updated crowbar circuit

crowbar circuit schematic

I have a couple of issues with the simulation;

  1. The zener diode has Vz at 4.7V, it is not behaving as expected since it is reducing the supply voltage by 4.7 instead of setting it to that voltage.
  2. The SCR is not triggering and short circuiting the PSU.
  3. Is there a way to add a spice model to a thermal fuse in capture.
  4. Any recommendations on the values of the resistors and capacitor used?

Updated circuit: the current triggering the SCR is negative for some reason, in order to be triggered it should be 2mA.


I'm going to have to disagree here- there should be huge gate current into the thyristor and it should certainly trigger.

The gate current is very low and it is flowing out of the SCR!

I wonder if there is some issue with the off-grid connections- try redrawing it with snap to grid on, and put R3 as something like 0.1 ohms on the other side of the SCR. I suspect the SCR cathode is actually not connected.

You can replace V1 with a ramp source and do a time- domain simulation when you get the first part working.

You can add PSPICE models as you like- see the manuals or contact Orcad support- there are fuse models out there.

Edit: Aside from the SCR not being connected, there's something wonky with the model of the antediluvian 2N1595 (Edit':- actually I think it's because it's an incredibly wimpy old SCR and unsuited for this kind of service).

Here's a time-domain simulation with the R3 0.1R placed as I suggested and the 2N1595 replaced with an (also ancient) C233 20A SCR (Green trace is input voltage ramped to 10V over time, Red trace is load voltage). I used 265R for R4.

enter image description here

  • \$\begingroup\$ +1 for cathode not connected. For one thing, there's no pink dot on that path in the diagram. The 15.34 fA current could be due to an implicit 2**48 ohm resistor to ground that the simulator attaches to all nodes (or at least otherwise "floating" nodes) so that the simulation can run. \$\endgroup\$
    – Dave Tweed
    Jul 15 '14 at 12:22
  • \$\begingroup\$ You're conjecture regarding the conncetion problems might have some merit. The anode current of the SCR exactly equals the gate current while the cathode leakage is calculated in fempto amps. Seems similar to "epsilon" results that are added to make mathematical sense of certain conditions that may arise. \$\endgroup\$
    – sherrellbc
    Jul 15 '14 at 12:28
  • \$\begingroup\$ @DaveTweed I thought that was an LTSpice thing.. Orcad/PSPICE used to throw errors when there is a floating node, so it's a bit strange, but I have not upgraded our %3$#$# dongle-based Orcad license in some time. \$\endgroup\$ Jul 15 '14 at 12:28
  • \$\begingroup\$ @SpehroPefhany, what is the working principle of this circuit? The crowbar here is much more intuitive as it uses a gate-controlled zener to turn on a TRIAC. How do we know the voltage at the zener's anode when the supply is not sufficient to force the device into conduction mode? Is V_scr governed by the zener's leakage current until Vdc > V_zknee? I can see after the zener begins conducting then sufficient current with turn on the SCR, but I am unsure what happens at each node until that occurs. \$\endgroup\$
    – sherrellbc
    Jul 15 '14 at 14:05
  • \$\begingroup\$ In the circuit I simulated the gate (which looks like a diode to the cathode) had a 265 ohm resistor same as the OP had. The trigger current for the C233 is around 5-10mA, and the resistor takes a couple mA at one diode drop (the gate trigger voltage), so around 7-12mA through the zener, the SCR latches on and shorts the supply. There is a very narrow band of voltage over which the SCR will act as a kind of transistor, but only up to some 10's of mA anode current, above which the regenerative action causes it to latch on. That can be ignored for the present purposes. \$\endgroup\$ Jul 15 '14 at 14:32

Look again at your circuit. It's quite obvious what is happening.

You have an effective resistive divider between the source voltage and your load, made by R3 and R2. So it's clear from nodal analysis that the voltage on the anode of your thyristor is going to be exactly half of the source voltage. That's all that is happening here. A 1k resistor is not a good model for a fuse.

And with the thyristor anode at 5V and its gate at 6V, it's not going to trigger.

  • \$\begingroup\$ What would be a good value to model fuse? \$\endgroup\$
    – user49395
    Jul 15 '14 at 10:09
  • \$\begingroup\$ and by removing the 1k resistor, the anode voltage would be the same as the supply which is still higher than the gate voltage, I am not sure how you would have the Vg higher than V anode. \$\endgroup\$
    – user49395
    Jul 15 '14 at 10:20
  • \$\begingroup\$ @user49395: in a crowbar circuit, the fuse needs to be placed before the crowbar circuit, so the anode of the thyristor and cathode of the zener are on the same node. I'm sure orcad has some kind of reasonable model for a fuse, maybe a PTC? \$\endgroup\$
    – user36129
    Jul 15 '14 at 10:38
  • \$\begingroup\$ It does have a thermal fuse but without a spice model, I wasn't able to simulate. \$\endgroup\$
    – user49395
    Jul 15 '14 at 10:55

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.