Pressure-loss characteristics of a variety of single- and double-substrate metal-foil and ceramic-substrate converters with tapered and truncated inlet and outlet headers were measured in room-air flow, hot-gas flow, and engine-exhaust tests. Test data in the three different media correlated with the inlet-pipe Reynolds number when expressed as a loss coefficient, i.e., pressure loss normalized by the inlet-pipe dynamic head. Because restriction measurements made in different media correlate well as a Reynolds number-dependent loss coefficient, inexpensive room-air test data can be used to estimate converter pressure losses in the engine environment. The normalized losses in the substrate varied inversely with inlet-pipe Reynolds number, ranging from, e.g., 6 at Re = 30 000 to 2 at Re = 200 000. The remainder of the losses occurred in the inlet and outlet headers and in the section between the substrates. Normalized losses in the headers were essentially independent of Reynolds number and constituted a third of the total restriction at moderate flows and half of the total restriction at very high flows. A normalized loss of 0.2 was associated with flow adjustment in the center section between substrates in double-substrate converters. The pressure loss introduced by truncating both the inlet and the outlet headers was small, only 4-6% of the total restriction. Offsetting the outlet pipe by 20 mm increased the restriction by 2-6%.