Hmm 385 Fahrenheit thats about 200 Celsius I can see why you would remove plastic parts like the fans and so on. But im guessing it wont hurt to leave pure metal blocks on? think I might try baking several sets of OCZ reaper ram and a set of Corsair DHX dominators and maybe a 7800gtx.
maybe but I think it might be just as dangerous to remove the heatsinks on the ram sticks as the baking itself. I have seen several pictures of people having ripped the ram of the plate, because of the sticky tape stuff thats in between.
Heatsinks also works in reverse. Normally the air in your PC isnt 200 Celsius. It will focus the heat onto the chips very efficiently.
It is - this just goes to show you how bad surface mounting chips is that run at high temperatures. You will get flexing no matter WHAT you do. PS3s are failing due to this as well. Lead free solder is just a pain in the ass to deal with, due to it's lower specific heat. /me picks of the roll of solder on his desk....not lead free. Om nom nom nom! A word of warning to the wise: This will fix issues where the chip has bad solder points; that is a very specific issues, and most failures are not due that. Heating a chip won't revive it if there is actual damage to the silicon package. The oven isn't the cure all for electronics. A videocard generally runs hot enough to cause flexing, and so can hit this problem. RAM however doesn't run that hot, and likely won't be fixed by this. Remember kids, you can damage electronics that weren't designed to survive high temps.
but there are ram on the cards... and they dont die during baking..... Was thinking maybe they sucked from the manufacturing process since they all died after a while having worked fine. Except the dominators they were dead on arrival.
its al the green parties fault, they made lead in solding illegal but then again i wouldn't want to die from lead poisoning ether just because i welded stuff in my life
With regards to memory on card, GDDR5 is VERY different from DDR2 or DDR3 - heat tolerance is certainly one component. You're not comparing apples to apples. With system RAM that doesn't work, it's usually just poor quality control. Moreover, even if it IS a bad solder joint, you are relying on the heating to make it move in the right direction. I mean, think about it: It could move back into place, or the solder could bridge causing a short. It's hit or miss at best.
It's not as hit or miss as you might think. When you melt the solder, it doesn't just run everywhere like water. Or rather, it does - that is to say, it has surface tension like water does, so it tends to stay on the terminals it's already on. Applying heat generally like this is how the joints were made in the first place - the solder is applied to the board as a paste, the components are all sat in the paste, and when it is heated up to the right temperature, the paste goes liquid and separates itself out into individual connections. It's a lot more efficient than trying to apply solder / solder paste to each connection individually. Reheating is more likely to remake a joint that is there already than it is to make a bridge.
My old PC's northbridge heatsink literally fell off when we opened the casing, the heat must've molten one of the pins holding it in place, and the paste that sticks it to the chip was all dried up, and not very sticky. it still works, but there are lines all over the screen, and the screen goes dark every now and then.
I was an EE/CE major until I came to my senses...I do know how this works. I worked in a plant that made some of the guidance systems for some of the US' guided missiles - so I know what a pick and place machine is and such. However, the paste doesn't just separate itself out, you usually only apply the solder paste to the exposed copper pads...you don't just coat the board in solder. You have a jig/stencil through which you put the solder. That doesn't change the fact that heating up solder on a chip that has flexed over time, that is to say it moved and no longer makes good contact with the PCB (given that some pieces are surface mounted, and have no thru pins.) Sure, solder adheres to copper...but again, heat it up on something under pressure (as GPUs generally have some retention bracket around it) or worse something with no pressure on it will just shift even more. All you're doing it heating it up and hoping it'll remelt the solder, then allow the surface mounted package to move into place. It just won't always work...moreover, it doesn't fix the issue: the same problem WILL reoccur again....all you did was move the chip back into place, you did not prevent it from flexing each time the chip heats up and cools down again.