20 Unijunction Transistor Applications
June 1968 Radio-Electronics

June 1968 Radio-Electronics [Table of Contents] Wax nostalgic about and learn from the history of early electronics. See articles from Radio-Electronics, published 1930-1988. All copyrights hereby acknowledged.

Unijunction transistors (UJT) were relatively new semiconductor devices in 1968 when this article appeared in Radio-Electronics magazine. Of course for that matter most semiconductor devices were still relatively new then. Some commercial products were still being made with vacuum tubes. The "transistor" part of the name is somewhat of a misnomer due to its physical construction, in that there is no rectifying PN junction between the two bases, only a single PN junction (emitter) embedded part-way into the channel between the device's two "base" connections. There is no collector. The UJT is sometimes referred to as a double-based diode, although there is no PN junction separating the two base connections. It usefulness lies primarily in a negative resistance region which make it a prime candidate for a simple reflex type oscillator. It also is useful as a triggered switch due to a characteristic where current between the two base connections increases suddenly at the "peak-point voltage." The UJT operates somewhat similar to the thyristor, although without a PN junction in the current path.

20 Unijunction Transistor Applications

Part 1 of 2 parts-to acquaint you with this versatile solid-state device

By R. M. Marston

All that most electronics amateurs know about the unijunction transistor is that it is sometimes used as a simple code practice oscillator or as a trigger for SCR's. Actually, the device has many more uses: It can be used as a very stable wide-range oscillator, and can be made to generate a whole range of different waveforms. It can also be made to act as an analog-to-digital converter, as a frequency divider, as a lamp flasher, as a time-delay unit, or as a number of other useful circuits.

In this article we'll show you what the unijunction transistor is and how it works. Then we'll introduce you to 20 or so circuits you can build around this amazing little device.

Basic Theory

The unijunction transistor is a very simple device. Its symbol is shown in Fig. 1-a and resembles the actual construction, as shown in Fig. 1-b. The device is made of a bar of n-type silicon material with nonrectifying contacts (base 1 and base 2) at both ends, and a third, rectifying, contact (emitter) alloyed into the bar part way along its length. The third contact forms the only junction within the unijunction transistor (UJT).

Since base 1 and base 2 are non-rectifying contacts, a simple resistance appears between these two points. This interbase resistance is that of the silicon bar and is given the symbol R_BB. Normally the value of R_BB is between 4000 and 12,000 ohms, depending on the construction of the UJT. It measures the same in either direction.

In use, base 2 is connected to a positive voltage and base 1 is connected to ground (or the negative side of the supply). Thus R_BB acts as a voltage divider with a gradient varying from maximum at base 2 to zero at base 1. As the emitter junction is at some point between base 1 and base 2, some fraction of the applied voltage also appears between the emitter junction and base 1. This fractional part of the applied voltage is the most important parameter of the UJT and is called the intrinsic standoff ratio, or η. The value of η is usually between 0.45 and 0.8.

The equivalent circuit - Fig. 1-c -of the UJT clearly illustrates the above points. Symbols r_B1 and r_B2 represent the resistances of the silicon bar, and diode D1 represents the rectifying junction formed between the emitter and the bar. When an external voltage (V_BB) is applied between base 2 and base 1, a voltage equal to η times V_BB appears across r_B1.

If a positive input voltage (V_E) is now applied between the emitter and base 1, and is less than η times V_BB, diode D1 will be reverse-biased, and no current will flow from emitter to base 1. Thus, under this condition, the emitter input appears as a very high impedance. This impedance is that of a reverse-biased silicon diode, and typically has a value of several megohms.

Table 1 - 2N2646 Characteristics

When V_E is increased above η times V_BB, a point will be reached where D1 becomes forward-biased and current starts to flow from the emitter to base 1. This current consists mainly of minority carriers injected into the silicon bar. These carriers drift to base 1, causing a decrease in the effective resistance of r_B1. This decrease in resistance causes the forward bias of D1 to increase, thereby causing the current to increase even more, and in turn causing r_B1 to fall even more. A semi-regenerative action takes place, and the emitter input impedance falls, typically, to about 20 ohms.

Thus, the unijunction transistor acts as a voltage-triggered switch. The precise point at which triggering occurs is called the peak-point voltage (V_P). It is given by V_P = η x V_BB + V_D, where V_D = forward voltage drop of the emitter diode (usually about 600 mV).

