Glenn Ashmore has accurately described the problem - load
dump. Depending on standard, load dump on 12 volts can be 60
volts or as high as 270 volts. Two standards are SAE J1455
and ISO 7637-1. As Glenn has accurately described, load dump
can be created by disconnecting a major load from alternator -
such as battery.

Another has too much experience without underlying theory.
His proof that something does not exist is that he
disconnected a battery and failure did not occur. Reasoning
equivalent to walking with only one leg which is why his
response is:
Absolutely bogus assumption about power output remaining constant. That leg called underlying theory is essential to understanding how things really work. Load dump means automotive type electronics must be designed beyond just the 12 volt power.

SGS Thompson defines load dump as:
- Peak voltage 80 to 100 volts
- Duration 300 to 400 milliseconds
- Series resistance 0.2 to 1 ohms"

Even laptop power supplies for mobile power cost more money
because load dump protection is required.

SG Thompson makes load dump protection circuits such as
LDP24 or RBO series. But they admit:
The protection at the alternator level is a quite new concept and
all the technical problems do not seem to be completely solved.

Yes, you were lucky in not damaging the alternator if
disconnecting when alternator was outputting power.

BOEING377 wrote:
Are zap stoppers really needed on alternators? These aftermkt devices claim to
portect alt. diodes against damage from transients. I can't imagine that
engineers at places like Motorola would design alternators that needed external
aftermarket devices to effectively protect them from transients. I have never
blown up an alternator from switching it in and out of a load, although people
say this is often fatal to the diodes. I have run alternators open circuit (no
load at all) with no problems. Was I just lucky?