Designing Next Generation Surge Current Testers

The most important parameter of power semiconductor devices reflecting their overload capacity is the surge current – the maximum permissible current amplitude of a semi-sinusoidal shape lasting 10 ms. Increasing the power capacity of semiconductors and designing devices with a rectifier element diameter of 100 mm or more both require a surge current testing system that can form current pulses up to 100 kA.

To solve this problem, a number of requirements must be taken into account. First, during the pre-breakdown state with the so-called direct current filamentation [1] the voltage drop across the test sample sharply increases and can reach 60V for high-voltage semiconductors. Secondly, the tester must ensure high accuracy in setting the current amplitude, since forming a current pulse with an amplitude of the order of 100 kA requires absolute deviation from the set value within 2-3 kA. Thirdly, it is necessary to ensure structural rigidity, taking into account the enormous dynamic forces arising during the current pulse.

The main approaches to the design of surge current testers are based on forming the surge current pulses either using a stepdown transformer [2,3] or by discharging a storage capacitor in an oscillatory circuit [3]. Using step-down transformers to form high-amplitude current pulses has a number of disadvantages, including an unacceptable effect on the power supply from an asymmetric current load, low setting accuracy with the impossibility of smooth current amplitude adjustment, the effect of supply voltage fluctuations on the set amplitude value, excessive weight and size. Today it is more common to form surge current pulses using an oscillatory circuit.

However, obtaining a pulse shape close to a proper half-sine wave requires a high-quality factor of the oscillating circuit, which is a problem in case of high losses. Losses in the oscillatory circuit due to losses in the conductor system and a significant voltage drop across the test sample lead to the emergence of an aperiodic component in the form of a current pulse, which is unacceptable under the test conditions. In addition, generating current pulses with an amplitude of tens of kiloamperes requires storage capacitors with capacitance up to thousands of microfarads. And finally, the stability of the generated current pulse in the oscillatory circuit is determined by the stability of the capacitor’s properties.

The need for storage capacitors of significant capacity with stable parameters and larger cross-section of the busbars in the tester to reduce losses cause a significant increase in the weight and dimensions of the tester. In [4] it was proposed to make test equipment based on active shapers – current sources with powerful MOSFETs. The principle of using a transistor as the current source lies in its operation on a linear portion of the transfer curve. A typical MOSFET curve is shown in Figure 1.

 

Figure 1: MOSFET transfer curve

 

It shows that a change in the gate-source voltage of the transistor within 5-6 V leads to a change in the drain current by about 40A. As mentioned above, the voltage on the semiconductor can increase to 60V during the pre-breakdown state. One of the conditions for the operation of a current source is that the load voltage should not exceed the supply voltage of the current source. In other words, the current regulator must have a certain voltage margin. The supply voltage of the current source of 100V would be optimal. This value is chosen on the basis that it is the minimal value sufficient for stable operation of the current source, and the fact that the industry offers electrolytic capacitors with an operating voltage of 100V used as energy storage devices. Being in the linear region of the transfer curve, a certain amount of power is emitted on the transistor. The amount of this power is determined by the transistor’s area of safe operation (SOA). The SOA diagram of the IRFPS3815 transistor is shown in Figure 2.

 

Figure 2: SOA of a MOSFET

 

The graph shows that at a drain-source voltage of 100V and a 10ms long rectangular current pulse the safe current amplitude is 10A. For a semi-sinusoidal shape, the permissible amplitude can be increased to 16A taking into account the shape factor. In addition, partial discharge of storage capacitors throughout the formation of a current pulse will decrease the drain-source voltage at the transistor and shift the operation mode of the transistor further into the safe area. Given this margin, the current value was selected at 13A.

