NVIDIA Graphic Card disabled - error Code 43

[stingray77 4]

The screen on my HP-notebook started flashing and I needed to restart.

At first try the symptoms were gone - but soon they came back...

Unfortunately this screwed my whole display (two identical / mirrored displays on screen with weird colors, icons way too big and everything definitely out of order), not only in Windows (Vista) but on booting already.

 

Investigating showed the GC (GeForce 8600GS on latest drivers) disabled and an error code 43.

 

Google brings up several of these errors - and proposed solutions to revert back to older drivers or replace GC...

Reverting does not really seem to work (and the distorted screen already at booting makes me believe more in some hardware-problem), but do you have any other recommendations / suggestions before I send my notebook in?

 

(Unfortunately this is 3 mths after the warranty expired...)

 

I hate not being able to fly.........

 

 

Here's how the mess looks like

Edited September 17, 2010 by stingray77

[xclusiv8 35]

looks like a temperature problem to me. is the lappy hot?

[JediMaster 451]

You could have a failed fan or the heatsink came loose on the GPU or its RAM. If you get distorted graphics BEFORE you load windows, just in the BIOS, it's 100% a hardware problem.

 

The 3 months OOW part is miserable, because this looks like that's exactly what it needs...to be repaired at the hardware level.

 

 

[stingray77 4]

Not more than it was before - and the fan didn't kick-in harder or longer than before (not that I noticed much a difference)...

 

I often read about the temp reading, but how/where can one see the temp of it?

 

Then, after the problem remained after days of not using the machine, the prob remains as was - does that sound more like a permanent problem?

 

Btw: a guy in a computer-shop around the corner said, as it was a notebook, there was noch real graphics card in it anyhow that could be exchanged!?!? Hey, that thing even has its own NVIDIA specification and name as a "card" - I certainly hope that man just didn't want to dive into a deeper discussion.....

[stingray77 4]

...If you get distorted graphics BEFORE you load windows, just in the BIOS, it's 100% a hardware problem. ...

Okay, that's not what I wanted to hear, but pretty much reflects what I was thinking.

 

Right after pushing on/off I am getting a green screen horizontally divided even before ANYTHING appears on the screen (that BIOS "menu") - all that was black before and (of course) not divided...

 

So, guess I'll have to pack the machine to send it in..... Darn f***in' s***!

 

 

Guess that will mean Stingray77 signing off for quite a while.....

But thanks for all your input.

[stingray77 4]

Oh, another question:

 

I always tried to back-up as good as possible - but you never know... From your experience: do the drives might get "damaged" while the machine is back at the yard?

 

There's quite a load of SF2V-data on one of my external drives, but all my FS9-data (including all of its registries) is on C, and only the original unzipped and not applied planes and sceneries are backed-up (as there was not much of an alternative - I failed to copy the flightsimulator-folder with all its add-ons)!

I can envision re-applying all that again to plain vanilla will take me a year.............

[JediMaster 451]

A small number of laptop models have a GPU-on-a-card that looks like a small circuit board that can be swapped out/upgraded. Most likely yours is not, it's not a popular option.

The fan is usually linked to the CPU temp, so if the GPU overheats you wouldn't necessarily have any indicator. I have an 8600GT on my Dell laptop, but it has just one fan.

 

Most likely there is a heatpipe arrangement that brings the heat from the CPU and GPU to the same place where the fan can cool it, hence my statement that perhaps the thermal connection between the GPU and heatsink has failed. If the laptop has suffered any jarring impacts lately, or been in a hot dry place, that could cause the thermal paste/pad to break down. When that happens, it's going to overheat within a second or two of power-on. The instances when it would work for a short while before failing was likely because the connection was still partly there, so it took a little while to overheat.

 

Your only (free) recourse is to disassemble the bottom of the laptop, find the GPU, and check the connection between it and the heatsink/pipe. Not knowing your model's design, I'm not sure how user-repairable it might be, to get a new thermal pad in there and reconnect it. That said, it may be "too late" for the GPU anyway, and while that would keep it from overheating again, it could already be permanently damaged.

 

 

 

As for data, they don't need your HDD to fix it. Most laptops have an HDD that is easily removed after you undo a couple of screws.

 

REMOVE THE DRIVE BEFORE YOU SEND IT AWAY. Otherwise, who knows what could happen?

[Slartibartfast 153]

As to the drives getting damaged back at the yard here's some advice

 

Advice 1 copy the Hard Drive...

 

Advice 2 Copy the Hard Drive

 

Advice 3 See advice 1 and 2...

 

You never know what they are going to try they may well strip the drive and reinstall to make sure its not a software issue... even though this is 100% a Hardware fail...

[stingray77 4]

Well I never opened anything on my notebook - and still am quite reluctant to do so... So I think I wont go the diy-direction, you describe - I VERY MUCH WELCOME THAT INFO THOUGH!

But I am quite sure, there's only one fan...

 

When reading your post, one of my first ideas was trying to cool down the machine in the fridge before starting it once again (adding some more temp-gap for the GPU to overheat) - although things like that never sound like pro-ideas, I must admit... So, I skipped that idea right away. The more as your post sounds like it might already too late for the GPU anyhow...

