Comparative Study of Switches of Powerful Nanosecond Pulses Based on Blocks of Series-Connected Shock-Ionized Dynistors

Abstract

Two types of switches for high-power current pulses with an operating voltage of 24 kV based on blocks of series-connected shock-ionized dynistors are described. The switching processes of these switches are considered, and the results of their comparative study are presented. The possibility of efficient switching of nanosecond current pulses with an amplitude of 6 kA and a rise rate of 150 A/ns is shown. The prospects for increasing the switched power are determined.

Energy losses in switches of high-power pulses with a submicrosecond rise time significantly depend on the time of their switching to a high-conductivity state. Therefore, the four-layer silicon structures described in [1–4], which have a subnanosecond turn-on time due to the use of the method proposed in [5] for triggering by a nanosecond high-voltage pulse, are of great interest. With such a duration of the triggering action, the avalanche breakdown does not have time to develop, and the field strength in the base regions of the semiconductor structure increases to a value sufficient to initiate the impact ionization of silicon. This process provides a very high rate of increase in the conductivity of the silicon structure, which is not currently available for alternative semiconductor switches.

The work [6] shows the high switching capabilities of a semiconductor switch with an operating voltage of 12 kV, which is made in the form of a block of series-connected shock-ionized dynistors (SIDs) described in [7–9]. The disadvantage of such a switch is the large amplitude of the output voltage of the control circuit (~27 kV), which provides switching of all dynistor structures. This drawback makes it difficult to develop SID switches with higher operating voltage, but it can be eliminated by using several serially connected SID blocks with independent control circuits.

Figure 1 shows a diagram of an experimental generator containing an SID switch with an operating voltage of 24 kV, consisting of SID₁ and SID₂ blocks, which are switched on in a relay-race mode. Block SID₁ is switched on by the CC control circuit, and block SID₂ is switched on as a result of the discharge of the capacitor С₃ due to the switching of the block SID₁. In the initial state, the capacitor С₃ is charged to charging voltage U₀ of storage capacitor С₀; voltage is applied to blocks SID₁ and SID₂, which is determined by the number of dynistor structures used in these blocks; capacitor С is charged to the voltage applied to the block SID₁. The control circuit contains the switch Sw, capacitor С₁ charged to a small voltage U₁, magnetic compression link С₂–Tr, and a block of drift step-recovery diodes (DSRD), first described in [10]. Сircuit C–D and resistor R eliminate the possibility of applying high voltage U₀ to the control circuit.

Fig. 1.

Electric circuit of the generator based on relay-switched SID blocks. DSRD—Block of diode structures with a diameter of 20 mm; SID₁, SID₂—blocks of dynistor structures with a diameter of 24 mm; D—block of diodes K100F (two in series); D₁—K100F; Tr—transformer (ferrite core N87 (EPCOS), four rings 29.5 × 19 × 20 mm in size, w₁ = 1, w₂ = 3); L₀—choke (ferrite core 600 NN, four rings 31 × 20 × 6 mm in size, one turn).

Switching of blocks SID₁ and SID₂ is carried out as follows.

After turning on the switch Sw, a current with a fundamentally short duration (~300 ns) flows through the DSRD block in the forward direction, which ensures charge accumulation in diode structures and charges the capacitor С₂. At the end of the charging process, the core of the transformer Tr is saturated and the inductance of its secondary winding decreases sharply. As a result, the capacitor С₂ is discharged through the DSRD-w₂ circuit and a reverse current flows through the DSRD block, which, in a fundamentally short time (~120 ns), removes the entire accumulated charge from the DSRD structures and ensures their turn-off. In the process of turning off the DSRD unit, an output current pulse of the control circuit is generated. Its amplitude is determined by the amplitude of the current flowing through the winding w₂, and the rise time is determined by the turn-off time of the DSRD block (~2.5 ns).

