The big number and the five averages

← The counters and the dials · ↑ Reading the window · The status line →

The heart of the instrument: one measurement, taken over five windows that each get ten times longer, and how well each one knows its number.

The big number

The counts per minute over the last 30 seconds, averaged across the counters. Three seconds is too jumpy to read as a headline and five minutes too slow to react to anything you are doing with your hands. Beside it:

CPM ±6 (17.4%)
0.211 uSv/h

±6 is one standard deviation, in CPM, and 17.4% is the same thing as a fraction of the reading. uSv/h is the reading divided by the tube factor (151.5 for the M4011 in a GMC-320; --cpm-per-usvh sets another).

Under it, with more than one counter, each counter's own 30-second figure in its own colour: A 18 B 48. The big number is the mean of these, not the sum. It is the number one counter would report, measured from twice the arrivals.

The five averages

    3s    34.3   0.223   50.0%
   30s    32.5   0.211   17.4%
  300s    27.1   0.176    6.1%
 3000s    filling  2044s
30000s    filling 29044s

Window	What it is for
3 s	To see a source come and go under your hand.
30 s	To read the room. The big number.
300 s	Five minutes: a figure worth writing down.
3000 s	Fifty minutes: what the background here actually is.
30000 s	Eight hours and twenty minutes: a working day.

Each row has three numbers.

The first is CPM, the mean rate over that window, coloured by the band it falls in. The second is the same rate in µSv/h, by the tube factor.

The third is how well the window knows its number. Radioactive decay is Poisson: the whole of the uncertainty in a rate is the number of counts behind it. N counts give a relative error of 1/√N. So the figure is

precision  =  100% / √(counts the window holds)

and the counts a window holds is roughly CPM × seconds × counters ÷ 60. In the block above, the 30-second row rests on about 32 counts, so it is known to 17%; the 300-second row rests on about 270, so to 6%. The 3-second row had four counts in it and is known to 50%. Every step down the block is ten times the counts and three times the precision. Without this column the only visible difference between the 3-second figure and the working-day figure would be that one of them jumps about.

This is also what a second counter buys. Two tubes put twice the counts behind the same mean, so every window is known a factor of √2 better. The column is the only place that benefit is visible.

A window that is still filling

It shows filling and the seconds it still needs, and no number at all, rather than a number made from part of a window. The 3000-second row fills fifty minutes after the service starts; the 30000-second row fills eight hours and twenty minutes after. Restart the service and they start again from empty. A window opened later inherits the service's history, so it shows the same rows as one that has been open all along.

← The counters and the dials · ↑ Reading the window · The status line →