The two commercially most important forms of bipolar logic are emitter-coupled logic (ECL) and transistor-transistor logic (TTL or T²L). In this chapter, we examined simple prototype circuits for the ECL and TTL gates, and then we explored the full gate structures.

ECL: ECL logic has traditionally operated from a negative supply voltage, typically -5.2 V,and V_H and V_L are therefore negative. The voltage transfer characteristic for the ECL gate was investigated, and the ECL logic swing ΔV is relatively small ranging between 0.2 V and 0.8 V with noise margins approaching ΔV/2. The ECL gate introduced two new circuit techniques,the current switch and the emitter-follower circuit, and also requires a reference voltage circuit.ECL logic gates generate both true and complement outputs, and the basic ECL gate provides the OR-NOR logic functions. Standard ECL unit-logic families provide delays in the 0.25- to 5-ns range with a power-delay product of approximately 50 pJ.

• Current switches: The current switch consists of two matched BJTs and a current source.
• This circuit rapidly switches the bias current back and forth between the two transistors,based on a comparison of the logic input signal with a reference voltage. In the ECL gate,the transistors actually switch between two points in the forward-active region, which is one reason why ECL is the highest speed form of bipolar logic. A second factor is the relatively small logic swing, typically in the 0.4- to 0.8-V range. The low ECL logic swing results in noise margins of a few tenths of a volt. ECL is somewhat unusual compared to the other logic families that have been studied in that it is typically designed to operate from a single negative power supply, historically -5.2 V, and V_H and V_L are both negative voltages.
• Emitter followers: In the emitter-follower circuit, the output signal replicates the input signal except for a fixed offset equal to one base-emitter diode voltage, approximately 0.7 V. In ECL, this fixed-voltage offset is used to provide the level-shifting function needed to ensure that the logic levels at the input and output of the gates are the same. The emitter followers permit additional logic power through the use of the "wired-OR" circuit technique.
• Reference circuits: Temperature-compensated reference circuits are used to provide the reference voltage required in the ECL gate. Temperature compensation is achieved using diodes that closely match the bipolar transistors in the circuit. In integrated circuits, diodes are usually realized by connecting the collector and base terminals of a transistor together.
  
  TTL circuits: Classical TTL circuits operate from a single 5-V supply and provide a logic swing of approximately 3.5 V, with noise margins exceeding 1 V. During operation, the transistor sin standard TTL circuits switch between the cutoff and saturation regions of operation. Basic TTLgates realize multi-input NAND functions; however, more complex gates can be used to realize almost any desired logic function. Standard TTL unit-logic families provide delays in the 3- to 30-ns range, with a power-delay product of approximately 50 pJ. Schottky diodes are used to prevent BJT saturation and speed up the TTL logic circuits.
• BJT saturation region: The collector-emitter saturation voltage of the BJT is controlled by the value of the forced beta, defined as β_FOR = i_C/i_B. The transistor enters saturation if the base current exceeds the value needed to support the collector current (that is, i_B > i_C/β_F so that &946;_FOR < β_F). An undesirable result of saturation is storage of ex-cess charge in the base region of the transistor. The time needed to remove the excess charge can cause the BJT to turn off slowly. This delayed turnoff response is characterized by the storage time t_S and is proportional to the value of the storage time constant τ_S, which determines the magnitude of the excess charge stored in the base during saturation.
• Schottky-clamped transistors: The Schottky-clamped transistor merges a standard bipolar transistor with a Schottky diode and was developed as a way to prevent saturation in bipolar transistors. The Schottky diode diverts excess base current around the base- collector diode of the BJT and prevents heavy saturation of the device. Schottky TTL circuits offer considerable improvement in speed compared to standard TTL for a given power dissipation because the storage time delays are eliminated.
• Inverse operation: The input transistors in a TTL gate operate in the reverse-active mode when the input is in the high state. This is the only use of this mode of operation that we encounter in this text.
• Fan out: The TTL gate has relatively large input currents for both high- and low-input voltages. The input current is positive for high-input levels and negative for low-input levels. This input current limits the fan out capability of the gate, and the fanout capability of TTL was analyzed in detail. At the output of the TTL gate, another emitter follower can be found. The emitter follower provides the high-current drive needed to support large fanouts as well as to rapidly pull up the output.
• TTL family members: TTL gates are available in many forms, including standard, low-power, high-power, Schottky, low-power Schottky, advanced Schottky, and advanced low-power Schottky versions. Standard TTL has essentially been replaced by low-power Schot-tky (54LS/74LS) TTL, which provides similar delay but at reduced power. SchottkyTTL (54S/74S) provides a high-speed alternative for circuits with critical speed require-ments.

Power delay products: The standard ECL and TTL unit-logic families have relatively large power-delay products (20 to 100 pJ), which are not suitable for high-density VLSI chip designs.VLSI circuit densities require subpicojoule power-delay products; simplified circuit designs with much lower values of power-delay product are required for VLSI applications. Low-voltage forms of TTL and ECL have been designed for VLSI applications, but for the most part they have been replaced by CMOS circuitry.

BiCMOS: BiCMOS is a highly complex integrated circuit technology, but it provides the advantages of both bipolar and MOS transistors. Full Bicmos technologies provide NMOS,PMOS, npn, and pnp transistors. Thus the circuit designer has maximum flexibility to choose the best device for each place in a circuit. In BiCMOS logic gates, MOS transistors are typically used to provide high-impedance inputs with the simplicity of MOS NAND, NOR, and complex gate implementations. Bipolar transistors are used to provide high-output-current capacity for driving large load capacitances. BiCMOS is also highly useful for "mixed-signal" designs that combine both analog and digital signal processing. S implied BiCMOS technologies add npn transistors to the CMOS process, and the resulting circuits are often referred to as BiNMOS instead of BiCMOS.