The circuits described in the previous section operate on the average-magnitude-squared (AMS) principle and are

therefore frequency-sensitive Adjustments must be made in the band-pass filter to make the system fit a particular reactor The lower cutoff at 4 kHz, indicated above, corresponds to about 25,000 radians/sec, this is below the

‘Neither system described allows final use as a current chamber for overpower monitoring

0/7 cutoff of a fast reactor (where the roll-off in the transfer function is at about 70,000 radians/sec.)

The principle involved is to use the lower frequency components of the band to assist in pulse overlap Thus there is an incentive to use a lower frequency cutoff (The high-frequency cutoff is selected for noise rejection.) As the low-frequency cutoff is raised in the AMS system, the overlap with the pulse-counting range decreases

In a test run in 1967, the AMS and true MSV systems were compared using the instrument setup shown in Fig. 5.24. The data shown in Fig 5.25 were obtained with a band pass of about 250 to 500 kHz The EBR-2 was put on a 59-sec period, and, because of its wide flux range before reaching power feedback conditions, this is constant over four decades or more. As shown by the data, the MSV system had a correct signal for nearly two decades before the AMS signal became correct. Had a lower band pass been used, this could have been corrected, but the low pass of
the band selected is not important to the true MSV system and bands above 250 kHz are used for other reasons

If on Fig 5.21 the blocks representing the band-pass filter and rectifier were replaced with a Hewlitt—Packard true RMS convertor, the signal would be as shown in Fig 5 25 At present, the use of this converter adds materially to the number of components in the system, and commercial suppliers are hesitant to supply such systems.