The final piece to the nVidia puzzle - voltage
Well, today was a tough day.
I thought i'd finally get around to fixing my dv9800, since the new GPU had been sitting here for a while.
Just yesterday (or, well, 2 days ago since it's almost 6am...) i had done a chip replacement in a Benq R56. I always reball even new chips with leaded solder, but on that one i got lazy and put it in the way it came from the factory, adjusting temperatures to match. No sweat, that one worked great and it's already back to the owner.
Today (well, yesterday), i decided to do the same with my dv9800 and skip the reballing of the new chip. Only the $45 bastard decided to blow itself up at 230C... !#*%.
I'll never solder a lead-free BGA again. I've had ZERO chips die in the last 2 years, and now i fried a BRAND NEW one. Plus, rendered my main laptop inoperable.
I went outside with someone to clear my mind for a bit. When i came back home, i decided to go back to fixing the dv9000. I had another intel/8600gs board, which was in a pretty sorry state (someone went over it with a wind-proof lighter IIRC) but i'd lifted the shorted GPU a couple months ago and the board powered up fine. And i still had the old, semi-crapped GPU that i pulled from my board.
I did my trickery on the GPU (covering everything but the die in lots of aluminum foil and bringing the die to some silly temperatures), and it came out alive. Of course, the bad old chip just doesn't want to die. Then i reballed it and soldered it to this other board. It came to life.
I've been long wanting to try undervolting a nvidia GPU from the bad lot to see the difference it makes. I'd never gotten around to actually trying it though. So i did the mod on this dv9000 board. First by using a pencil, then once i've settled on a value, by soldering in the appropriate resistors. To finish it off, i removed the original aluminum foil/thermal material for the GPU, and applied Arctic MX-4. The CPU got the same treatment.
I was able to undervolt the GPU from 1.2v/1.15v (high/low) to 0.95v/0.89v. That is 250mV, which represents a 21% reduction in operating voltage. All with the GPU not only still operating correctly, but still presenting some overclocking headroom!!! Under FurMark the GPU Vcore measured 0.943v.
Since power increases (or decreases) squarely with voltage, resistance being constant, this 21% reduction in operating voltage results in a 37% reduction in power consumption (and thus, heat dissipation) of the chip. The chip now uses only 63 percent of the power it normally used, and thus it's only dissipating 63% as much heat.
So, to sum it up: Nvidia underrated the current capability of the bumps inside most of their chips manufactured 2006 to 2009, they used the wrong materials for bumps and underfill, and to top it off, the parts were overvolted by 20%, thus using almost 40% more power and putting away 40% more heat than they could have. That's one hell of a booboo if you ask me.
With the results i have gotten now, i would say that if the operating voltage was more carefully selected for these parts (all other things staying the same), 95%+ of them would have made it the useful life of the laptops.
You may ask, how are the actual numbers looking? I'll just say this: My dv9000 is now cooler under FULL GPU LOAD than it used to be when completely IDLE. You'll have to do the guessing. I'll post pictures and screenshots after i get a good rest. I guarantee you'll be shocked. I still am, too.
I thought i'd finally get around to fixing my dv9800, since the new GPU had been sitting here for a while.
Just yesterday (or, well, 2 days ago since it's almost 6am...) i had done a chip replacement in a Benq R56. I always reball even new chips with leaded solder, but on that one i got lazy and put it in the way it came from the factory, adjusting temperatures to match. No sweat, that one worked great and it's already back to the owner.
Today (well, yesterday), i decided to do the same with my dv9800 and skip the reballing of the new chip. Only the $45 bastard decided to blow itself up at 230C... !#*%.
I'll never solder a lead-free BGA again. I've had ZERO chips die in the last 2 years, and now i fried a BRAND NEW one. Plus, rendered my main laptop inoperable.I went outside with someone to clear my mind for a bit. When i came back home, i decided to go back to fixing the dv9000. I had another intel/8600gs board, which was in a pretty sorry state (someone went over it with a wind-proof lighter IIRC) but i'd lifted the shorted GPU a couple months ago and the board powered up fine. And i still had the old, semi-crapped GPU that i pulled from my board.
I did my trickery on the GPU (covering everything but the die in lots of aluminum foil and bringing the die to some silly temperatures), and it came out alive. Of course, the bad old chip just doesn't want to die. Then i reballed it and soldered it to this other board. It came to life.
I've been long wanting to try undervolting a nvidia GPU from the bad lot to see the difference it makes. I'd never gotten around to actually trying it though. So i did the mod on this dv9000 board. First by using a pencil, then once i've settled on a value, by soldering in the appropriate resistors. To finish it off, i removed the original aluminum foil/thermal material for the GPU, and applied Arctic MX-4. The CPU got the same treatment.
I was able to undervolt the GPU from 1.2v/1.15v (high/low) to 0.95v/0.89v. That is 250mV, which represents a 21% reduction in operating voltage. All with the GPU not only still operating correctly, but still presenting some overclocking headroom!!! Under FurMark the GPU Vcore measured 0.943v.
Since power increases (or decreases) squarely with voltage, resistance being constant, this 21% reduction in operating voltage results in a 37% reduction in power consumption (and thus, heat dissipation) of the chip. The chip now uses only 63 percent of the power it normally used, and thus it's only dissipating 63% as much heat.
So, to sum it up: Nvidia underrated the current capability of the bumps inside most of their chips manufactured 2006 to 2009, they used the wrong materials for bumps and underfill, and to top it off, the parts were overvolted by 20%, thus using almost 40% more power and putting away 40% more heat than they could have. That's one hell of a booboo if you ask me.
With the results i have gotten now, i would say that if the operating voltage was more carefully selected for these parts (all other things staying the same), 95%+ of them would have made it the useful life of the laptops.
You may ask, how are the actual numbers looking? I'll just say this: My dv9000 is now cooler under FULL GPU LOAD than it used to be when completely IDLE. You'll have to do the guessing. I'll post pictures and screenshots after i get a good rest. I guarantee you'll be shocked. I still am, too.
Interestingly enough, i had to lower the PCI clock a bit, otherwise the GPU started acting up. This board does not seems to have a PCI/PCI-E clock lock, it changes with the FSB when the fine adjustments are used. It goes up to 240-245 without crashing but it isn't stable.
I was just pointing out that there is yet another side to the big nVidia issue that very few people have investigated.
Comment