Thyristors or silicon controlled rectifiers (SCRs) provide a convenient latching mechanism for delivering power to a load. However, once the thyristor is turned on, it cannot control the current and only external components can limit the current.On the other hand, the circuit in Figure a allows a constant current to be supplied to the load by triggering with a low level pulse; on the other hand, it can be used as a latched current sink. In the figure, TR1 and TR3 are connected in a "pseudo SCR" configuration. When the input terminal is low, all transistors are normally closed. Apply a pulse Vp with a sufficiently high amplitude to turn on transistor TR3. This also turns on TR1 and brings additional base current to TR3. Now the circuit is latched, and when the input terminal returns to a low level, TR1 and TR3 remain in the "on" state. However, it is different from the traditional SCR. The added transistor TR2 limits the emitter current of TR3 through the current sensing resistor R5 and the base-emitter voltage V VE2 of TR2. Assuming that the h FE (forward current transfer rate) of TR3 is large, then: I C3 ≌I E3 =V VE2 /R5 If TR1 also has a large hFE , then virtually all of IC3 flows through LED1. The LED has two functions. In addition to indicating that the circuit is locked, it also provides a constant reference voltage to the base of TR1. Therefore, no matter how much the power supply voltage (V S ) changes, the voltage across R3 is only VBE less than the forward voltage (V F ) of LED1 and remains constant. Therefore, TR1 forms part of the SCR structure and acts as a constant current source. If TR1 has a large h FE , its collector current is approximately: I C1 ≌I E1 =(V F -V BE1 )/R3 The appropriate value of R3 should be chosen to ensure that I C1 meets the base current requirement of TR3 and supplements the current shunted by TR2, R1 (or R2 mentioned below). The total current I SINK flowing through the load is: I SINK ≌I E1 +I C3 =[(V F -V BE1 )/R3]+(V VE2 /R5) Because all these parameters are constants at a given temperature, the absorption current remains constant even if VS changes considerably. The prototype circuit (R3 = 2.2kΩ, R5 = 62Ω, LED1 was tested with HLMP-1000 (3mm, red light), and the absorption current was 10.01mA when VS = 5V. When VS = 35V, the current increased to 10.05mA, which is equivalent to a 600% increase in VS and only a 0.4% increase in I SINK . As long as the input is at 0, R2 is not needed. However, if the input is left floating, R2 is necessary to reduce the noise at the base of TR3. R4 provides the same function for the base of TR1 and also serves to reduce the photovoltage generated by LED1. Otherwise, this photovoltage would latch up the circuit under bright light. Some LEDs can generate large photovoltages when exposed to sufficient light. In testing the prototype circuit, HLMP-1503 (3 mm, green light) was used for LED1. The experiment proved that if the LED is exposed to strong light, the circuit can act as a locked light detector. However, the effect will disappear completely if a 100kΩ resistor (R4) is connected in parallel across the LED. The results are naturally similar when the LED is isolated from the light source. Since most of the sink current flows through LED1, the circuit current is limited to less than or about 50mA. However, if other voltage references are used instead of LEDs, such as a low voltage Zener diode or two series diodes, the current can be increased. In this way, I SINK is only limited by the current rating of the diode and TR3. Once the circuit has locked, it can be reset by cycling power or by shorting the base of TR3 to 0 V. When the HLMP-1000 is used for LED1 (where V F ≌1.6 V), the circuit itself (minus the load) operates down to about 3 V. This allows ample power to put a lot of voltage on the load. The maximum voltage on the circuit is determined primarily by the VCE(MAX) ratings of TR1, TR3, and the power rating of TR3, which carries most of the current. The value of R1 depends largely on the trigger pulse amplitude. For pulses of a few volts, 10kΩ is sufficient, but larger values are acceptable. This circuit can be used to lock a constant current to an LED or relay coil. It also acts as a good "pulse grabber": the prototype circuit responds to low amplitude pulses (3V) as narrow as 100ns. Considering the circuit's ability to lock on very narrow pulses, it is necessary to connect capacitor C1 to the input to suppress noise to a certain extent. The circuit in part b of Figure 1 is a variation of circuit a, which uses an optocoupler instead of an LED. In addition to providing a stable voltage reference to the base of TR1, the optocoupler also provides an isolated signal indicating the locking of the current sink. |