Reballing does NOT mean it'll last long!
I replaced the motherboard in my dv9000 4 months ago. The GPU was fine thank you (i had reflowed it once), but the wired LAN had died, the audio input was fried, and the keyboard was acting up (chipset issue). I specifically asked for another board, not a reball of mine. I got another board alright, but it didn't work on battery. Went back with it, turns out a little diode got knocked off the board when he reballed it... I got it exchanged and the new board worked. Anyway, the point is the board i got wasn't new, but reballed.
Turns out the 8400M i got with this board was only 128MB instead of 256MB i had on mine. Not wanting to return this board as well i said screw it, i don't game anymore so this oughta do. And i hadn't opened a single GPU-intensive game on it...
Until Monday evening. I had a friend over and he played some Red Alert 3 while i was playing some oldschool Carmageddon on the old laptop. RA3 worked fine for about half an hour, and then it crashed with a screen full of artifacts. Restarted, again worked fine for a while, then crashed again. It was getting late so we shut it down and went to sleep. On Tuesday evening, it started crashing at the desktop then died entirely. I called the guy who sold me the board and asked about the 6-month warranty... he started talking back at me. I was like "fine, will you get this damn thing off my hands then?" He said he isn't interested in buying it, he's already got 2 for testing, and he isn't selling those (since everyone already knows they're junk).
Fine. Screw you... Back to our old friend the oven. I love the smell of freshly baked nVidia in the morning...
Took the board out, turns i went a little bit overboard with the heat as the connector for the internal modem fell off. I just soldered it back on. I also put some cardboard in the case 
where the GPU sits so the heatsink is held on tight and presses on the chip, turns out there wasn't anything to support it, and the GPU heatpipe on the DV9000 is only held by two screws to the GPU. Looks like whoever designed this failed their physics exam.
Put the thing back together, working like a dream. Passed 20 minutes of FurMark with the overclocked settings that i was using with my old board. Maximum temperature the GPU reached was 76C, and then it went down to 70 as the fan kicked in harder. Unfortunately the fan is bound more to the CPU than the GPU temp, but i doubt this is the cause of failure - my old board lasted a good 2.5 years with a failing GPU that would artifact at startup if it was cold, and i gamed a lot on it. And as i mentioned, the GPU on my old board wasn't dead when i replaced it!
The guy who sold me the board told me that it is recommended to mod the fan so it runs maxed all the time, but that is something i didn't want to put up with and i believe he was BSing me because the fan is temperature controlled after all. Besides, the fan acts like a dust sensor for me - if i hear it running too often it means i need to put the vacuum cleaner in there and pull out the dust bunnies. If it'd be running maxed, besides being annoying, i'd forget it needs cleaning, it would clog up and fail.
So, not only was i ripped off because he didn't bother to test the new board to see how much video RAM it has, but a "profesionally reballed" board lasted 4 months. That just sucks. I paid $80 plus my old board for it... And i knew the guy from before, i didn't imagine he was such a crook. Note that i'm not dissing reballing as a procedure by any means, i'm just pointing out that like all other jobs, it can go wrong if the user is inexperienced or the materials are poor. I think it's a case of the former here.
Let's see if you can guess how much my oven reflow will last... Bets are on. I think i posted the oven procedure somewhere, if anyone needs it i'll detail again.
Turns out the 8400M i got with this board was only 128MB instead of 256MB i had on mine. Not wanting to return this board as well i said screw it, i don't game anymore so this oughta do. And i hadn't opened a single GPU-intensive game on it...
Until Monday evening. I had a friend over and he played some Red Alert 3 while i was playing some oldschool Carmageddon on the old laptop. RA3 worked fine for about half an hour, and then it crashed with a screen full of artifacts. Restarted, again worked fine for a while, then crashed again. It was getting late so we shut it down and went to sleep. On Tuesday evening, it started crashing at the desktop then died entirely. I called the guy who sold me the board and asked about the 6-month warranty... he started talking back at me. I was like "fine, will you get this damn thing off my hands then?" He said he isn't interested in buying it, he's already got 2 for testing, and he isn't selling those (since everyone already knows they're junk).
Fine. Screw you... Back to our old friend the oven. I love the smell of freshly baked nVidia in the morning...
Took the board out, turns i went a little bit overboard with the heat as the connector for the internal modem fell off. I just soldered it back on. I also put some cardboard in the case where the GPU sits so the heatsink is held on tight and presses on the chip, turns out there wasn't anything to support it, and the GPU heatpipe on the DV9000 is only held by two screws to the GPU. Looks like whoever designed this failed their physics exam.Put the thing back together, working like a dream. Passed 20 minutes of FurMark with the overclocked settings that i was using with my old board. Maximum temperature the GPU reached was 76C, and then it went down to 70 as the fan kicked in harder. Unfortunately the fan is bound more to the CPU than the GPU temp, but i doubt this is the cause of failure - my old board lasted a good 2.5 years with a failing GPU that would artifact at startup if it was cold, and i gamed a lot on it. And as i mentioned, the GPU on my old board wasn't dead when i replaced it!
The guy who sold me the board told me that it is recommended to mod the fan so it runs maxed all the time, but that is something i didn't want to put up with and i believe he was BSing me because the fan is temperature controlled after all. Besides, the fan acts like a dust sensor for me - if i hear it running too often it means i need to put the vacuum cleaner in there and pull out the dust bunnies. If it'd be running maxed, besides being annoying, i'd forget it needs cleaning, it would clog up and fail.
So, not only was i ripped off because he didn't bother to test the new board to see how much video RAM it has, but a "profesionally reballed" board lasted 4 months. That just sucks. I paid $80 plus my old board for it... And i knew the guy from before, i didn't imagine he was such a crook. Note that i'm not dissing reballing as a procedure by any means, i'm just pointing out that like all other jobs, it can go wrong if the user is inexperienced or the materials are poor. I think it's a case of the former here.
Let's see if you can guess how much my oven reflow will last... Bets are on. I think i posted the oven procedure somewhere, if anyone needs it i'll detail again.
I guess they finally figured out that people want to actually use their laptops on the go for work and not just at home with dim lighting watching a movie. As soon as my budget allows i'm going to buy one.
Looks like someone figured how to make the GPU temperature matter more in fan speed settings. 
<----Computer says I need more beer.
Comment