1. Using the Texas photometer as a basis, choose two apertures
(one smaller, one larger) and arrange to illuminate them both
with a light source that shows stability of counting rate for
several minutes. I found an LED that had been on for long
enough to come to temperature equilibrium works fine, if it
is powered by a regulated power supply.
2. Take two readings of the counting rate, one through each
aperture, the higher chosen so
that the dead time should be significant at that counting
rate -- bewteen 500,000 cps and 2,000,000 cps worked for me.
3. Now decrease the illumination intensity, so the counting rate
drops about a factor of 10 or more. After the LED is stable
again, take two more readings, through each of the same
apertures.
4. There are four data points (the readings) and four unknowns:
a. The ratio of the aperture areas
b. The ratio of the lower to the higher intensity
c. The "true" counting rate if the detector were ideal
d. The dead time
5. I worked out the equations for this, but I don't have them
in front of me ... I can probably find them somewhere. But
the exercise is not very hard, and is useful to know about.
I recommend, as an exercise to the student, that you work out
the simultaneous equations and see if you get something you
can understand ... and believe in. They might even be the same
as the ones I worked out many years ago ...
To first order, if the dead time is negligible at the lower
counting rate, then the observed ratio calibrates the aperture
area ratio, which should be the same at the higher rate. Any
discrepancy can be blamed on the dead time. The proper way
to do it, though, is to solve the simultaneous equations for
the dead time explicitly. I got the 20 ns dead time figure
for the current system that way. Repeated measurements gave
the same dead time to within 10 percent.
ADVANTAGES: - Takes little time
- Can be done any time
DISADVANTAGES: - Less accurate
- Sensitive to measurement errors
- Requires constant artificial light source