Laptop upgrade advice try #2

[Per Hansson]

Re: Laptop upgrade advice try #2

I'm sorry RD but a metal spacer will not under any circumstance become thicker as it gets heated
What it might become is slightly wider but that will be negligible at the temperatures we are talking about

A metal spacer has another advantage btw, if made slightly too thick (in your case they should all be the same extra thickness of course)
This will put some extra pressure on the BGA under the chips so that if bad solder was used it might make proper contact again...
I have a crappy Acer here with that issue on it's ATI chipset, damn RoHs

[PCBONEZ]

Re: Laptop upgrade advice try #2

I'm sorry RD but a metal spacer will not under any circumstance become thicker as it gets heated
What it might become is slightly wider but that will be negligible at the temperatures we are talking about

A metal spacer has another advantage btw, if made slightly too thick (in your case they should all be the same extra thickness of course)
This will put some extra pressure on the BGA under the chips so that if bad solder was used it might make proper contact again...
I have a crappy Acer here with that issue on it's ATI chipset, damn RoHs

Actually he's kind'a right.
It will get 'thicker' when it heats:
... but to such a minute extent that it doesn't make a lick of difference.

While were at it:
The integral heat spreader on the IC chips gets thicker, the heat sink does too, and the retaining screws get a tiny bit longer.
All of those will have much more total expansion than a thin-ass shim.
- None of the above is enough to matter if it's properly assembled using a thermal compound.
[Above -and- below the shim if a shim is used.]
Thermal compound ALSO expands and contacts to keep full contact.
That's one of the reasons it's used.

~
You know, despite the fact that integral heat spreaders in IC chips do in fact expand and contract a TINY bit with temp changes I have yet to see one fall off a chip due to thermal stress cracking at the bond.
- And they are way thicker than a shim, meaning there is more there to expand.
.

[Th3_uN1Qu3]

Re: Laptop upgrade advice try #2

I don't know about the Linux drivers, but the Nvidia Windows drivers automatically vary the clock speed of the GPU as needed.

Yes, and they produce some big ass DPCs when doing so... which screw up any realtime tasks, especially sound. It took me 2 damn months to figure what was doing it, the sound skips and stutters were driving me mad.

It doesn't help that the two low power states are tuned so well that having Aero enabled switches between them every 30 seconds. I have created profiles for desktop and games, using the very-buggy-but-at-least-it-does-the-job nVidia Performance utility. That made life a lot better.

[Per Hansson]

Re: Laptop upgrade advice try #2

What I meant was that it will expand more on the sides than on the thickness
On the thickness we are talking way less than 0.001mm
On the sides maybe it would be 0.001mm maximum difference if heated from room temp to 90°C

[PCBONEZ]

Re: Laptop upgrade advice try #2

I didn't think you were really confused, but the way you said it....

For kicks I got the real numbers:

Expansion Coefficient of copper: 9.3 x 10-6 in/in per 1F
Assuming 100F temp rise: 70F [21C] to 170F [77C] = 100F
Assuming a 1" dimension on the copper:
.0000093 x 100F = .00093 inch
=> New dimension 1.00093", a gnat wouldn't notice the difference.
---
So lets say the shim is .010 inch thick.
Divide number for 1" by 100 to get number for .010" => .0000093 inch thicker.
~~ Whoop dee doo
Dust is said to be .00002" in diameter so a particle of dust is twice as thick as the shim would change in 100F.
.

[ratdude747]

Re: Laptop upgrade advice try #2

I didn't think you were really confused, but the way you said it....

For kicks I got the real numbers:

Expansion Coefficient of copper: 9.3 x 10-6 in/in per 1F
Assuming 100F temp rise: 70F [21C] to 170F [77C] = 100F
Assuming a 1" dimension on the copper:
.0000093 x 100F = .00093 inch
=> New dimension 1.00093", a gnat wouldn't notice the difference.
---
So lets say the shim is .010 inch thick.
Divide number for 1" by 100 to get number for .010" => .0000093 inch thicker.
~~ Whoop dee doo
Dust is said to be .00002" in diameter so a particle of dust is twice as thick as the shim would change in 100F.
.

reason i brought it up was thermal expansion is the reason the nvidia series 8 bug happens in the first place. the epoxy could not resist and thermal expansion breaks the connections... it all comes down to a poor choice in epoxy.