 

So, if there is no "seperate" card, it would need the whole circuit board (with the graphics-whatever-component on it) to be replaced?

 

Any input on the back-ups and affecting the drive(s)?

And (although this of course depends on shop and area) can I expect to receive my (unscathed and pampered) "design-notebook" (shiny surfaces) in its "as new"-condition, or do I have to expect some scratches and dents from the "usual works it takes"???

 

 

THANKS FOR YOUR INPUT! Though I am pitying myself a lot, at least I can leran a bit.........

[stingray77 4]

Ah yeah, cool advice.

 

When doing my first set of back-ups (at least I DO have an external data storage since 2 months now) I tried to copy all my SF2-stuff - but aside from showing some 22hrs of duration, after a while the system refused to copy any more data and showed my some wi-fi-connection-fail every 5 to ten minutes - had to click it active again.... That took me "ages" just to copy my final modded SF2-install!

 

But can it be so hard? On the bottom of my machine I see a door hiding the drive. Why dont they "simply" unhook it before doing ANY work to it???

[+suhsjake 11]

Just a heads up, I am dealing with the same problem. I have a Sager NP9262 with a dead 8800M GTX. Unfortunately, due to Nvidias subpar construction and use of shotty materials during the construction of the 8000M series which cuases them to die. You could try baking it (voids warranty), but if it is still under warranty have it replaced.

 

Of course mine died two months after my warranty expired and an 8800M GTX isn't cheap.

[stingray77 4]

Let's get drunk TOGETHER, pal... Yup, my warranty expired three months (or so) ago, too....

And I always understood NVIDIA to be one of THE top supplyiers of a part like that - sigh.

 

 

I think I wouldn't do that, but what's that "baking", you are talking of.....?

[+suhsjake 11]

Its because Nvidia used subpar solder in the construction of the 8000M series (they got sued over it too, but we missed that train).

 

Pretty much all you do is remove the card from the laptop. Remove any thermal grease and plastic bits and place it on four balls of aluminum foil (each holding up a corner) a cookie sheet lined with foil. Bake at 380F for ten minutes. Pull out and let cool. Reapply thermal grease and plastic bits and install into laptop to test.

 

Be forewarned, this isn't the final solution, but a temporary stave before you have to buy a new one. I have heard of people baking at least 3 more times and lasting another 6-8 months before the card finally bit the dust for good.

Edited September 17, 2010 by suhsjake

[stingray77 4]

1) thanks for moving to proper thread

 

2) thanks for explaning - I love building aircraft-models, but that's far beyond my reach... But good to hear/get to know about things like that...

When reading about "sloppy" driver design I already thought "man, there must be someone on their heels to deliver or pay back to their customer - the more so on the ("sue-everybody-for-nothing-minded") US-market"... But no?

 

When talking to the HP-support here in Germany I already got the info, that even if they replace a/that part there would be a warranty of 3 mths (instead of the usual 24 mths) only...

[stingray77 4]

Okay, took my rig to a local (Hamburg) repair shop around the corner yesterday. They confirmed this is a common "glitch" with the 8600 (and Pavilion DV9er series).

 

Relatively easy fix though: disassemble, take the graphics chip off the motherboard, resolder it back on, check/replace the heat sink(s), check and reassemble if everything checks out fine.

Takes around two weeks, cost below 200EUR* flat and that should provide me a reliable machine - after resoldering the chip, no such problem should occur again in future (don't know if he just told me, what I'd want to hear, but he sounded very convincing ). Sounds great to me.

He even showed me another mainboard of a DV9er he had at at hand, explaining what they are going to do - left a very good impression that they really know their job.

 

And they won't touch - say: they unhook - the drives for the repair (other than the shop doing the warranty repairs for hp here in Germany, who reportedly format the drives and send back the rig in "as was new"-state and condition).

 

Wish me luck guys, that my very good impression of their performance will keep up with what they deliver, and that everything will be a-okay in two weeks then.

 

 

* for THAT amount no way would I have been able to do anything near "repairing" or baking - the more when I saw the mass of those "micro-solders".

[stingray77 4]

Okay, after a little more than two weeks of "down time" I received my notebook back for "testing" over the weekend.

 

Every looks good again, so far.

 

Now, I am a bit reluctant to restore all my before-the-crash-settings - like going back to the latest NVIDIA-drivers. (My machine went back to 195. or 197.something drivers, after manually deinstalling the latest 258.96)

As the problem was related to the chip running to hot and losing contact with the MB, my question is: does the driver influence the temperature the chip is running at? Or doesn't it have any effect whether I use rather old or the newer NVIDIA-drivers???

 

(Maybe I am a bit over-cautious right now, and that will droop over time again...)

[JediMaster 451]

There have been rare instances of "bad drivers" in the past that either disabled the onboard fan or made it run too slow or conversely made the chip work far harder than it should.

Those have been pulled from distribution AFAIK, so you should be safe with the latest WHQL nvidia drivers or whatever HP has for DL.

[stingray77 4]

HP seems to bit slow there: the only one available to download from their pages is of early 2009 (or 2008) - whereas the latest ones from NVIDIA are not older than three months...

 

So, listening to what you're saying, I will give the 258 "another try".

 

Thanks again for your help/advice.