One of the most common applications of the UJT is the relaxation oscillator shown in Fig. 2-a. Here, when the supply is connected, C charges exponentially toward V_BB via R, but as soon as the capacitor potential reaches V_P the unijunction fires and C discharges rapidly into the emitter. Once C is effectively discharged, the UJT switches off, C starts to charge up again, and the process is repeated. A sawtooth waveform is generated between Q1 emitter and ground.

In this circuit, final switchoff actually occurs in each cycle when the capacitor discharge current falls to what is known as the valley-point current (I_V), generally a value of several milliamps. A minimum current is needed to start the switch-on action; it is known as the peak-point emitter current (I_P) , typically a value of several microamps.

The frequency of operation of the circuit is given approximately by f = 1/CR, and is virtually independent of supply line potential. A 10% change in supply voltage results in a frequency change of less than 1%. The actual value of R can be varied between a minimum of about 3000 and a maximum of about 500,000 ohms. Hence a very attractive feature of the circuit is that it can be made to cover a frequency range greater than 100 to 1, using a single variable resistor.

Frequency stability is very good with changes in temperature, being about 0.04%/°C. The main cause of this frequency variation is changes in V_D with temperature, these changes being about -2mV/°C. If better frequency stability is required, it can be obtained either by wiring a couple of diodes in series with base 2, or by connecting a stabilizing resistor (R_S) in the same place.

The interbase resistance of the unijunction increases by about 0.8%/°C, so the fall in V_P (with rising temperature) can be fully counteracted by the rising voltage on base 2 resulting from the changing potential-divider action of R_S and the interbase resistance. The correct value of R_S is given by

R_S = 0.7R_BB/η.V_BB + (1-η)R_B/η

where R_B = external load resistor (if any) in series with base 1. The exact R_S value is not, however, of great importance in most applications.

In some circuits, R_B is wired between base 1 and ground, as shown in Fig. 2-b, either to control the discharge time of C or to give a positive output pulse during the flyback period. A negative-going pulse is also available, if needed, across R_S in the flyback period.

The unijunction transistor used in the circuits below is a 2N2646. Fig. 3 shows its base connections, while Table I lists its characteristics.

Similar to the oscillator shown previously, the pulse generator of Fig. 4 gives a large-amplitude, negative-going pulse across R4, and a positive-going pulse across R3. Both pulses have a voltage amplitude of about half the supply-line value, are of similar form, and are low impedance. The R4 pulse is suitable for triggering an SCR.

With the component values shown, the pulse width is constant at about 30 μ sec over the frequency range 25 to 3000 Hz (adjustable with R1 ). The pulse width and frequency range can be altered by changing the value of C1. Reducing the value of C1 by 10 (to 0.01 μF) reduces the pulse width by a factor of 10 (to about 3 μsec) and raises the frequency range by a decade (250 Hz to 30 kHz). C1 may be from about 100 pF to 1000 μF.

A sawtooth waveform is generated at the emitter, but is at a very high impedance level and is thus not readily available externally.

Wide-Range Sawtooth Generator

In this circuit (Fig. 5) the saw-tooth waveform from the emitter of Q1 is fed to emitter follower Q2. Hence the sawtooth appears at the Q2 emitter at an impedance of about 10,000 ohms. Output coupling may be made, either directly or via a coupling capacitor, to an external load of 10,000 ohms or greater, without adverse effects on the waveform or the operating frequency.

Frequency range is about 20 to 3000 Hz with the values shown, so the range is greater than 100 to 1 via R1. If a smaller range is required, reduce the value of R1. Operating frequency can be varied from less than one cycle per minute (0.017 Hz) to over 100 kHz by suitable choice of C1.

If an output impedance lower than 10,000 ohms is required, wire a second emitter follower with an emitter load of 2700 ohms to the emitter of Q2.

Linear Sawtooth (Time-Base) Generator

The sawtooth at the emitter of the basic UJT oscillator is exponential (nonlinear). In some applications - such as an oscilloscope time-base circuit - a perfectly linear sawtooth is required. This can be obtained by charging the main timing capacitor from a constant-current source, as in Fig. 6.

