I this note we report preliminary work on a modification of the expansion tube. This modification is suggested by our experiences with secondary diaphragm breakage and its effect on flow quality. It appears that diaphragm fragment size significantly affects flow uniformity. Our experiences imply that large diaphragm "petals" of order of the tube diameter introduce disturbances in the same order of magnitude, and the subsequent expansion process does not sufficiently smooth the flow before the gas reaches the test station. On the other hand, small fragments such as result from shattering by the incident shock introduce disturbances of a correspondingly small order. These disturbances are smoothed out further by the expansion, so that a rather smooth flow results. Accordingly, flow quality is improved by forcing the diaphragm to break into small pieces. If at the same time the diaphragm could be retained, additional improvement in flow contamination could be achieved. Preliminary tests with the modification reported here show that such improvements are indeed achieved. The modification consists of a "nozzle plate" installed at the secondary diaphragm location. The plate is shown in Fig. 1. It contains a large number of individual nozzles that are initially sealed against the driven gas by a common dia-