T61p graphics problems

[abarbieri]

From the BIOS Release Notes:

<2.07-1.08>
BIOS: 2.07 / ECP: 1.08
- (New) Added the new error code 1830 to indicate that a single DIMM
is not installed in the memory slot 0.
Solution for the error code:
With a single DIMM, install the DIMM into the memory slot 0
(a slot different from having used it).
- (Fix) Unexpected interrupts from the USB controller may occur.
- (Fix) The system with NVIDIA graphics chip may cause hang.

<1.27-1.08>
BIOS: 1.27 / ECP: 1.08
- (Fix) Bluetooth (WPAN) LED not turned off even if a BDC (Bluetooh
Daughter Card) was not found.
- (Fix) POST 1805 error due to unknown Wireless USB card
- (Fix) Intel Video driver is always loaded as maximum brightness
even if user set minimum brightness.
- (Fix) On PC Doctor for Windows PE, Multicore test result with CPU
T7800 system is fail.
- (Fix) Blank display occurs when the system boots up with LID close,
which an external display is attached.

With regards to the variety of problems when running intense 3D apps this update didn't fix things in my case


one interesting thing to notice is that now if one uses a single memory module only slot 0 (the one nearer the motherboard?) can be used.

in a previous post in this thread it seemed one could use any slot in this situation (one memory module).

[Hellbore]

No, I actually have asked Lenovo this question, you can not put 1 memory module into the slot farther from the motherboard. If you are using only 1 memory module, you must put it into the slot closest to the motherboard.

Before this bios update, the computer would let you boot up with 1 memory module in the farther slot, but next time you rebooted (soft reboot), the computer would hang with a BIOS beep code and lock up. The beep code would be the memory error beep code.

Now I believe with this new bios version, it has a new error that will tell you right away if you put the memory in the wrong slot, instead of booting 1 time and then telling you the memory was bad. It is just clearer and more correct error reporting to detect the memory module position.

[Florian]

First of all I want to confirm the fact that bios 2.07 does not solve the problem; even worse.

My T61p (T7500 2x1GB ram, FX 570M 256MB) was manufactured on 2007-10-26 and it is an excellent machine.

I had the chance to play Crysis demo on the same kind of TP (even the memory type), one month older ... but this game can not even start on my machine initially.

Probably I never noticed that it has a problem if it were able to run Crysis. I can run my favourite old games, Unreal Tournament 2004 and Return to Castle Wolfenstein, without any problems under Vista 32 and XP, both clean installs. Now I run XP.

I have managed to start Crysis after I enabled Intel virtualization technology ... strange. It simply stops in the first stages or, if lucky, I can play for 5 min. I noticed two things: sometimes a colored bunch of pixels appear on the top of the screen and I have to shut down the machine, and, if I am able to close the game with ctrl+f4, the bluetooth swith to on ... this is strange, also.

For about a week I have tested everything possible, as follow:
- Nvidia drivers from Lenovo and Windows update (no fix);
- bios, display to PCI express ... (no fix);
- bios, single core (no fix);
- bios, Intel virtualization technology enabled (initially solved the problem because I was able to start Crysis but no fix in the end);
- bios, every device closed ... network cards, usb, etc. (no fix);

The only fix is to run on single RAM card. The last bios (2.07) let me use only slot 0 and seems to slow down the game.

I think Lenovo guys know well the issue but they just do not care. If you noticed something they solve the problem by changing the mobo. If not, good for them

I am not so happy to change the motherboard ... you know, it is like repainting a new car.

[alacrityathome]

Florian,

First of all, it is shame that you are having this frustrating problem.

You should not be having these problems.

I am sure Lenovo definitely cares.....but may not know themselves how to resolve it for all PCs. They do know that a full motherboard replaces may fix the problem.....but they may not know why that is the case.

Unfortunately, I see that you live in Romania.....but I still recommend you have Lenovo replace the mobo if you can spare the lack of PC usage.

Lenovo needs to make your situation a good one.

Good luck!

[Florian]

I am wondering what RAM brand do you have in your ThinkPad. Maybe it is only a compatibility problem. I have "hynix" memories (1GB 2RX8 PC2-5300S-555-12).

Is there a t61-t61p specification for memory compatibility?

I still can not believe there are different kind of motherboards or some with a hidden defect ! How should I know if replaced, the new mobo will perform as it should?

