One of the problems with ionization chambers is that the detector is indiscriminate and will detect any ionizing radiation If, for example, neutrons are to be detected m the presence of a strong gamma field and the neutron flux is to be related to the average current, then it is necessary to take into account the component of the current that is due to the gamma field

Part of the gamma-induced current is due to prompt gammas and is proportional to the neutron induced current This part is indicative of reactor power and is not detrimental However, the remainder of the gamma-induced current is relatively unchanging and creates a spurious signal, і e, a signal not indicative of the reactor power level

This, in turn, is not a problem at high power levels when the neutron field is much more intense than the back­ground gammas At low power levels, the gamma contribu­tion to the chamber current may be a large fraction of the chamber current and might exceed the neutron-induced current Thus the range of reactor power in which an ionization chamber can be used to measure the power level may be severely reduced

Two ionization chambers can be used to decrease the effect of the gamma background If an ionization chamber sensitive to gamma radiation only is installed near an ionization chamber that is primarily sensitive to neutrons, the signal from the gamma chamber may be used to cancel the gamma contribution to the neutron-chamber signal In practice, the chambers must be carefully matched and their relative positions must be properly fixed Chamber pairs for this purpose are commercially available

A common way of neutralizing or compensating for the effect of gamma radiation is to combine the two ionization chambers into a single unit called a compensated ionization chamber,1 4 1 s frequently abbreviated CIC A typical CIC is shown in Fig 2 8

A compensated ion chamber is essentially two ioniza­tion chambers m a single case One chamber collects the

total current due to both neutrons and gammas. The other chamber is identical to the first in sensitive volume but lacks the neutron-sensitive materials. If electrons are col­lected in one chamber and positive ions in the other and if the resulting currents are summed, the gamma-induced currents are cancelled. In practice, it is normally immaterial whether the signal is obtained by collecting electrons or positive ions.

Theoretically, the cancellation could be complete. In practice, cancellation can be made complete at a given reactor power level. However, over the range of reactor power levels, most of the gamma effects can be nulled but a residual should be expected. The residual may be less than 1% of the signal or as great as 10% and depends on the
specific detector and the effort made to achieve good compensation. A reasonable state-of-the-art number is 2 to 3%, i. e., 97 to 98% compensation.

It is also possible to overcompensate. Figure 2.9 illus­trates the manner in which compensation varies as a function of operating voltage. Because of this voltage — sensitive feature, it is possible to use variable com­pensation.16 The variable-compensation feature has been commercially used.

Figure 2.10 shows one way in which compensation might vary over a portion of the reactor range. With fixed voltages, compensation is exact at only one point. Figure 2.11 is an illustration of the improvement gained from using a compensated ionization chamber. The figure

also provides insight for measuring and testing com­pensation in an operating reactor

The use of a gamma-compensated detector extends the range, compared to that of an uncompensated detector, by about two decades Depending on the reactor, an uncom­pensated detector may be expected to cover three or more decades, a properly located compensated chamber can reasonably be expected to measure the reactor power over at least five decades Recent practice has generally been to operate with fixed voltage and to design safety systems that avoid dependence on as r ach as two decades of compensation