To calculate the thermal mode of the transistor operation, we assume that forming a current pulse of 10 ms causes a drop of voltage across the capacitors from 100 V to 60 V. At the same time, average value of the drain-source voltage on the transistor is about 80 V. Since the current has a sinusoidal shape with an area 0.63 times smaller than the area of the rectangle, the effective value of the current equals 8.2A. Then the pulse power on the transistor will be:

$$P_i = V_{DS} \times I_D = 80V \times 8.2A = 656W$$

 

In order for the transistor structure to cool down between the current pulses and the storage capacitors to charge, it is necessary to provide a pause between current pulses with a duration of at least 60 s. With this in mind, the average power on the transistor will be equal to:

$$P_a = \frac{P_i \times t}{T} = \frac{656W \times 10ms}{60s} = 0.11W$$

 

To determine the temperature of the transistor case at the highest possible ambient temperature of 35:

$$T_J = P_i \times Z_{th} + T_C = 656W \times 0.15^{\circ} C/W + 35^{\circ} C= 133^{\circ} C$$

 

Then the maximal temperature of the transistor chip will equal:

The maximum operation temperature of the chip is 175 °C, therefore, the temperature margin in the maximal load mode exceeds 40 °C. It was confirmed in practice when determining the maximum permissible current amplitude of the transistor, which made 20A.

A voltage source with low internal resistance is required to form a current pulse. As mentioned above, electrolytic capacitors were used for this. To calculate the required capacity of the storage capacitor, it must be taken into account that the voltage should not fall below 60V. At an initial voltage level of 100V, the voltage margin will be 40V. The capacitor's capacity is:

$$C = \frac{I \times dt}{dU} = \frac{13A \times 10ms}{40V} = 3250 \mu F$$

 

A capacitor with a capacity of 3300 μF per 100V was selected from the standard offer on the market.

In addition to the high accuracy of adjustment, the advantage of using current sources is the almost unlimited possibility of their parallel connection with identical sources to achieve a required value of the current amplitude, as shown in Figure 3. The functional chart of the surge current tester designed on the basis of current sources is shown in Figure 4.

 

Figure 3: Parallel connection of current sources.

 

The functional chart shows the voltage source 1 connected to the storage capacitor 2. The positive terminal of the storage capacitor 2 is connected in parallel to drains of N MOSFETs 3. Resistor 4 is installed at the source of each MOSFET. A common point of the resistors 4 is connected to the anode of the device under test 5, the cathode of which is connected through a shunt (current sensor) 6 to the negative terminal of the storage capacitor 2. The measuring leads of the shunt (current sensor) 6 are connected to the measurement unit 7. The interconnected gates of the MOSFETs 3 lead to the output of the amplifier 7. The inverting input of the amplifier 8 is connected to the source of the first MOSFET 3, while the non-inverting input of the amplifier 8 is connected to the output of the reference signal shaper 9.

 

Figure 4: Functional chart of the tester

 

The first switch 10 is installed between the inverting input of the amplifier 8 and its output, while the second switch 11 and the voltage limiter 12 are connected between the gates of the MOSFETs 3 and the common point of the resistors 4. The control inputs of the first and second switches 10 and 11 are connected to corresponding outputs of the synchronization circuit 13. Its third input is connected to the input of the driver of the reference signal 9, and the fourth input is connected to the input of the control generator 14. The output of the control generator 14 is connected to the control terminals of the tested semiconductor device 5.

 

The next article in this short series will explain the inner-workings of the surge current tester.

 

About Proton Electrotex

Proton-Electrex is one of the Russian leaders in the design and manufacture of power semiconductor diodes, thyristors, modules, coolers, IGBTs (IGBTs), as well as power units for use in various electric energy converters. The company was founded and began production in 1996 on the leased premises of the Proton plant. Since then, Proton-Electrotex has developed its own infrastructure for the entire production cycle. Production is equipped with modern production lines, measuring equipment of our own production and areas for “clean technologies” with full compliance with the requirements for electronic products and microelectronics.

 

Sources

1. Burcev E.F., Grehov I.V., Kryukova N.N. Lokalizaciya toka v kremnievyh diodah pri bol'shoy plotnosti toka. – FTP, 1970, № 10, p.1955 – 1962.
2. V.M. Bardin, L.G. Moiseev, Zh.G. Surochan, O.G. Chebovskiy. Apparatura i metody kontrolya parametrov silovyh poluprovodnikovyh ventiley. Moscow: Energia. 1971.
3. R. Lappe, F. Fischer Izmereniya v energeticheskoy elektronike. Moscow: Energoatomizat. 1986.
4. Utility model patent no. 185719 “Stend dlya ispytaniy silovyh poluprovodnikovyh priborov na stoykost' k vozdeystviyu udarnogo toka”.

 

This article originally appeared in the Bodo’s Power Systems magazine.