[toastygoodness]

Re: Laptop upgrade advice try #2

how about making your own heatsink? your uni has a shop with mills right?

jk with yeah, but i did make my own gpu heatsink in the shop, was kinda fun.

[ratdude747]

Re: Laptop upgrade advice try #2

how about making your own heatsink? your uni has a shop with mills right?

jk with yeah, but i did make my own gpu heatsink in the shop, was kinda fun.

yes but i don't have classes that allow me access. back where I used go to high school in west lafayette, I had full access to a shop (I was one of the best machinists in the school)

biggist issue is the lack of mountpoints to mount heatsinks with.

[toastygoodness]

Re: Laptop upgrade advice try #2

what about with arctic ceramique? thermal glueish should be enough if the heatsink isn't intensely heavy

[ratdude747]

Re: Laptop upgrade advice try #2

what about with arctic ceramique? thermal glueish should be enough if the heatsink isn't intensely heavy

yeah... but due tot he laptop's design, it would have to be passively cooled. this is quickly becoming a moot point... with the laptop inthis good of condition, I don't wanna f**k it up over a silly crazy mod.

It runs fine now, If it goes poof ill get a new mainboard.

[dood]

Re: Laptop upgrade advice try #2

RD- my primary traveling laptop is a Dell d620 with the NVS 110M. We bought it used, and I've been using it heavily for 2 years with no problems.

Upgraded to 2gb of RAM, a 250GB 7200RPM WD Black drive, and Windows 7... I don't baby it, so it's a little beat up, but it performs flawlessly...

[ratdude747]

Re: Laptop upgrade advice try #2

odd, i read the d620 had the problems too yet the nvs 110m is a 7300 based design. that is very odd.

[Th3_uN1Qu3]

Re: Laptop upgrade advice try #2

All nVidia mobile chips from 6000 series up to 8000 series have issues.

[ratdude747]

Re: Laptop upgrade advice try #2

UPDATE

I was trying debian out yesterday and well, dummy here mistyped some stuff and nuked the xp partition. (the debian partition trainwrecked later, thatls another story),

I ended up putting vista home premium 32 bit sp2 and fedora 15 x64 on it. even with aero, it runs well...

[mockingbird]

Re: Laptop upgrade advice try #2

Let this thread die.

Don't use Debian, they're eons behind Ubuntu in their stable release, last time I checked.

[ratdude747]

Re: Laptop upgrade advice try #2

fyi it was debian testing... stable is equivalent to the ubuntu LTS releases.

once gnome 3 and its dependencies escape experimental and move on to unstable (sid), it may be be worth trying again... right now, it is just too buggy.

I did try ubuntu, it sucked. they still are dumping features onto their steaming pile without fixing half the bugs... and unity... FAIL.

[ratdude747]

Re: Laptop upgrade advice try #2

It's almost been a year... and I have a major update.

I saw this thread:


https://www.badcaps.net/forum/showthread.php?t=19719


and since my laptop was due for heatsink cleaning, I took a look... and To my surprise, it had a "10" chip in it. Combined with the fact that it had a 2010 dated HDD, I think that sometime in 2010 it bombed the GPU and was refurbed under warranty anfd it was given a new board with a "fixed" gpu (good underfill).

What a relief...

On a side note, I also upgraded the RAM in it... I happened to have a $20 Fry's gift card and I used it for a 4gb PC2-6400 kit... What an improvement (no more swap area slowdowns).

[mockingbird]

Re: Laptop upgrade advice try #2

It doesn't matter... nVidia never fixed the underfill/lead issues with any of their chipsets... They simply just stopped making them.

[ratdude747]

Re: Laptop upgrade advice try #2

Whatever. the chip looked to be in good condition (no underfill cracks or the like) and it has been very stable... I read that the lead issue was never fixed but the underfill issue was unoficcially fixed; many repairpeople have noted that while many 08 and 09 chips have died, they have not seen any 10 chips with underfill issues.

[mockingbird]

Re: Laptop upgrade advice try #2

It's not just the underfill... It's also the design and the lead composition... The chips weren't engineered properly in the first place to dissipate heat properly, and they used an inappropriate combination of solders for the leads and contacts. The undefill issue you're talking about is that the one they used was too "tacky" and wouldn't allow the die to "float" under high heat... I don't think you'd see the underfill cracking, there'd just be massive internal chip damage that you wouldn't see.