There is a very common problem that only applies to these GM generators. Due to their design, they develop very high voltage spikes. Those spikes often destroy the built-in voltage regulator or one of the diodes. It is very rare to find an intermittent diode but that is possible for the voltage regulator. You also can have nothing more than worn brushes but they typically last longer than the other parts that fail.
Diodes usually short, then burn open. When you lose one of the six of them, you lose exactly two thirds of the generator's output capacity. That means you'll get about 30 amps from the common 90 amp unit under a load test. 30 - 35 amps is enough to run the car under most conditions but not all. The electric fuel pump draws 8 - 10 amps, the heater fan motor can draw another 10, then you have daytime head lights with another 8 - 10, and there's nothing left over to recharge the battery. This can go undiagnosed for a long time if the battery doesn't run down while driving.
Also, all AC generators develop three-phase output which means one of them is always producing near the maximum voltage. As one phase is just nearing its highest voltage another one has just left its highest voltage, and in between the output voltage drops just a little. That variation is called "ripple". Most professional load testers display ripple, not as a number, but as a bar graph. All you care is that ripple is low. When one diode fails, output voltage drops real low when its time for that phase to produce its output. That will cause ripple to be very high. The voltage regulator may respond to that momentary drop in voltage by trying to increase generator output. That can actually cause system voltage to go up a little, but the proof is the maximum output current that can be obtained is way too low. I know I didn't explain that real well. If you care to know more about ripple, this page might help:
Due to the design of these generators, they produce a lot of voltage spikes from switching the field current on and off hundreds of times per second. That always happens when current flow is forced to stop suddenly through a coil of wire. Developing those spikes is the goal in an ignition coil but not in a generator. Chrysler, Ford, and the '86 and older GM systems work the same way but they don't have this problem. I don't know why but it only affects GM cars with this "new" design that started with the '87 model year. It has been very common to go through four to six generators in the life of the vehicle but to prevent those repeat failures, the fix is to replace the perfectly good battery at the same time. Part of the battery's job is to dampen and absorb those voltage spikes. As they age, they lose their ability to do that. Your old battery will still crank the engine just fine; in fact it will work in an '86 or older car. It can still deliver the needed current but it can't deliver it as long as a new battery will. That's the secondary characteristic. The main problem is that inability to absorb the voltage spikes.
So, . . . if you do find the generator is causing the problem, replace the battery too unless it's less than about two years old.
The other issue with these voltage spikes is they cause current spikes in the wire going back to the battery. When current flows through a wire, it sets up a magnetic field around it and that will "induce" a similar voltage spike in other wires running alongside it. It's when one of those wires is for a sensor that problems occur. To some computers, a change in voltage from a sensor of a few hundredths of a volt is significant, and the computer will react in some way to that. The way a lot of people find that is to unplug the small connector on the side of the generator while the running problem is occurring. That will turn the generator off and stop the production of those voltage spikes. The proof is the engine will run better, but only if the battery is fully charged.
Keep in mind I'm only suspecting the generator because it is such a common problem and the symptoms you described could be the result. The proof is in the load test results while the problem is occurring. The test won't identify the problem if it's intermittent and not acting up during that test.
As far as plugging something into the lighter socket, I doubt that caused a problem, especially an intermittent one. There are some plug designs that will short out inside the socket if they are oriented just right but that blows a fuse. That would cause a permanent problem, not an intermittent one. Also, all these computers on today's cars DO fail too often but they do that on their own. You have to really try to cause damage on purpose. If simply plugging something into the lighter, (which is what it was designed for), really did cause a problem, we would have heard of it many times before now.
Thursday, March 29th, 2012 AT 7:29 AM