In reflow we trust

I was lucky enough to be given a dead Macbook Pro at the weekend; its owner had long since given up on it. The thing booted fine, made all the right noises but the screen was dead. The Geniuses at Apple centres were quick to try and upsell a new display assembly of course, but they have bills to pay and product to shift. I have steely determination and a miserly outlook.

As usual for these types of thing I had the machine in bits looking for any obvious signs like disconnected cables or fried components. Predictably, from the parts I could actually reach that weren’t hidden behind obscure tri-wing or torx screw heads, it all checked out.

I verified that it wasn’t just the backlight and tried an external display: nothing. That was kind of a relief because it meant the screen assembly was probably fine and would have been an expensive mistake to listen to the so-called Genius bar employees. Thus it was likely the graphics controller on the motherboard (sorry, logic board in Applespeak).

So I went into deep research mode, skipping past all the usual forum riff-raff, naysayers and fanboys who have more money than sense — you know, the clueless ones who recommend buying all new hardware and software just because an app stops working or Apple sneeze a new product.

Anyway, after Googling for the model number (A1229) and the symptoms, I stumbled upon a thread buried in Google’s results whereby someone claimed it was a known fault with this particular model. Back in the day the machines had been recalled or reworked by Apple when they discovered the manufacturing fault. As luck would have it, I had direct experience of such problems when I worked in manufacturing: dry joints.

Tell me dries, tell me sweet little dries

The trouble with chips such as BGAs or LGAs (the Land equivalent that require paste on the pads) is that they need to be perfectly seated when they are flowed or else they are prone to dries. With normal components you can spot dry joints with a decent light source and they can be fixed with a careful bit of fine-tipped soldering. But the BGA’s pins are all directly beneath the chip casing so it’s impossible to see without an x-ray machine and, further, nigh-on impossible to conventionally reflow the joints at home.

This problem affects loads of devices, perhaps as a result of the drive towards cheaper, outsourced manufacturing. The X-box famously had this problem and some people resorted to putting them in the oven wrapped in a towel to reflow the joints instead of having to buy new machines. I wasn’t quite ready to incinerate the Macbook, but wanted a solution.

All you need to reflow chips is heat. Lots of it, but not too much. This guy on the forum suggested — based on some info from ifixit.com apparently — that since Macbooks get so damned hot, with a bit of creativity one could use its own heat to reflow the joints. In a twisted way, it made sense and I’m all for experimenting, especially as there was nothing to lose in this situation and I’m reluctant to pay for anything that still has life left in it.

So I followed his advice:

1. Opened the lid and switched the machine on, with battery in and power supply connected.
2. Put a towel on the desk.
3. Turned the machine upside down so its keyboard was resting on the towel and the lid hung over the desk edge.
4. Folded up a nylon blanket as thick as I could and completely covered the machine to trap heat inside. Duvets also work well, I’m told.
5. Left it like that for four or five hours, only poking my hand under every fifteen minutes to tap a key and prevent it going to sleep.

The machine got roasting hot and the fan cried, but against every computery instinct, I ignored its whimpering. After it was cooked I turned it back over, let it cool for a few hours and switched it on. It only bloody worked! Boo yah, in your face, Genius bar.

I don’t know how long it’ll last, but for now I feel vindicated that I have a functioning Macbook Pro from whence there wasn’t one. Thank you Internet: sometimes you can be a Really Useful Engine.