3

Several times I have seen an SCR described as similar to a PNP and NPN transistor linked together. But any attempt to replicate the idea seems to fail. In the circuit below, for example, the LED is always on.

schematic

simulate this circuit – Schematic created using CircuitLab

What am I doing wrong?

Serodis
  • 139
  • 1
  • 4
  • 1
    This is similar to the "why can't I make a transistor from two diodes" query. the two bjt model are to help with functional understanding. What is missing is the electric field at the P-N junction that would exist in silicon –  Dec 08 '14 at 09:14
  • possible duplicate of [Transistor(BJT) as Two Diodes](http://electronics.stackexchange.com/questions/92131/transistorbjt-as-two-diodes) –  Dec 08 '14 at 09:15
  • 1
    @JonRB: No, that isn't the case here; the circuit should work. But it may be necessary to deal with leakage currents by putting a high-value resistor (10K - 100K) across the B-E junction of each of the transistors. – Dave Tweed Dec 08 '14 at 12:48
  • I've attempted to add 10k resistors, but that didn't correct the issue. I will see if I can get a couple 100ks to try. – Serodis Dec 08 '14 at 13:18

1 Answers1

7

This can be made to work as an SCR type circuit. There are couple of things to consider though.

One item, as pointed out by David Tweed, is the need to put bleeder resistors across the BE junctions of both transistors to keep leakage from instantly turning on the coupled PNP-NPN pair.

Another item is that a fast turn on rate of the supply voltage can couple through the transistor junction capacitance and cause the pair to come on. Real SCRs also have this same problem if the dV/dT is too fast.

Here is an LTSPICE simulation of the circuit working as expected. Note the gentle 100usec turn on of the power supply and the added BE resistors. Note I also replaced the switch with a transistor to simulate the switch.

enter image description here

enter image description here

Michael Karas
  • 56,889
  • 3
  • 70
  • 138