simulate this circuit – Schematic created using CircuitLab

I built this circuit (real not simulated.) With a 3V DC source, the output voltage is about 37V using a 1:14 transformer. I think this is a "self-oscillating" circuit.

When I try to simulate the circuit in Falstad or EveryCircuit, the circuit does not appear to oscillate and so the transformer and booster circuit don't work.

I don't understand how to use CircuitLab, so I didn't test whether that would successfully simulate the oscillation.

I built the circuit in CircuitLab in order to insert it here.

Can anyone tell me why simulators have trouble with this circuit? Will CircuitLab successfully simulate this? If so I'll learn how to use it. I'm a beginner so I might not have described this correctly.

Here's a photo of the Snap Circuits part I'm using as the transformer: enter image description here

  • \$\begingroup\$ can you provide a link to the falstaad simulation? \$\endgroup\$
    – FrancoVS
    May 9, 2022 at 15:47
  • 5
    \$\begingroup\$ Have you tried adding a little AC noise to the source. It could be failing the way bistables sometimes fail in simulation, because the initial conditions in the simulator coincide with a metastable state which is exceedingly unlikely in the real world. \$\endgroup\$
    – Dannie
    May 9, 2022 at 15:49
  • \$\begingroup\$ The inductance of the transformer affects the oscillation - try to match that in your sim. \$\endgroup\$ May 9, 2022 at 16:01
  • 1
    \$\begingroup\$ This answer might be useful: electronics.stackexchange.com/questions/350823/… \$\endgroup\$ May 9, 2022 at 16:03
  • \$\begingroup\$ @FrancoVS Here's a link to the falstad simulation: tinyurl.com/y3gr2uxg \$\endgroup\$ May 9, 2022 at 18:13

3 Answers 3


It's not the simulator that's isn't working. The challenge is to accurately model your circuit: not just its connections, but also the circuit elements. Even then, simulations can bomb out when they encounter cases their internal modeling can’t handle, as we’ll see below.

The thing you're making has a complex circuit element in it: a transformer. It will have its own characteristics besides its turns ratio. These include, among others: inductance, coil resistance, and parasitic capacitance.

If you know your transformer's characteristics you can add these to your simulation.

With your 'audio' transformer parameters (3.5mH primary, 1:14, reference example here), here's what we get (simulate it here):

enter image description here

There's these differences between your diagram and this sim:

  • Inductance added in
  • Coil resistances added (see the datasheet reference) but you could also measure these with a voltmeter to confirm
  • Primary is flipped to make the feedback work
  • Reverse diode on the transistor base

Why the diode? The sim fails because the transformer flyback when Q1 shuts off makes a huge negative voltage spike, which causes the sim to fail to converge.

In the real world, probably the transistor is getting fried with excessive reverse Vbe (over 600V in the sim, until it stops.) You can remove the diode in the sim and see for yourself.

The LED isn't very happy either. It's also seeing large reverse voltage (-38V with the diode, much more without.) It won't live long either.

Maybe Snap wants to sell you some components when they die?

  • \$\begingroup\$ Are the two secondaries each 140 turns ? \$\endgroup\$
    – Antonio51
    May 10, 2022 at 10:22
  • \$\begingroup\$ 70 + 70, split in the middle. \$\endgroup\$ May 10, 2022 at 13:07
  • \$\begingroup\$ Why do the transformer primary terminals need to be flipped in the simulator when they are not flipped in the actual circuit? \$\endgroup\$ May 10, 2022 at 13:42
  • \$\begingroup\$ I would bet that they are flipped in the actual circuit too. Post a photo maybe? But that’s just one of several reasons why your sim isn’t working. \$\endgroup\$ May 10, 2022 at 13:56
  • 1
    \$\begingroup\$ In the same Falstad circuit, if I flip the primary connections, the simulation works: link \$\endgroup\$ May 10, 2022 at 14:50

The transformer winding in the collector is the wrong way round. You need to swap the two connections such that when the collector goes towards ground, the base goes positive.

The dots on the windings indicate the polarity. If a terminal with a dot is connected to the collector, the base needs to connect to a terminal opposite to the dotted terminal.

Simulating oscillators can be tricky as there is no noise to start the oscillation. You may need to inject a pulse or experiment with the start-up conditions to achieve oscillation even if the circuit is correct.

Also with a 1:14 transformer the breakdown voltages on the LED and the transistor base-emitter junction will be exceeded.

  • \$\begingroup\$ By reversing the transformer connection as discussed, and changing the voltage source to a step input, it oscillates nicely: circuitlab.com/circuit/ckpz9arm4q69/converter Feel free to paste this into your answer if you like. Just run the time-domain simulation with "skip initial'. \$\endgroup\$
    – Theodore
    May 10, 2022 at 19:06
  • \$\begingroup\$ To clarify: It oscillates and generates the expected voltage at its output. How accurately it models the physical circuit is another matter. \$\endgroup\$
    – Theodore
    May 10, 2022 at 19:14
  • \$\begingroup\$ Once I got the transformer polarity right (out-of-phase) by flipping the connections to the primary coil, the simulation worked. \$\endgroup\$ May 11, 2022 at 13:07

These sorts of simple self oscillating schemes rely on transformer saturation to define the on time by volt seconds applied to the winding. This is a close relative to the blocking oscillator. You must work on your transformer model to include saturation.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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