4.16. BJT QUIRKS273tic power transistor packages like the TO-220 and TO-247 dissipate well over 100 watts, ap-proaching the dissipation of the all metal TO-3. The dissipation ratings listed in Figure 281,4.117are the maximum ever encountered by the author for high powered devices. Most power tran-sistors are rated at half or less than the listed wattage. Consult speciﬁc device datasheets foractual ratings. The semiconductor die in the TO-220 and TO-247 plastic packages is mountedto a heat conductive metal slug which transfers heat from the back of the package to a metalheatsink, not shown. A thin coating of thermally conductive grease is applied to the metalbefore mounting the transistor to the heatsink. Since the TO-220 and TO-247 slugs, and theTO-3 case are connected to the collector, it is sometimes necessary to electrically isolate thesefrom a grounded heatsink by an interposed mica or polymer washer. The datasheet ratings forthe power packages are only valid when mounted to a heatsink. Without a heatsink, a TO-220dissipates approximately 1 watt safely in free air.Datasheet maximum power disipation ratings are difﬁcult to acheive in practice. The maxi-mum power dissipation is based on a heatsink maintaining the transistor case at no more than25oC. This is difﬁcult with an air cooled heatsink. The allowable power dissipation decreaseswith increasing temperature. This is known as derating. Many power device datasheets in-clude a dissipation versus case termperaure graph.• REVIEW:• Power dissipation: maximum allowable power dissipation on a sustained basis.• Reverse voltages: maximum allowable VCE , VCB , VEB .• Collector current: the maximum allowable collector current.• Saturation voltageis the VCE voltage drop in a saturated (fully conducting) transistor.• Beta: β=IC/IB• Alpha: α=IC/IEα= β/(β+1)• TransistorPackages are a major factor in power dissipation. Larger packages dissipatemore power.4.16BJT quirksAn ideal transistor would show 0% distortion in amplifying a signal. Its gain would extendto all frequencies. It would control hundreds of amperes of current, at hundreds of degrees C.In practice, available devices show distortion. Ampliﬁcation is limited at the high frequencyend of the spectrum. Real parts only handle tens of amperes with precautions. Care must betaken when paralleling transistors for higher current. Operation at elevated temperatures candestroy transistors if precautions are not taken.