[alacrityathome]

Florian,

I think the forum users in different threads have exhausted the possibilities with the RAM variations. I don't think the type makes any difference.

Hopefully Lenovo is well aware of this issue and when they replace the mobo, they are also testing more heavily for 3D intensive applicaitons.

The mobo reason could be something as simple/complex as a cold/bad solder joint. Very difficult to find or confirm but occurs with heat changes like gaming use.

[Hellbore]

The mobo reason could be something as simple/complex as a cold/bad solder joint. Very difficult to find or confirm but occurs with heat changes like gaming use.

This happened to a TON of Xbox 360's, it could happen to T61P's... It's possible! It might explain why they are having souch a tough time finding the issue. You could have 2 identical motherboards and one works, while the other doesn't, and you would have a beast of a time figuring out why...

[thomas565]

ok, mine ran bioshock demo without error. tested it for an hour at 1920x1200 and all possible other stuff 'on'. i have he mushkin pc2-6400 2x2gb from newegg. using bios 1.21, but will upgrade at some point. was wondering if lenovo ships pressed CDs every once in a while to get the whole package of, first the bios, then drivers, linuxdrivers.

tomorrow i will see if I can get the tests from www.spec.org to run (linux and xpp) and if it gives me something meaningful, seems its a hardcore test. did you guys use the free demo versions of burnintest and 3dmark to demonstrate the problem?

suse64 10.3 installed ok with some glithces solved in other threads.

I have to say that the display is impressive, I like it, have only an 8yr old yellowish 1400x1050 to compare with.

[alacrityathome]

Toby,

Congratulations. Sounds like you have a good one. I was also impressed with my display and confirmed it was a Philips. Crisp and brilliant.

For testing, I started with PCDoctor video tests, then 3DMark05, then 3DMark06, then Gaming without crashing, then Gaming w/FRAPs.

Bioshock, Crysis, C&C3.

Good luck.

[dikrek]

I run Vista 64 and have 4GB RAM.

Before installing the new BIOS I was running 1.26.

Crysis would crash after loading and asking me to press the ANY key.

After flashing, I powered off, then powered on, went to setup, loaded the defaults.

Now crysis works MUCH better but still crashes eventually.

So, the BIOS does do something useful, but doesn't quite fix the issues.

D

From the BIOS Release Notes:

<2.07-1.08>
BIOS: 2.07 / ECP: 1.08
- (New) Added the new error code 1830 to indicate that a single DIMM
is not installed in the memory slot 0.
Solution for the error code:
With a single DIMM, install the DIMM into the memory slot 0
(a slot different from having used it).
- (Fix) Unexpected interrupts from the USB controller may occur.
- (Fix) The system with NVIDIA graphics chip may cause hang.

<1.27-1.08>
BIOS: 1.27 / ECP: 1.08
- (Fix) Bluetooth (WPAN) LED not turned off even if a BDC (Bluetooh
Daughter Card) was not found.
- (Fix) POST 1805 error due to unknown Wireless USB card
- (Fix) Intel Video driver is always loaded as maximum brightness
even if user set minimum brightness.
- (Fix) On PC Doctor for Windows PE, Multicore test result with CPU
T7800 system is fail.
- (Fix) Blank display occurs when the system boots up with LID close,
which an external display is attached.

With regards to the variety of problems when running intense 3D apps this update didn't fix things in my case


one interesting thing to notice is that now if one uses a single memory module only slot 0 (the one nearer the motherboard?) can be used.

in a previous post in this thread it seemed one could use any slot in this situation (one memory module).

[abarbieri]

I run Vista 64 and have 4GB RAM.

Before installing the new BIOS I was running 1.26.

Crysis would crash after loading and asking me to press the ANY key.

After flashing, I powered off, then powered on, went to setup, loaded the defaults.

Now crysis works MUCH better but still crashes eventually.

So, the BIOS does do something useful, but doesn't quite fix the issues.

D

so... really it does not resolve the NVIDIA graphics problems once and for all, since "almost" a fix is not a fix.

from my current limited experience since the new BIOS I tend to get now more of 'the graphics device driver stopped responding' and the Vista 32 OS manages to recover rather than freezes, BSODs and reboots.

