The avalanche mode of operation of bipolar transistors, also known as avalanche generation, is a special mode in which an avalanche breakdown occurs in the transistor junction. This phenomenon is accompanied by a significant increase in current and generation of charge carriers. The latter circumstance is extremely important for understanding this effect, since it allows us to assume the presence of additional, or free, energy in it.

In this work, we will examine some types of bipolar transistors for their operation in avalanche mode, calculate their efficiency when operating in a special circuit, and even obtain overunity results for several types of transistors. But first, let us recall some features of this effect.

The avalanche mode occurs when the reverse voltage at the collector (or emitter) junction reaches a critical level, at which the electrons, accelerating in the electric field, acquire sufficient energy to knock new electrons out of the crystal lattice. This process is self-sustaining, causing an avalanche-like increase in current [1-2].

The main characteristics of the avalanche mode:

• High reverse voltage on the collector-emitter or collector-base;
• Avalanche-like multiplication of charge carriers inside the semiconductor;
• A sharp increase in current, which is limited only by the resistance of the external circuit;
• Possible transition to avalanche-injection mode, in which the transistor operates as an avalanche diode with amplification.

Difference from secondary breakdown
Avalanche breakdown differs from thermal breakdown (secondary breakdown). In the avalanche mode, conductivity is determined by physical processes in the semiconductor, and not by local overheating, which makes it partially controllable if the current is limited.

Avalanche breakdown, which is carried out according to the standard circuit, may work unstably. Its start and steady state largely depend on the parameters of the elements included in the circuit. That is, if you take a transistor from the same brand, but a different batch, the avalanche mode may not start. In such circuitry, all elements, as a rule, require careful selection.

In our subsequent experiments, we will use the KT315A as the base transistor. This transistor was the most common and in demand in the USSR. It had a number of excellent characteristics that made it practically irreplaceable [3]. Its complete analogue is the BFP719. A more modern version of this transistor, made in a TO-92 case, is the KT315A1-P1 [4]. Its analogues are the 2SC544 and 2SC546 brands.

The circuit diagram for connecting a transistor in avalanche mode, which has already become classic, is shown in Figure 1a. Its main disadvantage is the difficulty of selecting a transistor for correct operation in this mode. From a batch of 100 pieces, you can find only a few suitable for use. More stable and less demanding circuits are shown in Figures 1b and 1c. In addition to additional resistance in the base of VT1, they add a storage capacitor C2, or a pair of C2-C3, which, among other things, pushes this transistor to self-excitation. The new circuit design allows us to use various types of transistors for our research.

The operating principle of such circuits is to accumulate a voltage on the capacitor C2 sufficient to break down VT1, while creating a field strength of about 10⁸ V/m in the zone of this semiconductor. If C2 is absent, then accumulation occurs in the diffuse capacitance of the transistor itself. But what is even more significant for our research is the formation of additional electron-hole pairs in the semiconductor during this process.

The operating point is found by the trimmer resistor R1, and the pulse repetition frequency is also regulated. This frequency is also affected by the capacitance C2. Capacitor C1 serves to smooth the supply voltage Up during the pulse. It is recommended to install it as close as possible to the avalanche circuit. The power supply circuit is shown in Figure 1d, where the supply voltage is taken from the LATR transformer (VVT) after its rectification by diode D1. It is recommended to select a LATR designed for an effective voltage in the secondary circuit of 300 V, since the standard value is usually 250 V.

The following oscillograms show the graphs of operation at differentx circuit solutions. The oscilloscope probe is connected between the OS1 circuit output and the common wire. The oscillogram in photo 2 reflects the operation of the KT315A transistor according to circuit 1a, and the oscillogram in photo 3 shows the operation of this transistor in the improved circuit 1b. It is interesting that the pulse in both operating modes does not exceed 5 ns in its duration, but in the second case a “tail” is added to it, which wails in proportion to the product of the capacitance C2 and the resistance R3. Its attenuation time is found as follows: \[ \tau \approx 0.7\cdot C_2 R_3 \tag{1}\] The front of these pulses is no more than 2 ns.

