Traditionally, many nuclear sensors provide short pulses that have an amplitude proportional to the event you are detecting (hence, proportional counter, etc.). Others just provide a constant pulse height for any event (i.e. a GM tube, or a neutron scintillator). In some cases, the pulses coming out are rounded or misshaped due to cable capacitance, non-ideal amplifiers, or just inherently in the detector. A lot of nuclear detectors are inherently a "pulse discharge" kind of waveform (i.e. a small capacitance discharging through a small resistance... exp(-kt) trailing edge).
Also, recall that a lot of the technique of nuclear metrology was developed before the advent of inexpensive, fast, high gain amplifiers, and a LOT of time was spent compensating for the pecularities and limitations of what was available in 1930 or 1940. Modern equipment often duplicates the functions of what was used 50 years ago, more for familiarity and convenience, rather than from any inherent measurement requirement.
OK.. lets take a look at grass, discriminators, single and multichannel analyzers...
Say you've got that proportional ionization chamber detector (and amplifier) hooked up to the oscilloscope. You see the noise at the bottom (grass), and occasionally, a big spike pokes up. You want to count the spikes, but not the grass. A basic counter is going to trigger on any zerocrossing, so you would get a lot of spurious counts. Enter the "discriminator", which is a threshold circuit (called "trigger level on modern counters and scopes) that only produces a pulse out when the input goes a certain amount above the floor. Now, only the big spikes count.
In the "good old days", the counters weren't all that hot at triggering on a short pulse, so you'd add a shaping amplifier, which would turn the spike with exponential decay into a nice square pulse, with a bit slower rise time, that the counter could reliably trigger on.
Say, further, that you are only interested in spikes of a given height (not too big, nor down in the grass, nor too small). This would be typical if you are analyzing particle energies in an ionization chamber, where the height of the pulse is determined by the energy of the particle. Add another threshold circuit... now you've selected a band of levels where you'll get counts. This is a "single channel analyzer".
Now, hook up a whole raft of these in parallel, with all the levels set in a regular progression.... A "multi channel analyzer"...
Today, though... Take the output, run it into a run of the mill video a/d (say 40Mhz at 12 bit), and hook some logic on the output, and you've got yourself a 4096 channel analyzer, with all the bells and whistles you could imagine.
Created on Tuesday, April 10, 2001 12:35 PM EDT by James Lux