I (incorrectly*) modeled this simple answer in LTspice IV and am seeing seemingly random instantenous voltage spikes up to +/-1kV on the output of the depletion nMOSFET. Are these real?

If so, are they significant, and if so what standard practices are missing in this design to control them?

Spikey voltage in LTSpice

* Here is the correct model and output, which exhibits no voltage spikes. (But I believe the question on the original circuit is still valid.)

  • \$\begingroup\$ Which diodes are you using? \$\endgroup\$ – W5VO Mar 11 '16 at 19:11
  • \$\begingroup\$ Also, you didn't make the schematic correctly - your drain and source on M1 are swapped. \$\endgroup\$ – W5VO Mar 11 '16 at 19:15
  • \$\begingroup\$ You are using the "ideal diode" model which is basically useless. \$\endgroup\$ – PlasmaHH Mar 11 '16 at 20:00
  • \$\begingroup\$ @W5VO - Oh ... well, if I do it correctly then it works properly, and the transients go away! The diodes are default -- apparently no breakdown voltage or forward current. \$\endgroup\$ – feetwet Mar 11 '16 at 20:00

The significance that I usually associate depends on the magnitude of the voltage. I think a good measure of how suspicious/skeptical you should be is how many multiples of the expected value you see (and how many ideal components you use).

  • Circuit behavior matches expected results: 3/10
  • Output voltage exceeds supply voltage by 50%: 4/10
  • Output voltage exceeds supply voltage by 110%: 5/10
  • Output voltage exceeds supply voltage by 500%: 8/10
  • Output voltage is measured in kV or MV: 10/10
  • You have to look up the SI prefix for your voltage or current: 11/10

An extremely relevant quote that applies to modeling is

"Everything should be made as simple as possible, but not simpler."

With ideal circuit elements, it is easy to omit a lot of the parasitic effects in a circuit, such as capacitance, inductance and resistance. Ideal models may not switch operating modes smoothly, and abrupt transitions/discontinuities are also sources of more weird solver behavior. Another thing not handled well is multiple high impedance devices in a voltage divider. Another gotcha is operating a model outside its characterized region.

Beware of simulating circuits you don't understand, as you may not understand the expected behavior or be able to spot when it's misbehaving.


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.