W500 - Shuts down within a few seconds after reassembly

[nirvana80]

Hello, I'll start with saying that I'm going completely mental with an issue with my W500.

I'm no newbie with computer HW and SW so I decided to task myself with replacing the yellow power connector which had become loose over time.

I bought the correct connector from ebay disassembled the W500 and reassembled it, turn it on ones logged into windows all good, then shut down and went to sleep, and I thought great.

Following morning went to power it on and didn't power on. Disconnected the battery tried again and it was on again, however after a few seconds or after the first reboot it would just shut off instantly and decide not to power on again until I removed either the power supply or the battery.

Opened it again, reassembled it, and things got a little worst in the sense, that now the laptop does the same but it only stays on for a few seconds normally even before the laptop gets a chance to start booting the operating system.

If left alone for a while, it will get a little bit further and start booting the OS only to either freeze or shut off again refusing to power on until the power had been removed again. tried again disassemble, removed individually components to try and isolate the issue but nothing same behavior.

So I thought, that's it, the motherboard is damaged, although I wasn't convinced because the issue to me could have either been related to a overheating issue, although not completely likely cause the laptop would only be on for a few seconds before it was permanently shutdown and in refuse to power on mode.
Can components under the heatsink fan overheat that quickly???

I thought maybe there is some sort of short circuit or something exposed, also unlikely cause I checked everywhere.

If you've managed to read so far my story gets even weirder now

Cause I had to isolate a motherboard issue, I decided to order a used motherboard online. I bought an half w500:

Full Base (motherboard, cpu, and wireless card, modem, etc, keyboard, mouse pad)
Not included (HD, RAM, LCD)

Anyway I took the motherboard with CPU, etc + the main gray inside structure and put it inside my bottom plastic (in a better state).
I reassembled the laptop neatly and it still does exactly the same and I was like WTF is going on .

I then reassembled the old parts with the left over of the new base I built and tried to boot it on external screen, it boots for a few senconds then again same symptom, it just shuts off and refuses to come back on until power is removed.

So now I have 1 and a half laptop, technically I could use them both if I only manage to resolve whichever issues is causing the laptop to shut itself down and remain off until power is removed (a protection mechanism against HW damage I suppose).

Sorry for the long post but does anybody have any idea on what I might have done wrong during the reassembly of both laptops.

I just wonder why I didn't try to power on the base I bought in the state I received it and test it (unfortunately didn't think about it), all I had to do was to put some ram in it to get to boot it into the bios.

I wonder if it could be an overheating issue, but only a few seconds from cold, I really doubt.

I am left with nothing else to isolate or test, (oh I didn't mention I changed RAM as well so it's not faulty RAM).

Thanks in advance to anybody who can give me an advice.

Nirvana80

[RealBlackStuff]

Welcome to the forum.
Did you clean up and replace the thermal paste between the CPU and the cooler? (I guess not...)
If old/dry/missing, the laptop can/will overheat in seconds and switch off.

[nirvana80]

Hello RealBlackStuff,

Thanks for your quick reply, much appreciated.

I did not replace the thermal paste as I saw there was still some and an overheating issue within 10 seconds seemed far fetched. I'll try and update the forum.

Thanks

Nirvana80

[holoman5]

I had this same problem with two customer's W500. I fixed the issue on one by updating bios. The other had mild corruption in windows and a fresh install fixed it. Hope this helps you.


_____________________

W520: I7-2760QM, 32GB, Crucial 240GB SSD(OS); Seagate 1TB SSHD (Default Files Location), Docked, NexStar 8TB Ext. Array.
W500: T9800, 8GB, 500GB Seagate SSHD
T510: I7-640M, 8GB, 1TB Samsung
T410: I5-560M, 8GB, Seagate 500GB
R500: T9800, 8GB, Seagate 320GB

[holoman5]

Forgot to mention that i used a intel 575 cpu to do the above fixes then reinstalled customer's cpu.

[nirvana80]

Ok I applied the paste and things started working again, and as long as I use Integrated Graphics is all good, however as soon as I switch to Discrete Graphics I get overheat problems again, however this time is different as I manage to boot into windows and then the computer just goes off.

The main difference is I don't need to pull the power to get it going again, it just powers on again and if hot shutsdown again.

If I let it cool and switch back to integrated graphics no problem, no overheating.

Obviously the problem now is with the Discrete Graphic GPU.

I was looking at the manual http://download.lenovo.com/ibmdl/pub/pc ... 3y6630.pdf and on page 89 it shows where I should apply the paste.
In my case the lower picture, is shown as a rubber rubber material (which is the case) in which I'm not supposed to apply paste, however that is supposed to be the Integrated Graphic Chip if I understood this correctly, with the discrete graphic being shown as [a left]. I just wonder why in the integrated graphics model thermal paste needs to be applied in the same area where and the rubber is in the discrete graphics model.

Anyway the chip shown under the rubber is that the inegrated graphics in the discrete graphics model. should I not apply paste under the rubber?

I will reopen and clean reapply paste again I just need to understand which of the three is getting hot (bottom b, or left a).

By the way I did upgrade the bios to the latest.

Thanks a lot again for all your advices.

