Half wave recifier Simulaiton problem

[Muffins]

Hi guys,

I'm using SImetrix to simulate a half wave rectifier. There is an RL load. I seem to be getting an unexpected result, after the diode stops conducting there is a ringing in the voltage.

I've attached a screenshot of my results. My question is why is there ringing in the voltage?

[Ian.M]

I'm not a SImetrix user, so YMMV.

You'd expect ringing if there was a significant capacitance somewhere in the circuit.   Take a careful look at your 'ideal' diode model.

However the decay and the varying period don't look right for 'real life' ringing, so I would suspect some sort of simulation artefact.  Try simplifying the sim by removing (shorting) R1, and try reducing the maximum timestep.

[Muffins]

Thank you.

I had a look through the simulation settings. I also changed the diode after playing around with the ideal one. I reduced the maximum time step and changed the integration mode from trapezoidal to Gear, though I'm not sure what that is exactly. I've attached the change settings and the new waveform

There is still a positive spike after the diode stops conducting but it has a much smaller time signature. When the spike occurs, what is causing it? Does the inductor give some of its energy to a capacitance in the circuit?

[Ian.M]

There's no energy in the inductor at zero current. The 1N4007 has a little bit of junction capacitance so I would expect to see some damped oscillation due to the step change in voltage across it when the diode cuts off.  Also, if the diode model is good enough you may be seeing the snap recovery effect, which even in a diode not optimised for the purpose can produce a very short reverse current pulse with harmonics extending into the HF or even VHF bands. See https://en.wikipedia.org/wiki/Step_recovery_diode

Zoom in on the spike, check if it is a damped oscillation, and if so find the period.  then see if its consistent with F=1/2*Pi*L*C, the inductance value and the junction capacitance with 300V reverse bias from the graph in the 1N4007 full datasheet.  N.B. many manufacturers only publish short form datasheets for their diode which don't include junction capacitance vs reverse bias graphs so yo'll probably have to hunt for a full one.  The OnSemi datasheet has the required graph, but it only goes up to 200V bias.  Extrapolating,  it looks like around 2pF would be realistic.  If the answer agrees within an order of magnitude, you are probably on the right track.

[Muffins]

Thank you for the direction here, I'm going to go look through everything you mentioned. 

[floobydust]

In real circuits and some simulators, I found it due to the diode's reverse-recovery spike, which is several usec for an 1N4007. This starts the inductor ringing and resonating with any parasitic capacitances.