January 04, 2001, 12:02 PM — I've talked about software to a great extent lately, so I'm going to leaven the mix up by talking about hardware. Hardware problems, to be exact. The kind that drive support people nuts.
It's safe to say that hardware problems are a pain to diagnose and treat compared to software problems. If one comes to the understanding that they can be due to so many different things, it's really hard to throw out any instructive example and have it stick. What worked in one case may not carry over to the next. But having said that, I'm going to make the attempt to show how hardware problems can manifest themselves.
Hardware problems tend to first show up in equipment that experience a varied working environment. Changes in environment can give rise to physical forces (like expansion and contraction) that have definite physical effects. The classic example of this is the user who brings in a laptop with a totally black LCD screen along with the explanation that he only left it open in the car when he went out for lunch and yes, the sun was hot and windows closed in the car and when he tried to use it after lunch the screen was all black. One must be very calm in explaining to the user that heat can mess up the LCD screen and that it's not a real good idea to leave the screen in places where the sun can heat it up. Screaming is not to be encouraged , even in the face of extreme obstinacy. [OK, maybe after the fourth time the laptop is brought in; but not until then.]
Hardware may also show intermittent symptoms. This explains why it may look fine on the workbench ; but the user says they have a problem. Here's one case involving a G3 Powerbook.: the display goes screwy. Icons on the desktop are surrounded by low-resolution artifacts; system fonts display either jaggy or as totally different
font; colors are somewhat off (but not severely). All of this is intermittent -- sometimes it comes and goes within minutes; sometimes hours.
In the course of backing up user's data to swap machines, it was noticed that the lower PC card bay acts strange -- first, it won't eject properly; then after encouraging the card to emerge, it won't accept any more cards. This behavior gives an hint ( a big old flare, actually) about where to check. Pop the keyboard, remove the heat shield, and look in the bay.
There's a spring-loaded ejector piece of metal in the bay, which apparently can become bent -- and did so. In this case, it had flexed downw ard at it's center, *just* enough to contact the motherboard below (no shielding between them), and there's the video circuitry. Getting shorted out. Diagnosis is made; and Apple gets one machine returned.