Freewheel Diode Recomendation

[MikeMl]

... Here's an o'scope shot of almost 30v developed by switching a 12v signal across a coil (5v/div).

And it is totally meaningless without knowing the timebase, the inductance, resistance, type of diode, what is doing the switching, how well regulated the supply is for negative currents, etc, etc.

 

[Mike odom]

the circuit is driving the inductor at the resonance point of the circuit, with the duty cycle adjusted for critical damping of the waveform (one overshoot, one slight undershoot, then nulls out). The 12V is HIGHLY regulated and isolated so there are no negative currents. I wish I could share more of the circuit with you, but this was a very special project, and not even my 80W minuature radar replacement switchers I did for the FAA took this much work or design effort. Leave it to say that under certain conditions, inductors do crazy things. Tesla knew this, very well. You don't always need a secondary winding to get a large voltage jump. For instance, the reason you need the snubber in the first place...

It is just as meaningless to blindly pick a snubber at 2x without knowing what frequency, inductance, etc of the circuit being driven, but it's done every day by 'rule of thumb'.

 

[MikeMl]

the circuit is driving the inductor at the resonance point of the circuit,..

So it is not just a simple snubber diode around an inductive coil, is it?

I rest my case.

 

[Mike odom]

actually, yes it is. Or at least, a simulation of it would be. But every circuit has inherent capacitance and the driving frequency was adjusted to make use of this. But, for the same reason a PSpice model of a transmission line doesn't give an actual waveform, it models it as a series of LC circuits. It doesn't give a true waveform, but it gives the best one that can be simulated.

So, in sumation, the entire drive side of the circuit, was coil, diode, and darlington. There was also a polyfuse on the 12V line to limit current, but only operated if the darlington shorted out to keep it from dragging the 12V line down and affecting the other drive circuits.

I rest my case.

 

[ronv]

Please share a few more details of the circuit. I'm dying to see how it fools the simulation.

 

[Mike odom]

Please share a few more details of the circuit. I'm dying to see how it fools the simulation.

I really can't, not without violating my confidentiality agreement with my customer. I just sent the reponse to a response directed at me....
What I will (can) say is that it didn't 'fool' the simulator, the simulator just wasn't playing with a full deck of cards. It really is just a simple drive circuit, driving a switching transformer, not unlike an ignition coil, but unlike an ignition coil it was driven at a very specific frequency, not swept through a band. Also, it was wound for a very specific frequency, that we weren't even driving it at, we were driving it at (close to) a harmonic. So I slightly misspoke when I said it was driven at 'resonance', but that was the closest term available, and the frequency was adjusted slightly to hit the resonance of the total circuit. A few HZ off, and it doesn't even supply an output, which was one of the design specifications....

 

[ccurtis]

Even in simulation the inverse voltage across the diode can be hundreds of volts if equivalent capacitance is added to the inductor being switched and loss resistance in the windings is negligible. You have a tank circuit with a resonant frequency and if the Q is high, and the switch is stiff and sharp, the inverse voltage is very high. Try adding a parallel capacitor to simulate the interwinding capacitance without the series resistance (or small amount of resistance) of the windings. The many volts of reverse voltage will be obvious. But, generally, the Q of a relay coil is very low.