Semiconductor Valves for High-Voltage Direct-Current Converters
The rapid growth in the use of h.v.d.c. since about 1980 has been due to the development of high-voltage, high-current semiconductor devices. These superseded the previously used complex and expensive mercury arc valves that employed a mercury pool as cathode and a high-voltage graded column of anodes, with the whole enclosed in steel and ceramic to provide a vacuum tight enclosure. Nowadays, the semiconductor devices are stacked to form a group which is able to withstand the design voltages and to pass the desired maximum currents - this group is termed a 'valve'.
Thyristors are manufactured from silicon wafers and are four-layer versions of the simple rectifier p-n junction, as shown in Figure 9.3(a). The p layer in the middle is connected to a gate terminal biased such that the whole unit can be prevented from passing current, even when a positive voltage exists on the anode. By applying a positive pulse to the gate, conduction can be started, after which the gate control has no effect until the main forward current falls below its latching value
Figure 9.3 (a) Structure of a four-layer thyristor. (b) Symbol, (c) Thyristor characteristic: lg gate current to switch thyristor on at forward voltage
(see Figure 9.3(c)). This current must be kept below the latching value for typically 100 ms before the thyristor is able to regain its voltage hold-off properties. (Note that in forward conduction there is still a small voltage across the p-n junctions, implying that power is being dissipated - hence the semiconductor devices must be cooled and their losses accounted for.)
In practice, many devices, each of rating 8.5 kV and up to 4000 A, are stacked in a valve to provide a rating of, say, 200 kV, 4000 A. Valves are connected in series to withstand direct voltages up to 800 kV to earth on each 'pole'. Each thyristor can be 15 cm in diameter and 2 cm depth between its anode and cathode terminals. A typical device is shown in Figure 9.4(a) and a valve in Figure 9.4(b).
Figure 9.4 (a) High-power thyristor silicon device (Reproduced with permission from the Electric Power Research Institute, Inc.). (b) Thyristor valves in converter station (Reproduced with permission from IEEE.)
Figure 9.5 Circuitry associated with each thyristor
In order to turn on a thyristor a pulse of current from a gate circuit that is at the same potential as the cathode is required. Alternatively, a light trigged thyristor is triggered by a light pulse sent through a fibre-optic channel.
When many thyristors are connected in series to form a valve, it is necessary to:
- 1. obtain a uniform voltage distribution across each thyristor.
- 2. retain uniform transient voltage distributions with time.
- 3. control the rate of rise of current.
These are achieved by the auxiliary electrical circuitry shown in Figure 9.5. The inductor, L, limits the rate of rise of current during the early stage of conduction. The chain Rj, Q bypasses the thyristor, thus controlling the negative recovery voltage. The d.c. grading resistor, R0, ensures uniform distribution of voltage across each thyristor in a valve.
Insulated Gate Bipolar Transistors
Recently the insulated gate bipolar thyristor (IGBT) has been used in h.v.d.c. schemes. The IGBT is a development of the MOSFET, in which removal of the voltage from the gate switches off the through current, thereby allowing power to be switched on or off at any point of an a.c. cycle. IGBTs have ratings up to 4kV and 1000 A, although the on-state voltage drop and switching losses are larger than with a thyristor.
When a sufficient voltage is applied to the Gate with respect to the Emitter, it inverts the p region below the Gate (shaded area) thus forming a diode between the Emitter (n region) and Collector (p substrate). As shown in Figure 9.6(c), then when
Figure 9.6 (a) structure of an IGBT (b) symbol, (c) characteristic
in addition the Emitter to Collector voltage is greater than 0.7 V, conduction between Emitter and Collector commences.