From the oscillograms presented above, it is clearly visible that due to the addition of capacitor C2, the pulse amplitude increases by about 10 times, the pulse itself almost does not change in shape, but a "tail" is added to it, forming a discharge circuit of C2 through resistance R3.

If we now measure the pulse parameters and the energy costs of its creation, we can find the efficiency of the avalanche mode circuit of various transistors. Here we need to take into account one point: when charging the capacitor, half of the energy is dissipated on the total resistance of the power source + R1 + R2 + open channel resistance of transistor VT1 [5]. Thus, the overall efficiency, taking into account the dissipation energy, will be doubled, which we will reflect in the following table. But first, let's clarify some of its parameters.

Up - maximum amplitude of the supply voltage of the circuit. Usually, at such an amplitude, the transistor enters the avalanche mode.

Urms - root-mean-square (effective) value of the voltage in the pulse on the resistor R3, which was measured at a sweep of 500 nanoseconds per division, and was practically independent of the supply voltage (starting from a certain minimum value). It turned out that the property is exclusively the transistor itself and the capacitor C2 (or C2 + C3). The smaller the Urms, the less the circuit consumed from the Up power source, and, as a rule, the higher the pulse frequency. But these parameters did not affect the efficiency, so they are not given in Table 1.

h₂₁ (h_FE) - transistor current gain [6]. Here we take some average value from several tested samples.

η/2 - transistor efficiency in avalanche mode.

η - efficiency of the entire circuit, taking into account energy dissipation on the internal resistances of the circuit.

The author almost by chance came across one copy of the KT315B transistor, which gave modest, but still superunity parameters. They are reflected in the last line of Table 1. These data have been repeatedly rechecked and can be considered confirmed. In other words, out of a hundred transistors included in the batch, only one can have such characteristics, and even this one can demonstrate them only if the operating mode is correctly configured.

Below are oscillograms of some measurements carried out according to the 1b scheme, at different device sweeps and for different transistors. The blue beam reflects the amplitude of the supply voltage (Up), and the yellow beam shows the pulse itself.

Here, special attention should be paid to the pulse amplitude, which exceeds the supply voltage amplitude, although there are no inductors in the circuit that could contribute to this. This is clearly visible in Figures 4, 6, 8. The author noted that the better the transistor operates in avalanche mode, the greater this excess in amplitude. In Figures 5, 7, 9, these pulses are visible against the background of the supply voltage.

A stand for research can be made with your own hands from a breadboard. The only recommendation is to place all the elements as close to each other as possible. It is better to make a stand on connectors so that its elements can be easily replaced.

All resistances must be calculated for a power of at least 0.5 W. This is especially true for R3, which dissipates a large pulse power. On the stand, the resistor Rb can be made adjustable.

The containers must be used for a voltage of at least 500 V, and they can be any, except ceramic.

1) The excess of the pulse amplitude over the amplitude of the supply voltage indicates an additional charge that appears in the avalanche mode when generating charge carriers. This implies the potential possibility of using this mode for super-unity devices, which is confirmed by an experiment with the KT315B transistor. However, at the moment, both these transistors themselves and the circuits based on them have a rather low efficiency, and the excess energy is dissipated as heat on the active resistances of the circuit.

2) The efficiency of the avalanche circuit can be approximately calculated. For a sweep of 500 ns per cell, the approximate formula for calculating the overall efficiency of the circuit looks like this: \[ \eta \approx 6.5 {U_{rms} \over U_p} \tag{2}\] It is assumed that the pulse fits completely on the oscilloscope screen. It follows from the formula that it is very important to obtain as high an effective voltage (Urms) on the load resistor R3 as possible in comparison with the supply voltage (Up). Obviously, such a formula shows the average value for a large sample of transistors. For example, individual samples of KT315E and KT315B dropped out of it in our experiments.

3) Urms, in turn, does not depend on the supply voltage, starting from a certain minimum value. This is a property exclusively of the transistor itself and the storage capacitor C2, if used.

4) The theory that the current gain of a transistor (h₂₁) affects its ability to operate in avalanche mode and its efficiency in this mode is not confirmed.