Sounds like you've had a lot of success but I think you're beginning to run into some of the limitations of using this method.
the Radeon 9600 (128MB) chip just went dead, the notebook starts, no beep, no screen, just the DVD spins up and reads something. Fan is running. The Thinklight does not work anymore.
I am dead sure I did not overheat the chip with the heat gun. I tried reflowing twice.
Do you have any idea what was wrong?
I'd suggest checking fuses - go to my sig, download my 'Docs and How To's List', and have a look at the T4x fuses location doc listed.
Other than that, I'd say the motherboard is damaged elsewhere.
On the Radeon 7500 I had an extremely strange experience. When I heated it up to just barely 230C, suddenly little balls of tin appear next to the graphics ram chips, out of nowhere, squeezed out from a virtual hole in the sealing of the graphics ram on top of the chip. I have never seen this before! The chips don't work anymore. Has anyone seen this as well?
This is due to too much heat being applied to the top of the chip and has damaged one of the GPU memory chip BGA's (ball grid arrays - the individual tiny solder joints on the chip), and is not uncommon if too much top heat is applied like this. What's needed is to remove that memory chip and reball it, but AFAIK no one does it, so it's a new graphics chip.
One of the reasons why infrared reflowing employs bottom heat as well as top heat is to precisely avoid too much heat from one source, and instead applies a certain percentage from below and so much from above.
The resultant combination of heat sources then just melts the BGA enough to allow the solder to just turn into a liquid state without it starting to run all over the place as has happened to you.
Regarding attempting to measure the temperature whilst applying heat, it is not actually possible to meaningfully measure the heat being applied to the chip whilst applying top heat from any source since too much intense heat being applied to one area which will heat up far more quickly than the thermocouple probe can track. If relying on this method inevitably the chip will be overheated.
I appreciate that it's not practical from a price point of view, but when reflowing subminiature high magnification devices are employed which enables the user to actually observe the BGA in real time and see when the solder melts.
This is the one of the most reliable methods and is far and away better than trying to guess at the surface temperature.
Hope that helps - maybe in future think about applying heat a lot more gently - or if you're in the States send the boards off to Superior Reballing
or Arndt Computer Services
- both owners are active members here too.