Determining the operating voltage of each counter is called "plateauing the counter." We want the counter to see as many cosmic ray signals as possible without recording extraneous noise. To determine the plateau the signal counts per minute is graphed verses the voltage. To more exactly find the optimum voltage performance studies were used. These studies were run using Quarknet's Cosmic Ray e-lab. These studies show the number of pulses or signals for each length of a pulse. At the optimum voltage the graph should ideally have a Gaussian distribution with the highest point near 10 nanoseconds. A voltage that is too high will have too many short pulses, there is too much noise. A too low voltage will not give enough pulses, we aren't recording enough of the cosmic ray pulses.
There are two ways to plateau a counter: using a single counter, or using multiple counters. Different counters were plateaued using the two methods.
With a single counter the singles frequency, or signal counts per minute, is determined at various voltages and then simply graphed. The voltage varied from the lowest possible while still receiving signals to the highest possible voltage in increments of about 0.1 volts. To determine the counts per minute at a particular voltage the counts over five minutes was determined then divided by the time to give counts/minute. The scalers display the counts in each of the counters and the command DS displays the scalers. RB resets the scalers to zero and DG displays the time from the GPS. The commands RB and then DG were used to reset the scalers and give a starting time. After about five minutes the commands DS and DG where used to display the scaler counts and the ending time. This was done at each voltage. The starting and ending times were subtracted to give the time passed and then the scaler counts were divided by that time to give the counts/minute. Next the counts/minute were graphed verses the voltage, with the counts scale logarithmic. The plateau is the most level part of the graph. It is usually towards the middle of the graph, the ends are not the plateau even if they are the most level.
Plateauing using multiple counters was similar to using a single counter. The difference is instead of measuring the singles frequency we measure the coincidence frequency. The coincidence frequency is the counts per minute for signals that are detected coincidently in two or three detectors. The counters were set up in a stack and for the sake of description we will label the top counter number one, then two and three with number four on the bottom. When plateauing counter one its voltage was changed just like when plateauing singly. The other two counters', say counters two and three, voltage was set based on the single counter plateau plots and kept constant through out the experiment. The counts were then recorded for the coincidence of one, two, and three at each voltage, as opposed to the counts for just counter one. The coincidence counts for counters two and three also had to be measured, but only a few times because their voltage does not change and so the counts should be about constant. The coincidence between counters one, two and three was then divided by the coincidence of two and three and was plotted vs. the voltage. This time the frequency scale was not logrithmic. The scale does not cover several orders of magnitude and so does not need to be logrithmic. Again, the place on the graph that is the most horizontal is the optimum voltage.