In this circuit, Q1 is wired as an emitter follower with emitter load R4, and feeds its collector current into the main timing capacitor (C1). The emitter current of Q1 - and thus the Q1 collector current and C1 charging current-is determined solely by the setting of R2. It is totally independent of Q1 collector voltage, C1 charging current is thus constant, and the capacitor therefore charges linearly up to the striking voltage of the unijunction. At this point Q2 fires and the capacitor discharges rapidly. Then the timing cycle starts over again.

The signal from Q2's emitter is fed to emitter follower Q3, giving a final linear sawtooth output at Q3's emitter at an impedance of about 10,000 ohms. This signal is suitable for feeding to the external time-base input of a scope. In this application, the flyback pulses from R6 can be taken via a high-voltage blocking capacitor and used for beam blanking.

This time-base oscillator can be synchronized with an external signal by feeding the external signal to base 2 of Q2, via C2. This signal, which should have a peak amplitude of 0.2 to 1 volt, effectively modulates the supply voltage, and thus the triggering point of Q2. This causes Q2 to fire in sync with the external signal.

C2 should be chosen to have a lower impedance than R5 at the sync signal frequency. It should also have a working voltage greater than the external voltage from which the signal is applied.

With the component values shown, the operating frequency can be varied over the approximate range 50 to 600 Hz using a 9-volt supply, or 70 to 600 Hz using a 12-volt supply. Operating frequency can be varied from a few cycles per minute to about 100 kHz by suitable choice of C1.

Analog/Digital Converter, Resistive

The circuit of Fig. 7 converts changes in light level, temperature or any other quantity that can be represented by a resistance, into changes in frequency. The resistive element (LDR, thermistor, etc.) is wired in parallel with R1, and so controls the charging time constant of C1, and thus the frequency of operation. A frequency range of 30 Hz to 3,700 kHz is available, the lower frequency being obtained with the variable element open circuit.

Output is taken across R4, and consists of a series of pulses. When fed to an earphone, these can be clearly heard, even at the lowest frequency.

The unit is of particular value in remote reading of such things as temperature, the output pulses being used to modulate a radio or similar link. At the receiver end of the link, the digital information can be converted back to analog via a simple frequency meter circuit.

Analog/Digital Converters, Voltage

These circuits have applications similar to those of the resistance-controlled circuit. However, their operating frequencies are controlled by voltage or by any quantities that can be represented by a voltage-photovoltaic cells, thermocouples, etc.

Figure 8 shows a basic shunt-controlled converter. Q1 shunts the main timing capacitor (C1 ) and so shunts off some of its charging current and affects the operating frequency. If zero voltage is fed to Q1's base, Q1 is cut off, and the circuit operates at maximum frequency (about 3.7 kHz). When a positive voltage is fed to Q1's base, the transistor is driven on and the operating frequency falls.

A restriction in this circuit is that, as Q1 is driven on, Q1's collector voltage falls; and when it falls to less than V_P, the circuit ceases to operate. The operating range is thus rather restricted, about 800 Hz minimum in this case.

The value of R4 is chosen, by trial and error, to suit the control voltage in use. Usually, it will have a value of a few hundred thousand ohms at potentials up to about 10 volts, and a few megohms at 100 volts.

Figure 9 shows a basic series-controlled converter. Here, the C1 charging current is controlled almost entirely by Q1. When Q1 is driven hard on (saturated) by a voltage applied to R4, the charging current is limited by R1, and the circuit operates at about 3.7 kHz. When zero voltage is applied to R4, Q1 is cut off, and C1 charges via R5, giving an operating frequency of about 30 Hz. Between these two extremes, the frequency can be smoothly controlled by the voltage applied to R4 (which controls the col-lector current of Q1 ). The value of R4 is found by trial and error, as in the case of Fig. 8.

In the circuits of Figs. 8 and 9, Q1 is cut off until a forward voltage of about 600 mV is applied to its base, so the operating frequency is not affected by voltages less than this. This difficulty can be overcome by applying a standing bias to Q1 base, as shown in Fig. 10. This modification allows use of input voltages right down to zero, or even reverse voltages.

Relay Time-Delay Circuits

These circuits enable time delays ranging approximately from 0.5 second to 8 minutes to be applied to conventional relays. That is, there is a delay from the moment at which the supply is connected to the moment at which the relay switches on. In Fig. 11, one set of normally closed relay contacts is wired in series with the positive supply line. Hence, power-supply current is fed to the unijunction circuit via these contacts. After a delay determined by the setting of R1 and the value of C1, the unijunction fires and drives RYI on. As RYI switches on, the supply to the UJT is broken by the relay contacts and the positive supply line is connected to RYI via R4, holding the relay on.

