In the hope that there la someldiing to be learned from niatakes, I shall out-;• line briefly ttie sitiiationa that hare led to aocidaital pron^jt-critical radiation bitrsta in critical asseniblies laboratories, and nentian a few others that ippeared to be rislQr. Enough has been said of two of the three wartime hand-assenibly accidents to Txnderline their tragic lesson. V^th the abundance of guiding information on eritieal isystems that now exists and the use of reiK>te3y-controlled f

more »

... for approaching critical, there is no excuse for recreating the situation ^ere a sup of the hand on part of an assembly makes the system supercritical. The third early incident. Illustrates an error that did not depend xipon hand operation. • A water-tank surrounding a metal assembly in a dry well was being filled at a / " rate that exceeded the o«paoity of an overflow opening, water-level indication '. ^ was inadequate, and the metal assembly flooded, developing a radiation bxirst of * 3 X ICr^ tlasloas, Thou^ the operator peeked into the tank just in time to see a blue glow, there was enough intervening water to pxt>tect him from s«riou« injujy, Hfflre (slide 1) is a foolish situation that we set up (most accidents are foolish in retrospect). The aquarium was desired originally for obtaining the neutron multiplication of sin^e pieces of fissionable metal in water. As one •f '-•^ scram, a pneumatio cylinder raised the unit out of the water} another, though . " slow, was draiidLng of the tank. A traveling support for a second unit was added ao that critical s^aratlon distances could be determined. A dropping Cd screen ^* also was added as a scram. As a flourish after the final measiu'eraent planned for the Aquarium, the system Illustrated was scrammed, local radiation detectors went off scale and a cloud of steam showed on the monitoring televislpa^^OMpu *" JS>^J^^$Hr\l^ \ t«o"zed person is prohibited.