The control circuit output current flows through the capacitor С and is switched to the SID₁ block and into the SID₂–L₃–C₃ circuit. Since, at the nanosecond front of the switched current pulse, the electrical resistance of the elements L₃ and C₃ is relatively large, the control circuit output current flows mainly through the SID₁ block. As a result, the own capacitances of the dynistor structures of this unit are quickly charged. Current flowing through the circuit SID₂–L₃–C₃ discharges the own capacitances of the dynistor structures of the SID₂ block and provides an increase in voltage across the capacitor С₃. When the voltage on the dynistor structures of the SID₁ block increases to the threshold value, which initiates the process of impact ionization, they switch to a high-conductivity state in a time <1 ns. As a result, the capacitor С₃ rapidly discharges through the circuit L₃–SID₂–SID₁, and the voltage on the SID₂ block increases sharply. Since the energy losses in the capacitor discharge circuit С₃ are small and its capacitance is much larger than the own capacitance of the SID₂ block, then the voltage on the SID₂ block can reach a value significantly exceeding the voltage across the capacitor С₃ at the time of switching on the SID₁ unit. Block SID₂ turns on when the voltage on it reaches the threshold value, which ensures the initiation of the process of impact ionization in the structures of the dynistors of this block.

In the process of switching the SID switch, the choke L₀ has a large inductance, which limits the rate of change of current in the circuit of storage capacitor С₀. A few nanoseconds after turning on the SID₂ block, the core of the choke L₀ saturates, the inductance of the choke decreases sharply, and the discharge current of the capacitor С₀ grows rapidly. In the interval between the moment of switching on the SID₁ block and the saturation moment of the core of the choke L₀, the output current of the control circuit flows through the diode D and through the dynistors of SID₁ block, providing high conductivity of the SID₁ block.

Figure 2 shows an oscillogram of the current through the SID switch, corresponding to the voltage U₀ = 24 kV. It is obtained using a pilot SID switch, in which the SID₁ and SID₂ blocks consisted of the same number of dynistor structures (six in each block). The DSRD block consisted of 25 diode structures. Capacitor С₁ was charged to voltage U₁ = 1.4 kV. The output current of the control circuit was ~350 A.

Fig. 2.

Oscillogram of the current through the SID switch.

As can be seen from the oscillogram, the current through the SID switch has an amplitude of ~6 kA and increases at a rate of ~150 A/ns. A indicates a low electrical resistance of the SID switch.

Studies have shown that the electrical resistance of the pilot SID switch increases significantly when U₀ < 20 kV. Since the trigger conditions for the SID₁ block did not change, then the result obtained indicates an increase in energy losses in the SID₂ block. This can be explained by the fact that, at low voltage U₀, the circuit of capacitor C₃ does not effectively enable the SID₂ block consisting of six series-connected dynistor structures.

To ensure efficient operation in the range of the voltage U₀ from 24 to 12 kV, the pilot SID switch has been optimized by reducing the number of dynistor structures in the SID₂ block. In an optimized SID switch, the SID₁ block contained eight dynistor structures and the SID₂ had four. Increasing the number of dynistor structures in the SID₁ block determined the need to increase the charging voltage of the capacitor С₁ (up to 1.6 kV) and increasing the number of diode structures in the DSRD block (up to 35 pieces).

Figure 3 shows an oscillogram of the voltage drop across the optimized SID switch, which was obtained during the study of its switching process. The amplitudes of the voltage peaks are determined by the turn-on voltages of the SID₁ and SID₂ blocks.

Fig. 3.

Voltage waveform on an optimized SID switch.

An alternative way to ensure the efficient operation of a high voltage SID switch in a generator with deep regulation of the capacitor charging voltage С₀ is the use of a modular design: serially connected SID units are switched autonomously using separate control circuits.

