This week’s repair boggled my mind. A 2.33GHz MacBook Pro came in with the simple problem of not booting up from its internal hard drive. I usually start up holding down the Apple and S keys to get into single user mode, and use fsck (file system consistency check; http://en.wikipedia.org/wiki/Fsck) to check and attempt repair to the hard drive’s file system. If it gives an i/o (input/output) error, I can generally assume that the hard drive needs repair. If it attempts repair, but fails, I move on to DiskWarrior (http://www.smalldog.com/product/41941), which repairs pretty much anything. If that fails, the file system is so far gone that it needs to be erased.
In this case, I went into single user mode and ran fsck and it stated that no problems existed. Since the machine booted past the progress bar, and stopped at the same point each time, I thought the issue was software based. I threw in a trusty Leopard disc and went to do an Archive and Install, a nondestructive reinstallation of the operating system, and once the laptop began to boot from the Leopard disk, the screen went dark.
I plugged it in, thinking the battery went dead, and tried again – but the computer never turned off. I forced a shut down by holding down the power button and then fired it back up, holding down the trackpad button to force the disk to eject before the startup process, then forced a shut down again. I inserted the machine’s original restore disc and tried to boot from that. Same thing: dark screen.
I then plugged the machine into our NetBoot system to see if any external boot volume would keep the screen on. Same thing happened.
Totally baffled, I put the laptop into target disk mode and hooked it up to my own MacBook Pro to capture a backup of the machine. I then reinstalled the operating system from scratch using target disk mode. When it finished, I tried to start it up, and… you guessed it, the screen stayed dark.
At a total loss, I opened the thing up and started removing components, stripping it to a totally minimal configuration of logic board, inverter, display, and a few other components. Gradually, I started plugging components in between restarts, using the process of elimination to figure out which hardware component was causing the problem.
When I finally arrived at the SuperDrive, I had my answer. The machine behaved normally until I reconnected the SuperDrive, so I then took a known good SuperDrive flex connector cable to see if that little cable could cause all this ruckus. Lo and behold, a $10 part needed replacing.
I restored the customer’s data from the backup I’d created and let him know it was ready for pickup. This was truly one of the most bizarre diagnoses I’ve ever come to.