[RealBlackStuff]

Many people overdo the amount of thermal paste, or do not apply it correctly.
FYI: best instructions here: http://www.arcticsilver.com/pdf/appmeth ... d_v1.1.pdf

The HMM part you referred to is 1150 Fan Assembly on pages 108-111 (116-119 in Foxit).
A discrete GPU model only needs paste on CPU and ATI/AMD GPU (next to it), and not on the Intel GPU.
What happens in your case is either an "overload" of thermal paste, not enough paste and/or bad contact between fan and GPU.
If you take it apart again, clean CPU and GPU, apply fresh paste, put the fan on with all the screws and brackets, then immediately remove the screws/brackets again and lift the fan off. Check that the paste on the fan-bottom matches the size of the CPU and GPU surfaces.
Most likely the fan's GPU surface is 'incomplete'.
In that case you need to (extremely carefully) adjust that GPU part so it sits absolutely flat on the GPU.

[nirvana80]

Thanks I'll try that, and thanks for the guide.

[nirvana80]

Ok so I disassembled and reassembled it again. The GPU overheat seems to be gone now.

I'm monitoring things with HW monitor, CPU Temperature stays around 50C - 55C. however if I run a benchmarch eventually the pc goes off.

Ati Mobility Radeon GPU stays around 65C.

I'm thinking maybe the paste has to set in, cool and warm up as I've read on some sites.

Should I open the laptop again and try again? Should I accept a 5 minutes prolonged 100% CPU possiple overhaeaing.

Is there any thruth to paste settling in and going through off and on cycle

[Neil]

I'm not sure I'm following just what you have been doing here. Have you already applied fresh Artic Silver compound when you "disassembled and reassembled"? If not, then do so, and I think you'll see the temps stay where they should. Assuming your fan is working properly, that is.

[nirvana80]

well I applied a compound but not Artic Silver 5, cause I found out about Artic silver only after I had already purchased this one:
http://www.ebay.co.uk/itm/221016709179? ... 1497.l2649

Is Artic Silver 5 really going to make that much difference that I should bother buying it and replacing it?

Thanks

[Neil]

Many folks around here wil argue that Artic Silver is much better, but I was unclear as to whether you had applied fresh thermal compound of any kind or not. Any kind of fresh compound is likely better than leaving the old dried up stuff on the components, but I can't honestly say whether Artic Silver wil make "that much" of a difference or not.

[nirvana80]

I suppose I'll have to join to the many folks out there, as after applying the Artic Silver 5 (with Artic Silver liquid to clean and remove the surfaces of the chips) I noticed a huge difference in terms of cooling and machine stability.

The cheaper compound I previously applied still made my machine heat quite badly and in some cases shutdown when put through CPU stress test with CPU at 100% at prolonged periods of time.

Things are running much better now and I can see an increased level of performance as well during normal use.

Even after I applied artic silver 5 I could see the compound looked much stickier and covered the surfaced more appropriately then the other compound did (the other compound being a bit watery and all (it must have been a pretty bad one).

Thanks all for your help.

Not having done any HW maintenance on computers for quite a long time, I forgot how important the use of fresh cooling compound on heat sinks is, I'm sure next time I won't underestimate it . Lesson learnt.

[precip9]

Artic Silver is not the solution to this problem. Colloidial silver is hazardous to use, because if a little gets on the board, electromigration occurs, with formation of microscopic silver whiskers that short things out. It is true that silver paste has the highest thermal conductivity, but the laptop was designed to work without it. There exist nonconductive compounds, made of ceramic particles, that are inherently safe.

The "rubber" material covering the GPU is not rubber. It is a gel, used because it requires less mechanical precision. Since I don't have the gel, I replace it with a typical heatsink compound.

The temperature you give, 55C is much too high for an idling W500. Is that an idle reading? I see 30-35C on mine.

The stainless steel spring plate that maintains pressure on the heatsink where it contacts the GPU and the Intel system chip is a little tricky to reattach. Misinstallation can cause overheating of those chips, but not the CPU.

No fancy chemicals are required to get 30C idle. Scrub the chip surfaces with drugstore 91% alcohol. Use any decent compound. High temperatures are the result of assembly error.

[nirvana80]

I'm not sure if I've achieve maximum contact between CPU / GPU and Heath sink, however things have improved a lot.

I do not use discrete graphic cause the chip still reaches around 60-63, it doesn't reboot the computer but it gets quite warm.

When using integrated graphic though things run quite smoothly, CPU is at 40 when idle, and 55 when put at a CPU stress test.

I'm just reporting on the improvement between my cheaper compound and artic silver, I'm aware that spilling conductive compound over the board might short components and fry the board.

In regards to the GPU, it might very well be that I haven't achieved complete surface contact, however this is the best I've got to and I'm not sure how else I can improve surface contact really. I do have a second half W500 (without a screen) as a result of my useless purchase, I'll assemble that and use it as a VMware ESXi machine, we'll see how that one goes.

Anyway thanks all again for your input.

[webgen]

I have similar problem with T500, my integrated card has a problem after assembly, I can confirm it is not because of thermal paste or overheating. It works fine on discrete graphics but on integrated graphics it shuts off just after it gets off of BIOS.