Electrical Engineering › Section Q · Thyristors, operational amplifiers, and power conversion › Topic Q.3
The UJT, the SCS, and field-effect-controlled thyristors
A thyristor is only as good as the circuit that fires it. This topic covers the unijunction devices that generate trigger pulses, the silicon-controlled switch that can be turned off from a gate, and the MOSFET-driven thyristors that need almost no drive at all.
Trigger devices and gate control
SCRs and TRIACs need a clean, repeatable pulse at a chosen moment. The unijunction transistor (UJT) is not a thyristor, but its sudden switching makes it a classic pulse source; its four-layer cousin, the programmable unijunction transistor (PUT), does the same job with adjustable settings. Two further devices attack a different weakness of the SCR, that its gate can turn it on but not off. The silicon-controlled switch (SCS) adds a second gate, and field-effect-controlled thyristors put MOSFETs in charge of triggering, and in one design of turning off as well.
Inside the UJT
A UJT is a bar of lightly doped N-type silicon with a contact at each end, base 1 (B1) and base 2 (B2), and a P-type emitter (E) joined to the bar between them, nearer B2. With the emitter open, the bar is just a resistance, the interbase resistance \(R_{BB}\), typically a few kilohms up to about 10 kΩ. The emitter divides it into \(R_{B1}\) and \(R_{B2}\), and the intrinsic standoff ratio
typically lies between about 0.5 and 0.8. With a supply \(V_{BB}\) from B2 to B1, the bar's divider holds the junction point at \(\eta V_{BB}\) above B1.
Peak point, negative resistance, valley point
The emitter-to-B1 path is a PN junction in series with \(R_{B1}\), so the emitter conducts only when its voltage exceeds the peak voltage
Below \(V_P\) only a tiny current flows; the peak current \(I_P\) is a few microamps. Past the peak, holes injected into the bar lower \(R_{B1}\), so current rises while voltage falls: a negative-resistance region. It ends at the valley point, voltage \(V_V\) and valley current \(I_V\), typically a few milliamps. If the emitter current falls below \(I_V\), the junction turns off and the emitter returns to its high-resistance state. That switching between two states is what makes oscillators possible.
The UJT relaxation oscillator
Connect a resistor \(R_E\) from the supply to the emitter and a capacitor \(C_E\) from the emitter to ground, with B1 to ground through a small resistor, often tens of ohms, and B2 to the supply. The capacitor charges through \(R_E\) until it reaches \(V_P\), then dumps its charge through the emitter and B1, producing a sharp voltage pulse across the B1 resistor, the pulse used to fire an SCR. Below the valley point the UJT turns off and charging begins again. Ignoring the diode drop and valley voltage, the period is about
with no dependence on the supply voltage.
Worked example: timing a UJT oscillator
Take \(V_{BB} = 12\,\text{V}\), \(\eta = 0.65\), \(R_E = 47\,\text{k}\Omega\), \(C_E = 0.1\,\mu\text{F}\), and assume \(V_V = 1.0\,\text{V}\), \(I_P = 5\,\mu\text{A}\), \(I_V = 4\,\text{mA}\). 1. Peak voltage: \(V_P = 0.7 + 0.65(12) = 8.5\,\text{V}\). 2. Charging from 1.0 V to 8.5 V: \(T = 4.7\,\text{ms} \times \ln(11/3.5) = 5.38\,\text{ms}\), about 186 Hz. 3. The quick formula gives \(4.7\,\text{ms} \times \ln(1/0.35) = 4.93\,\text{ms}\), within 10%. 4. Limits on \(R_E\):
so 2.75 kΩ < \(R_E\) < 700 kΩ. Check: 47 kΩ lies safely inside.
Why the resistor has limits
The limits come from the two ends of the negative-resistance region. The upper limit: \(R_E\) must pass at least the peak current \(I_P\) when the emitter sits at \(V_P\), or the capacitor never quite reaches the peak and the UJT never fires. The lower limit: after the capacitor has discharged, \(R_E\) on its own must not be able to supply the valley current. If it could, the supply would hold the UJT in conduction at the valley point, it would never turn off, and the oscillator would stall. Choosing \(R_E\) well inside the range also allows for spread between parts.
The programmable UJT
Discrete UJTs are now hard to find, but the PUT has outlasted them. It is a four-layer PNPN thyristor whose gate is the N layer next to the anode, an anode gate. Two external resistors set the gate voltage: \(R_1\) from gate to ground and \(R_2\) from gate to the supply, so \(V_S = \eta V_{BB}\) with \(\eta = R_1/(R_1 + R_2)\). The PUT fires when the anode rises a small offset \(V_T\) above the gate, \(V_P = V_T + V_S\), typically a few tenths of a volt to about a volt. Changing the resistors programs the trigger point, which a UJT's bar fixes; anode, cathode and gate play the roles of E, B1 and B2.
