The equivalent circuit shown in Fig.4 has been developed to explain how the device works, and it is necessary to define the terms used in this explanation.

R_{BB} is known as the __interbase resistance__, and is the sum of R_{B1} and R_{B2}:

R_{BB} = R_{B1} + R_{B2} (1)

N.B. This is only true when the emitter is open circuit.

V_{RB1} is the voltage developed across R_{B1}; this is given by the voltage divider rule:

RSince the denominator of equation 2 is equal to equation 1, the former can be rewritten as:_{B1}V_{RB1}=(2) R_{B1}+ R_{B2}

RThe ratio R_{B1}VRB1 =x V_{BB}(3) R_{BB}

If an external voltage V_{e} is connected to the emitter, the equivalent circuit can be redrawn as shown in Fig.5.

If Ve is less than V_{RB1}, the diode is reverse biased and the circuit behaves as though the emitter was open circuit. If however V_{e} is increased so that it exceeds V_{RB1} by at least 0.7V, the diode becomes forward biased and emitter current Ie flows into the base 1 region. Because of this, the value of R_{B1} decreases. It has been suggested that this is due to the presence of additional charge carriers (holes) in the bar. Further increase in V_{e} causes the emitter current to increase which in turn reduces R_{B1} and this causes a further increase in current. This runaway effect is termed __regeneration__. The value of emitter voltage at which this occurs is known as
the peak voltage V_{P} and is given by: V_{P} =
_{AV}V_{BB} + V_{D} (4)

The characteristics of the UJT are illustrated by the graph of emitter voltage against emitter current (Fig.6).

As the emitter voltage is increased, the current is very small - just a few microamps. When the peak point is reached, the current rises rapidly, until at the valley point the device runs into saturation. At this point R_{B1} is at its lowest value, which is known as the __saturation resistance__.

The simplest application of a UJT is as a __relaxation oscillator__, which is defined as one in which a capacitor is charged gradually and then discharged rapidly. The basic circuit is shown in Fig.7; in the practical circuit of Fig.8 R3 limits the emitter current and provides a voltage pulse, while R2 provides a measure of temperature compensation. Fig. 9 shows the waveforms occurring at the emitter and base 1; the first is an approximation to a
sawtooth and the second is a pulse of short duration.

The operation of the circuit is as follows: C1 charges through R1 until the voltage across it reaches the peak point. The emitter current then rises rapidly, discharging C1 through the base 1 region and R3. The sudden rise of current through R3 produces the voltage pulse. When the current falls to I_{V} the UJT switches off and the cycle is repeated.

It can be shown that the time t between successive pulses is given by:

V_{BB}- V_{V}t + R1C lnsecs (5) N.B. R measured in Megaohms. C in µF. V_{BB}- V_{P}

The oscillator uses a 2N2646 UJT, which is the most readily available device, and is to operate from a 10V D.C. power supply.

From the relevant data sheet the specifications for the 2N2646 are:

VIt is important that the value of R1 is small enough to allow the emitter current to reach I_{EB2O}I_{E}(peak) P_{TOT}(max) I_{P}(max) I_{V}(max) Case style TO18 30V 2A 300mw 5µA 4ma 0.56 - 0.75

VFrom the specifications for the 2N2646 the average value of is 0.56 + 0.75/2 = 0.655. Substituting this value in equation (4) and assuming V_{BB}- V_{P}V_{BB}- V_{V}R1(max) =and R2(min) =I_{P}I_{V}

So R1(max) = 10 - 7.25/5µA = 550K, and if VIf we choose a value for R1 somewhere between these limits, e.g. lOK, the value of C can be calculated from equation (5)_{V}= approx VBB/10, R1(min) = 10 - 1/4mA = 2.25K.

If f = 1MHz, t = 1/f = 1msec. V_{BB} - V_{P} = 10 - 7.25 =
2.75 and V_{BB} - V_{V} = 10 - 1 = 9

tBecause of component and UJT tolerances it is sufficient in most circumstances to use an approximate formula: f = 1/CR, which assumes that is 0.63 - well within 5% of the average value for the 2N2646. In practice one would use a variable resistance (or a variable resistance in series with a fixed resistance) for R1 so that the frequency of oscillation could be adjusted to give the required value.Rearranging equation(5) to make C the subject: C = V_{BB}- V_{V}R1 lnV_{BB}- V_{P}0.001 so C == approx 84nF. 10^{4}ln (9/2.75)

R2 is not essential; if it is included, a value of 470 ohms is appropriate for the 2N2646. The value of R3 should be small in comparison with R_{BB}, with which it is in series, so as to prevent it from affecting the value of the peak voltage. A value of 47 ohms or thereabouts is satisfactory.

__Editor's notes__: The above design points are illustrated in the circuit of the enlarger timer which was described earlier this year in the April Newsletter. In that circuit the UJT provides clock pulses at 20Hz. R1 is a combination of a 47K variable and a 150K fixed resistance; R2 is omitted and R3 is 33 ohms. The timing capacitor has a value of 220nF. In
addition to the 2N2646, the component list for this timer also includes the TIS43 and the 2N4891. Most suppliers list only the 2N2646, but Maplin also include the TIS43. This device was used with a transistor constant current generator as the sawtooth oscillator in the timebase of the "Student's Oscilloscope" published in "Practical Wireless" in August 1973.

In his book "110 Semiconductor Projects for the Home Constructor" (2nd edition 1978), R.M.Marston gives twenty circuits for UJT projects using the 2N2646. These include pulse and sawtooth generators, analogue/digital converters, relay time delay circuits and frequency dividers. If any member would like to experiment with UJTs there is a good number of 2N2646 and TIS43 in Cyril's stock, and Ray Marston's book can be borrowed from me for 38p postage.

There is also a device called a programmable UJT - the BRY39 is an example so called because its parameters can be set by external components. It is a PNPN device, similar in some ways to a thyristor, and can be used in applications similar to those for the UJT. Perhaps we could have an article about this in a future Newsletter.