In this circuit, the relay must be a fast-acting low-voltage type with a coil resistance of less than 150 ohms. The supply-line potential should be at least 4 times the relay operating voltage. Also, the value of R4 should be chosen to keep the "on" current within limits when the relay is fed from the positive supply line.

One difficulty with the circuit of Fig. 11 is that the relay type must be carefully selected. This trouble is overcome in the circuit of Fig. 12. Here, the relay is connected in the collector of Q2, and is normally unactivated. When the UJT fires, a positive pulse is fed from R4 to the base of Q2 via D1, driving Q2 and RYI on, and rapidly charging C2. At the end of the pulse, the UJT switches off and D1 is reverse-biased, so C2 discharges into the base of Q2, holding the relay on for about 100 msec. Thus C2 is used as a pulse expander, and eliminates the need to use fast-acting relays.

As soon as RYI starts to close, the negative supply (ground) line to the UJT is broken via the relay contacts, but is still connected to Q2. Once RYI is fully closed, the supply is connected directly across RY1, holding it on, and cutting Q2 out of the circuit.

The relay in this circuit may be any type with a coil resistance greater than about 100 ohms, and with a working voltage of 6 to 18.

In the two relay circuits considered so far, the relays lock on and draw current indefinitely once they have been triggered. Fig. 13 shows a different arrangement of the circuit of Fig. 11, in which two relays are used.

This circuit's positive supply is connected via the normally closed contacts of RYI and the normally open contacts of RY2. The R Y2 contacts are shunted by pushbutton switch S1. As soon as this button is pressed, the supply is connected to the UJT and to RY2, which instantly switches on. When RY2 is activated, its contacts close, keeping the positive supply connected once S1 is released. After the preset time delay, the UJT fires, driving RY1 on and thus breaking the positive supply line to both the UJT and RY2, which switches off and thus completely breaks the supply to the circuit. The output of this circuit can be taken from the spare RY2 contacts.

When fed with a series of constant-width input pulses, the circuit of Fig. 14 produces a linear staircase output waveform that has a repetition frequency equal to some subdivision of the input frequency. Alternatively, if the input frequency is not constant, the circuit "counts" the number of input pulses, and gives an output pulse only after a predetermined number have been counted. Thus, the circuit can be used as a pulse counter, frequency divider, or step-voltage generator for use in such applications as transistor curve tracers.

Circuit operation is as follows: In the absence of an input pulse, Q1 is cut off, and Q2's base is shorted to the positive supply line via R3, so Q2 is cut off also, and no charging current flows into C2. If a constant-width positive-going input pulse is now fed to the circuit via C1, Q1 and Q2 will be driven on and C2 will start to charge via the collector of emitter-follower Q2; the charging current can be con-trolled via R6. C2 charges linearly, as long as Q2 is on, and since Q2 is on only for the fixed duration of the input pulse, the C2 voltage will increase by a fixed amount every time a pulse is applied to it.

In the absence of the pulse, there is no discharge path for C2, so the charge voltage stays on C2. The next pulse again increases the C2 charge by a fixed amount, until, after a predetermined number of pulses, C2 voltage reaches the trigger potential of Q3, and the UJT fires, discharging C2 and restarting the counting cycle.

If the input pulses are applied at a constant repetition frequency, the signal across C2 will be a linear staircase waveform, and an output pulse will be available across R8 every time the UJT fires. If the input frequency is not constant, the staircase will be nonlinear, but the R8 pulse will appear after a predetermined number of input pulses have been applied. Stable count or division ratios from 1 up to about 20 can be obtained.

Finding Division Ratio

Important: this circuit must be fed with constant-width input pulses if stable operation is to be obtained. Also, the width of the pulses must be small relative to the pulse repetition period. The value of C2 is determined by these considerations, and is best found by trial and error. Once a value of C2 has been selected, the division ratio can be varied over a range of about 10 to 1 via R6.

Now you know how the unijunction transistor works, and you've seen the first 11 projects. In the next article we'll show 9 additional applications.

Continued next month

Posted August 5, 2024