Fig. 2. (a) Structure of the n-p-LDMOS using MP to control MN.(b) Schematic view of circuit diagram and the control method of the n-p-LDMOS.A.Three-Terminal n-p-LDMOS Using MP to Control MNFig. 2 shows the specific method to obtain the inner gate control signal for controlling MN. As shown in Fig. 2(a), a resistor R is connected between the MP and the electrode S to generate a pulse signal VR during the switching of MP. Then, the VR is sent to a low-voltage pulse signal processing circuit to obtain an appropriate square wave sig- nal for controlling MN, and the circuit schematic is shown in Fig. 2(b).The value of the resistor R should not be set too high, otherwise the p-top/n-well junction can be turned ON because of the value of VR is too high when MP is ON, it also increases its power consumption. The resistor R is needed to have a low resistance compared with the ON-resistance of MP and the load resistance ( RL). It is worth pointing out that the resistor R can be realized by a ploysilicon resistor or a diode-connected MOS transistor, and either of them can be implemented on the chip.To confirm the pulse signal of VR , a simulation for the structure shown in Fig. 2(a) is carried out by means of the circuit mode in TCAD MEDICI. In the simulation, the value of R is set to 1.0 × 1050. · μm, and a load resistor RL witha value of 6.8 × 1070. · μm is also used with one terminalconnected to the electrode D of the n-p-LDMOS and the otherterminal applied to a voltage of 310 V. To investigate the practical variation of VR during the processes of turning ON and turning OFF of MP, two specific cases are researched: one is that MN is OFF and the other one is that MN is ON. Fig. 3(a)shows the gate control signal VDG (VDG = VD − VGP) of MP.The simulation results of VR of these two cases during the switching of MP are shown in Fig. 3(b) and (c), respectively. It is seen that when MP is ON and MN is OFF, a value of0.44 V is obtained, whereas a value of 0.09 V is obtainedFig. 3. (a) Timing of control signal VDG of MP. (b) Simulation result of VR while MP is turned ON and OFF and MN is OFF. (c) Simulation result of VR while MP is turned ON and OFF and MN is ON.while MP and MN are both. Hence, it is shown that in the real situation, VR varies from 0.09 to 0.44 V with the switching of MP and MN. Such values, however, cannot be used to control the gate of MN directly, because the voltage level and the driving capability are both low. Thus, a low-voltage signal processing circuit is needed to amplify and drive the obtained pulse signal VR .To increase the voltage level and improve the driving capability of the pulse signal VR , a low-voltage pulse signal processing circuit is designed, and the circuit schematic is shown in Fig. 4(a). The amplifier A with a gain of Au is used for amplifying the pulse signal. Because of the turn-ON process of the device, the voltage drop across R (VR) increases from 0 to 0.44 V (MP is turned ON from OFF-state), and then decreases from 0.44 to 0.09 V (MN is turned ON following the turning ON of MP). In addition, in the turn-OFF process, VR decreases from 0.09 to 0 V (MP is turned OFF from ON- state), so the gain Au of the amplifier needs to be decided reasonably. In this paper, the gain of Au is set to 30. The comparator (COM) shown in Fig. 4(a) is not only used to eliminate the sharp pulse signal of VR caused by the p-top/n- well junction capacitance during the switching of the device, but also used to adjust the delay time between V2 and V1 by changing different reference voltage VRef . The higher of the value of VRef , the longer of the delay time between V2 and V1 in the rising edge, and the shorter of the delay time between V2 and V1 in the falling edge is shown in Fig. 4(b). Because of the fact that the turn-ON speed of the device is much faster than the turn-OFF speed, a relatively high value of 2.5 V for VRef (>30 × 0.09 = 2.7 V) is used in this paper to achieve a short turning OFF delay time between V2 and V1. The buffer driver B shown in Fig. 4(a) is used to improve the driving capability of the pulse signal.The A, shown in Fig. 4(a), is a conventional CMOS differ- ential amplifier. Fig. 4(c) shows a CMOS circuit schematic of