Reflowing T4x video chipsets

[small_thing_on_wax]

hi there.

1) heat the gas oven up to 250°C or or slighty more.
2) put your board on a baking dish, so you can get it out of the oven.
but put some cloths between the board and the baking dish. it prevents shocks while moving the board while its hot and assures that no "pieces" fall of the downward site of the board.
3) measure the chipsurface with a IR every now and then 30s/60s.
What i measured: when the boards temperature is about 90°C, it heatend up with ~20°C/min (but it depents highly on how long you open up the oven for measurements | and yeah i know its a bit fast). After 5 min my board was about 220°C. You can see that the chip sagged.
4) remove the board with caution and let it cool down to at least 70°C. Test it (you need the monitor, cpu, ram, cable , fan, hdd, speaker, keyboard, touchpad)

I measured the warm up of my gas oven, it is only about 10°C/min. So it is not good powerfull enough to put the board in from the beginning. I think it will be the same with a oven powered by electricity. The plastics are very insensitve to heat. Nothing happend to them, even to the black stickers. But remove everything just to be sure.

to richk's question
well, i used some "standard" sticks (those can you get in every do-it-yourself store) , i only made sure they wouldn't get elastic under 80°C, since this temps "can" occure under high load of the system. I sticked all bga chips on the board (twice on every edge and in the middle of the chip).
The board also becomes MUCH stiffer that way, and it secures the chip as the glue holds the pressure against the board. Also you can place a good heat sink on top of the chip and apply some additional pressure from the keyboard.

and don't use that heat gun method, you don't get the heat UNDER the chip...

[visionviper]

and don't use that heat gun method, you don't get the heat UNDER the chip...

I have to disagree - if you use the heat gun method properly the heat does penetrate the board.

I prefer the heat gun method because I get the ability target the BGA package with the heat instead of heating the entire board.

I am not saying the oven method is bad - I am just saying there are options and both work.

[small_thing_on_wax]

hi visionviper.

ok, I agree with you, the heat penetrates the board. But my concern is that you must stress the upper site of the chip too hard to bring the required temperaure under the chip. The space between the chip and the board is rather narrow and you can get your gun only at a point every cycle and don't cover the whole chip. Secondly, you must have a heat gun that hits the desired temperature in a good approximation, otherwise you must move the gun a lot what causes to much fluctuation in the temperature.
I didn't intend to discredit your guide, its well done. But my experience is that you must do the reflow via heat gun very accurate to get a lasting archievement. (I tried it several times with no lasting archievement). In my opinion the oven method is easier to use and more efficient. The whole board is able do overstand the process, and its better to do it once the hard way than try it multiple times with the heat gun.
But hey in the end what works - works, right?

[visionviper]

But hey in the end what works - works, right?

Yup! That's the main thing - help people on these forums find a way to be able to fix their board themselves.

[richk]

One thing I would add is that commercial infrared reflow machines heat the chips from above

[small_thing_on_wax]

...and after 3 months of hard use one of five is gone down the drain.
Well this one was one of the first and the temp was not as high as it should have been. Think i`ll give him another chance after toastin that lil fellow again. And by the way: What i read so far is that is not recommended to heat the balls any longer than 10min, cause than you'll get problems with oxydation on the surface. Oxidation is also limiting your tries. It will get harder to archieve a lasting reflow each try.
To optimize the overall process, you can heat the board in the oven and apply the last peak of heat with the heat gun.

[sjthinkpader]

One thing I would add is that commercial infrared reflow machines heat the chips from above

That was true when the IR oven first appeared in the early '80s. Today they are forced air. When we first use this type of ovens, the first few boards that went in got charred. We then use dummies as a lead in to stablize the oven temperature.

In those days, vapor phase reflow was developed at about the same time. These were abandoned in the mid-nineties due to the elimination of fluorocarbons (freon).

Reflow Profile

The shoulder area is called preheat. The actual reflow is very short. So WAX's oven bake first is like a preheat. Some plastic packages may absorb enough moisture to pop like a popcorn when heated. They require a pre-bake at lower temperature to dry out.

[small_thing_on_wax]

hello sjthinkpader.
So you think this method is too harsh and i should take care about moisture? Would it be better to preheat ("A minimal soaking zone of 20 to 30s at 170 to 175°c reduces voiding, poor wetting and opens") the board? I had no probs concerning this so far. But if you think it would be necessary, let me know. Thx!

[sjthinkpader]

Plastic components are shipped in sealed bag with desiccant inside to prevent this popcorning effect. But they are looking for a 99%+ first pass yield.

I think as a repair process, you don't have to worry about it.

[qwadro]

Put T41 MB in the oven, 450 F temp. 10 minutes. The RAM module fall down, as well as all other part soldered on other side of the board. ))

Soldered everything back, but the RAM module. Doesn’t start up.

I was wondering can i use an iron. Put it on a Cotton mode and heat GPU? Anyway a new board on the way and I’m just playing with this board around.

[small_thing_on_wax]

U need to to remove everything from the board, incl. Ram.
Even the black stickers, as some smd can get loose to board, while sticking on the sticker. smds won't fall of if u put a cotton cloth under the board to fixate them.

Also, I don't think that idea of yours will work.


(since then all IBms i've toasted are stable (except the one i had to lil fellow again, but that fellow behaves now too)

[lophiomys]

Hi ,

Thanks to small_thing_on_wax.
I just did a reflow of my 2373-7XU following the
convection-oven and it did actually work to the point
that all PC-Doctor tests have passed.

\Lophiomys