Issue link: https://iconnect007.uberflip.com/i/1053050
52 DESIGN007 MAGAZINE I NOVEMBER 2018 They are cooled by a combination of conduc- tion through the dielectric, convection through the air, and radiation. It had previously been assumed that convection conducted heat away from the traces and cooled the traces more effi- ciently than conduction through the dielectric. Hence, the assumption that internal traces were hotter than their external trace counterparts. It turns out that dielectrics cool traces more efficiently than does convection plus radiation, and internal traces are relatively cooler [3] . One question is, "Is there a predictable rela- tionship between the external temperature of a trace and the internal temperature of the same trace carrying the same current?" That is, is the internal trace 10% cooler, 20% cooler, or is there another predictor we can use to pre- dict internal temperatures? The data in IPC- 2152 makes it possible to explore this question. Internal and external temperature curves are provided for 1.0-, 2.0-, and 3.0-oz. traces and a variety of trace widths, so these relationships can be explored directly. IPC-2152, published in 2009, was the most thorough study of trace current and tempera- ture relationships ever undertaken [1] . It con- tains over 90 pages and charts that examine the relationships from numerous directions. Before IPC-2152, the industry used a set of charts that went back to a National Bureau of Standards (NBS) Report #4283 published in 1956 [2] . The charts were first published as a part of MIL-STD-1495 in 1973. They were later published in a subsequent series of standards that culminated in MIL-STD-275E in 1984. Fur- ther, the charts were published as part of IPC standard IPC-D-275, and IPC-2221A in 2003. One of the most interesting findings in IPC- 2152 was that internal traces are cooler than external ones for the same size and current. Independent experimentation was not done on internal traces in earlier charts. Internal traces were merely assumed to be hotter than external traces, and the external trace data was derated by 50% by that assumption. Traces are heated by Joule, or I2R, heating. Brooks' Bits by Douglas G. Brooks, Ph.D., with Dr. Johannes Adam