[SnakeJW]

Ok, I thought I should give you guys an update on my case. I received a new mother board last friday and lenovo send someone over to my office and finished replacing the board. I had the same 2x2G double stacked memory originally came with the laptop. I just ran some test after the mobo swap, everything seems working fine now, no more crashes and BSOD. It could be the problem with the mobo, but it could also be the video card that is fused with the mobo, and I'm not sure which one is causing the problem. But the memory probably is not the cause since I'm using the same ones.

So I think this problem is definitely hardware related, and Bios update probably won't fix the problem.

[abarbieri]

...
So I think this problem is definitely hardware related, and Bios update probably won't fix the problem.

thanks SnakeJW for the update... it all seems to be pointing at a bad batch of motherboards unfortunately

well, lets see what the guys at Lenovo come up with.

see:
http://forum.lenovo.com/lnv/board/messa ... read.id=55

please provide as much feedback as possible even if it is a repetition of what has been said/reported here.

[dikrek]

I run Vista 64 and have 4GB RAM.

Before installing the new BIOS I was running 1.26.

Crysis would crash after loading and asking me to press the ANY key.

After flashing, I powered off, then powered on, went to setup, loaded the defaults.

Now crysis works MUCH better but still crashes eventually.

So, the BIOS does do something useful, but doesn't quite fix the issues.

D

so... really it does not resolve the NVIDIA graphics problems once and for all, since "almost" a fix is not a fix.

from my current limited experience since the new BIOS I tend to get now more of 'the graphics device driver stopped responding' and the Vista 32 OS manages to recover rather than freezes, BSODs and reboots.

Actually I tried Half Life 2 and, while the laptop gets hellishly hot, it does NOT have issues of any kind. With the graphics settings at max.

So, it's also app-specific?

D

[alacrityathome]

dikrek,

I think HL2 is less 3D intensive than Crysis.....at least that is what the forums say. And, to support that, HL2 runs on older video cards.

But, one thing to try is to run FRAPs and ATI Tools to monitor both the fps and the graphics card temperature to see which game is pushing the FX570M. fps starts dropping and temp starts climbing.

I run HL2 Deathmatch at high settings with no problems.

Let us know what you find out.

alacrity

[dikrek]

dikrek,

I think HL2 is less 3D intensive than Crysis.....at least that is what the forums say. And, to support that, HL2 runs on older video cards.

But, one thing to try is to run FRAPs and ATI Tools to monitor both the fps and the graphics card temperature to see which game is pushing the FX570M. fps starts dropping and temp starts climbing.

I run HL2 Deathmatch at high settings with no problems.

Let us know what you find out.

alacrity

I run it at 8x aniso + 4x antialias, doesn't even run fast like that so I'd think the card is heavily pushed. But I'll try the others.

Edit: I tried Bioshock again and it crashes my PC (bluescreen) while loading, at least Crysis is more elegant in the way it crashes.

D

[SnakeJW]

Could it be that a batch of bad video card is causing the problem? Replacing the mobo is actually swaping the video card?

[SnakeJW]

ok I searched some chinese forums, and it looks like they are having the same problem as we do. And it's not a new one, T40s and T60s all have similar problems. One thing they found out was there's some soldering defects with the graphic chip. Mainly, the assembly line didn't solder the chip firmly and cause the symptoms very similar to ours.

[alacrityathome]

ok I searched some chinese forums, and it looks like they are having the same problem as we do. And it's not a new one, T40s and T60s all have similar problems. One thing they found out was there's some soldering defects with the graphic chip. Mainly, the assembly line didn't solder the chip firmly and cause the symptoms very similar to ours.

Very interesting. Yes, that would definitely cause the problem! Do you speak Chinese or are the Chinese forums in English?

When I got my paperwork last week with my T61p, I had the impression my PC was made in Singapore. Maybe that's why I have not seen any GPU problems. I wonder if the T61p is made in different countries?

[SnakeJW]

Yes, I'm native chinese and mine was made in China. The date on the roll cage marked Oct 11, 2007 as manufacture date.

[abarbieri]

ok I searched some chinese forums, and it looks like they are having the same problem as we do. And it's not a new one, T40s and T60s all have similar problems. One thing they found out was there's some soldering defects with the graphic chip. Mainly, the assembly line didn't solder the chip firmly and cause the symptoms very similar to ours.

this looks more and more an actual hardware problem that is triggered by the presence of two memory modules.

because the graphics hardware is integral part of the motherboard, nothing but a motherboard swap will cure the problem...

let's see if the Lenovo engineers will confirm this hypothesis.