Figure 4 shows a diagram of a modular SID switch with an operating voltage of 24 kV. It consists of two SID₁ and SID₂ blocks with an operating voltage of 12 kV connected in series. Triggering of blocks SID₁ and SID₂ is carried out using the same CC based on blocks DSRD₁ and DSRD₂, similar to the control circuit in Fig. 1. Transformers Tr have one well-insulated turn in the primary winding. As a result, the breakdown voltage between the windings w₁ and w₂ significantly exceeds the maximum voltage on the SID switch.

Fig. 4.

Electrical diagram of the generator based on the modular SID switch. DSRD₁, DSRD₂—25 diode structures with a diameter of 20 mm; SID₁, SID₂—12 dynistor structures with a diameter of 24 mm; D—block of diodes K100F (two in series), D₁—K100F; Tr—transformer (ferrite core N87 (EPCOS), four rings 29.5 × 19 × 20 mm in size, w₁ = 1, w₂ = 3); L₀—choke (ferrite core 600 NN, four rings 31 × 20 × 6 mm in size, one turn).

In the diagram in Fig. 4, the formation of nanosecond pulses triggering blocks SID₁ and SID₂ happens as follows. After turning on the switch Sw, capacitor С₁, initially charged to voltage U₁ = 2.8 kV, is discharged through the windings w₁ of transformers Tr. As a result, short current pulses flow through the circuits С₂–DSRD₁ and DSRD₂–С₂, which ensure the charging of capacitors С₂ and accumulation of charge in diode structures of blocks DSRD₁ and DSRD₂. When the capacitors С₂ are charged to maximum voltage, transformer cores Tr are saturated and capacitors С₂ are quickly discharged through the windings w₂ of transformers Tr. As a result, through blocks DSRD₁ and DSRD₂, rapidly increasing currents flow in the opposite direction, which remove the accumulated charge from the DSRD structures.

The entire accumulated charge is removed by the time the capacitors С₂ are completely discharged and the currents through the windings w₂ reach the maximum value (~300 A). At this point, the DSRD structures turn off and the currents flowing through the windings w₂ are switched in the SID₁–С and С–SID₂ circuits. These currents provide fast charging of the own capacitances of dynistor structures of SID₁ and SID₂ blocks up to the switching voltage. Through the use of a common switch Sw, almost identical conditions for the formation of trigger pulses of SID₁ and SID₂ blocks are created.

Figure 5 shows a typical voltage oscillogram on the SID switch, illustrating the process of turning it on when the voltage is U₀ = 24 kV. The step-wise shape of the voltage indicates that the SID₁ and SID₂ blocks turn on not simultaneously. After the leading SID block is turned on, the control circuit of the lagging SID block ensures its turn on with a small delay, which is determined by the technological spread of the parameters of the control circuit elements.

Fig. 5.

Voltage waveform on a modular SID switch.

The voltage amplitude on the SID switch is defined as the sum of the amplitude values of the voltage on the leading SID block at the moment of its turn on and the voltage that is applied to the lagging SID block at that moment. It is significantly less than the sum of the amplitudes of the switching-on voltages of the SID₁ and SID₂ blocks, each of which uses six dynistor structures with a turn-on voltage of ~4 kV.

Comparative study of generators made according to the schemes in Figs. 1 and 4 showed that, when using the same circuit С₀–L₀ and the same voltage U₀, discharge current of capacitor С₀ was practically the same, which indicates approximately the same energy losses in SID switches with autonomous and relay-race mode sitching of SID blocks.

Since virtually the same energy (~120 mJ) stored in the capacitors С₁ and С₃ in the diagram in Fig. 1 and in the capacitor С₁ in the diagram in Fig. 4 was also used to switch the SID blocks, we can assume that the considered high-voltage SID switches have approximately the same efficiency. Their main advantage is the possibility of increasing the operating voltage by connecting additional SID blocks.

In the considered generators, the switches Sw were made on the basis of pulse integrated thyristors with a working area of ~0.5 cm² described in [11]. The current pulses were measured using small-sized current sensors developed by the Pearson Electronics, Inc. Voltage waveforms were obtained using a Tektronix P6015A probe.