Worked example: a PUT oscillator
A 15 V supply feeds a PUT gate divider with \(R_1 = 15\,\text{k}\Omega\) and \(R_2 = 10\,\text{k}\Omega\). A resistor \(R\) charges 0.1 µF on the anode. Assume \(V_T = 0.5\,\text{V}\), \(I_P = 5\,\mu\text{A}\), \(I_V = 100\,\mu\text{A}\), and a valley voltage near 0 V. 1. \(\eta = 15/25 = 0.60\), so \(V_S = 9.0\,\text{V}\) and \(V_P = 9.5\,\text{V}\). 2. Upper limit: \(R < (15 - 9.5)/(5\,\mu\text{A}) = 1.1\,\text{M}\Omega\). 3. Lower limit: \(R > 15/(100\,\mu\text{A}) = 150\,\text{k}\Omega\). 4. With \(R = 330\,\text{k}\Omega\): \(T = 33\,\text{ms} \times \ln(15/5.5) = 33.1\,\text{ms}\), about 30 Hz. Check: 330 kΩ sits inside both limits.
The silicon-controlled switch
An SCS is an SCR with both inner layers brought out: a cathode gate on the NPN transistor's base, as in an SCR, and an anode gate on the PNP's base. Either gate can fire it: a positive pulse on the cathode gate or a negative pulse on the anode gate, relative to the nearby main terminal. Because SCSs are small, low-current parts, the gates keep enough influence to turn the device off again with the opposite pulses, or by briefly shorting the anode to the cathode, which starves the PNP of emitter current. Turning a thyristor off while its current is still above the holding current is forced commutation; waiting for the current to stop is natural commutation.
MOSFET-controlled thyristors
A MOS-gated thyristor places a MOSFET where the gate current would enter, so the trigger draws almost no current. Since ordinary SCRs are already easy to fire, the gain is modest, and nothing can now pull the gate to force turn-off: only low-current dropout stops it. The MOS-controlled thyristor (MCT) goes further with two MOSFETs on one gate. One gate polarity turns on the MOSFET that fires the thyristor; the opposite polarity turns on a second MOSFET that shorts the NPN's base to its emitter, forcing turn-off. With zero gate voltage it keeps its last state. In effect it is a latching relative of the IGBT, described in Insulated-gate FETs (MOSFETs) and IGBTs.
Common mistakes to avoid
First, forgetting the diode drop: a UJT fires at \(0.7 + \eta V_{BB}\), not \(\eta V_{BB}\). Second, defining η with the wrong resistor: it is the share of the bar, or of the PUT divider, between the junction and the B1 or ground end. Third, choosing \(R_E\) too small: the oscillator latches at the valley point and stops. Fourth, expecting an SCR gate to turn the device off; that needs an SCS, GTO or MCT. Fifth, assuming an MCT or SCS resets when the gate returns to zero: it holds its last state.
Key takeaways
A UJT is an N-type bar with an emitter junction dividing it in the ratio η; it fires at V_P = 0.7 + ηV_BB and resets below its valley current. An R-C network on the emitter makes a relaxation oscillator with T ≈ RC ln(1/(1 − η)), and the charging resistor must lie between (V_BB − V_V)/I_V and (V_BB − V_P)/I_P. A PUT is a four-layer device whose trigger point two external resistors program. An SCS has two gates and can be force-commutated. MOS-gated thyristors fire with negligible gate current, and an MCT turns on and off from one gate, holding its state at zero gate voltage.
Flashcards
Flashcard
Unijunction transistor: an N-type bar with bases B1 and B2 and a P-type emitter between them
Test your understanding
Question 1 of 5
A UJT with an intrinsic standoff ratio of 0.60 runs from a 20 V interbase supply. At about what emitter voltage does it fire?
Question 2 of 5
A PUT has R1 = 22 kilohms from gate to ground and R2 = 10 kilohms from gate to a 12 V supply, with an offset V_T of 0.6 V. What is its firing (peak) voltage?
Question 3 of 5
A UJT relaxation oscillator stops oscillating after its first pulse when its charging resistor is made very small. Why?
Question 4 of 5
What can a silicon-controlled switch (SCS) do that an ordinary SCR generally cannot?
Question 5 of 5
An MCT is latched on and its gate voltage is then returned to zero. What happens?
