>
>Suppose we got a small radioactive source, as is used in smoke detectors. I
>believe that the detector works by smoke particles absorbing the escaping
>particle and the count is lower than normal, and the alarm goes off.

The particles are alphas, +2 positive charge. Basically helium without electrons.  Smoke stops them. AIR stops them pretty quick.
AM241 emitter.

They sense like a high impedance DMM.

> But the time between particle emissions is truly random. Couldn't the detector be used on the INT pin of a PIC to extract the value of a counter rapidly running through all possible numbers? and the cost should be trivial.

Detecting individual alphas isn't that trivial, but it could be done.

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.