[alacrityathome]

Great input to this forum, Snake! I think it explains a lot for the GPU problem. Thanks for taking the time to read the Chinese forums and post here.

[YinLe]

I'm using T61p 6457AK1 also. Do u guys mean only 3D intensive games could cause the crash? I played some games like PES2008 using the highest quality, everything seems OK.

[alacrityathome]

YinLe,

Not sure how intensive PES2008 is. You can pull down 3DMark06 for free and test your PC capability.

With an intensive 3D game or 3D graphics test, you can actually see the GPU temp rise rapidly. When that happens, every thing gets stressed or tested on the motherboard/graphics card.

Higher temp is not a bad thing.....but a bad solder joint will be tested at the same time and fail.

[Hellbore]

LOL... Probably cold solder joints on the GPU. Just like the Xbox 360 had. How much you wanna bet...lol...

Wouldn't it be fun to try reflowing the solder on one of these T61P's and see if the problem is solved?

I dunno if it's the GPU though, why would it only be an issue with 2 sticks of RAM? That part is what is strange to me. Hmm...

Meanwhile, over on the Lenovo.com forums, people keep desperately trying to make it a driver or BIOS issue. Sigh.

Snakejw you have been super helpful!

[Hellbore]

I wonder, has anyone tried turning on the Windows driver verifier? I don't think anyone has. It would be nice to turn it on, even though the game will probably run much slower, and try to reproduce this crash, especially if you are getting a hard lock-up where you don't get an error message. The driver verifier might help us gather more information about the specifics of this issue. I doubt it will help much though. Just something you might try. I almost wish I had a broken T61P to play with lol... but "unfortunately" Lenovo already replaced mine with a good one which has never crashed yet

[badblood]

Saying its hardware is partly ambiguous. It can be hardware if there's an inherent engineering design flaw which is pretty serious. Or it can be hardware also if it was shoddy assembly. If its assembly like suggested by Snake, then its an issue of QA and an easier fix I reckon. Either way it needs to be fixed.

[Hellbore]

Saying its hardware is partly ambiguous. It can be hardware if there's an inherent engineering design flaw which is pretty serious. Or it can be hardware also if it was shoddy assembly. If its assembly like suggested by Snake, then its an issue of QA and an easier fix I reckon. Either way it needs to be fixed.

Yeah I wonder, I wish we could collect some information about the motherboards having these issues to find out if the bad ones come from a particular source, or if they are of a particular revision number or if the GPU's are from different runs, etc... I hope that's what Lenovo is doing! It would sure be nice to be able to identify which are bad without having to actually play games and look for a crash.

I dunno about you but I feel like it's more likely a QA issue not a design flaw...

[alacrityathome]

Hellbore,

In the interest of Lenovo's own profitability, i am sure they are trying to dissect this problem.

I would bet that there are multiple manufacturing locations and one of them is the culprit. Date codes are critical to narrowing down the manufacturing location and culprit dates.

For some reason on the paperwork I received, I had the impression my Pc was manufactured in singapore which, if so, may have given it a higher manufacturing quality. Maybe Snakejw can ask on his Chinese forum if there are multiple locations for the mobo assembly.

Alacrity

[Hellbore]

Well, Toby909 told me that there is a factory code, on one of the stickers on the bottom of the laptop, which tells you which factory your Thinkpad came from. His and mine were both factory code H, and both do not have the problem. There is another Thinkpad I can look at that did have the problem originally, I could see what factory code is printed on that unit. Also you guys could check your factory codes and see if there is any relationship. I kind of doubt it though.

That might just be the factory the machine was assembled in, and one factory might get parts from more than one source, or more than one factory from the same sources, or only some parts from a source might be bad, etc. It seems to me like it would be more important to know where your motherboard came from and other specifics about the board itself.

There have been a few people who had this problem and sent their machine to the depot, then the depot replaced the motherboard, and after that the unit worked fine. It is too bad that the depot did not record all the details about these machines so that this information could be used by Lenovo to track this issue. I wonder what they did with the motherboards that were bad? Surely someone must keep track of this stuff!

Later today I am going to test this other laptop (not my nice new one) that was originally bad, the depot replaced the motherboard but I don't know yet if the problem was solved. I almost hope it is still broken so I can participate in the detective work Ever since Lenovo got me a new laptop that works perfectly, I haven't been able to experiment with this issue any more becuase I don't have the problem now lol...