up pins or connections but they were all fine. So i decided to turn it on just to see how long it would take to shutoff again and to my surprise it hasnt turned off yet. and its been almost an hour. Since it has been on this long i took the chance to install Rive tuner and Coretemp. and i have a few screen shots of the info if you can figure something out from them. But i have noticed that the laptop is a good bit cooler when it isnt plugged in, which makes sense cause it doesnt have as much voltage available to get hot. But when it is plugged in the CPU temp is coming dangerously close to the TCase max which is 54*C ive seen it as high as 52*C , this is while it is plugged in. But while it was just on battery power i saw the CPU temp for Core#0 spike for just a split second to 260*C . which kinda worries me. Unfortunately this laptop has almost NO programs on it and i still havent found a driver for the Wireless card yet for some reason. But the worst of it all is that all these temps and screenshots are while the laptop is running at NO load, CPU load is 1% at most. >.< so what happens when i try to actually do something.?

P.S. It wont let me post those screen shots so ill put them on photobucket al post a link when i can.

 

Ok so it took me all of 10 mins after i wrote my last post for my hopes and dreams to be once again crushed under the colossal weight of the Clevo Co. D900K. I took it downstairs to hook it to the router so i could do some updates and hopefully let windows find a driver for the wireless card just to see it turn on and give me the very much hated error code lights and promptly shutdown. it didnt even have time to get into windows. >.< :cry: :cry:

so ill give it some time and start removing non essentials.

 

Wait a minute,
did you go to move your laptop while it was on?

If you did, and noticed it shut off when you moved it, than the problem is a connection issue somewhere along the lines.

Core temp reads the peak temp for AMD's as like 60C. That is not true. All socket 939 processors can be run up to 100C before any damage occurs.
However the cooler it runs the better. When I have my fans on full. My system peaks at around 136F, which translates to 58C
My idle temperatures are around 90-95F, or 35C

You should install CPU Rightmark when you get your system running properly, so you can undervolt your processor. AMD's undervolt really well, I knocked about 12 watts off my peak cpu usage by undervolting.

Other wise take both memory cards out of your system.
Look at the black memory chips on both cards.
Make sure the numbers on both are the same.
If they are different that is the problem. You need a matched memory kit or else your system will force dual channel mode and fail because both memory cells are not 100% the same.

To post an image, click on the "image" button in the post section. Than just copy the link from photobucket in. The Image button looks like a small mountain on a yellow background

K-TRON

 

Ok the screen shots are linked below

And the computer was off when i moved it.
Also those temps you listed sound about right for my laptop, but what is your Tcase max set to ? youll see it in the screens for mine.

 

Also if you look closely at the Core temp for the graphics on that chart it shows it at -157*C ?? so im guessing something is wrong there? also before i said i had a CPU temp spike on Core#0 for a split second that went to 260*C do you if that is a problem or was it probably just a glitch.

 

Your graphics is reading -157C because you did not apply the 200C offset. So in reality yours is running 43C, which is not bad at all.
You can apply this offset in rivatuner:
Click the charts button in rivatuner to bring up the monitoring screen.
At the bottom click "Setup"
Scroll down to "core temperature"
and click "setup"
In the top middle of the window lies the offset. Type in 200C

The jump in core temperature is simply a polling error. It may jump to 260C for one second, but than right back, as it corrects itself. The temperature monitoring is pretty accurate. core temp reads the same values as my infrared gun. My cpu sometimes jumps to 137C for a microsecond and than jumps right back down. It is just a thermal probe polling/read error.

Core temp reads my TCASE MAX of my Opteron 185 as 63C.
But if I am not mistaken, the AMD Spec Sheet says 80C as the max

K-TRON

 

Well then is it possible that the Tcase max being set wrong it my problem and the laptop thinks it is getting too hot and thats why it is shutting its self down?
And if so how do i change it? im assuming it is in the Bios which i havent updated yet. >.< if you can find it in your heart to post a link to a place i can download the Bois version you suggest i would greatly appreciate it.

thnx again and i hope were close to a solution.

 

Coretemp cannot override the onboard cpu shut off values. The system is set to shut the CPU off at 100C
I dont think you can change the Tcase Max whic core temp reads, because the program is reading it wrong

K-TRON

 

do you know where i can find out where i can find the exact ram that i need for this laptop. Im seeing some online that looks like it would work for about $22 so if it is that cheap then im just going to go get some just to be sure that the ram chip themselves arent the issue.

 

What memory do you have in it?
You need identical memory cards. You need to have two identical cards with the same numbers down to the individual black memory chips on the card

I am running two Corsair VS1GSDS400 cards in my laptop, and Gophn is using two 1GB Kingston cards

K-TRON

 

Did you know what this might have meant by the way, i read your post afterwards and im not sure if you said what it was. But i am about to test the ram again the way that you suggested to see if i get the same problems.

 

Usually the system picks up the change in memory on the fly. I am not sure why you got that error beep, it is probably just an anomaly.

K-TRON

 

So i just need 2x PC3200, 200 pin, SODIMM, 1GB @ latency 3-3-3-8
?

 

So i just need 2x PC3200, 200 pin, SODIMM, 1GB @ latency 3-3-3-8
?

 

Yes, but they need to be identical. Buy a paired kit.
For example, I have 6 Corsair 1Gb Pc3200 memory cards. only two of them have matching numbers on the memory cells. If you use unmatched memory cells than the system will shutdown whenever dual channel memory is needed, which is almost all of the time with the AMD socket 939 processors

K-TRON