Finding L & C for Circuit Protection: dv/dt & di/dt Values

[Johnie]

I'm trying to use this protection cct. What I do know is that (dv/dt)max = 50 V/us and (di/dt) = 10A/us (us = microseconds)
R = 1 ohm, R = 5 ohm, Vcc = 120 V
Can anyone give me an idea how to find the right L and C ?

Thanks

 

[turin]

What's a cct? Is that a zener? What is the application of this circuit/what are you trying to protect and where is it connected?

 

[Johnie]

snubber

cct = circuit

That is a thyristor (scr) (not a zener)

A snubber is used mainly to protect a device from large rates of anode-to-cathode voltage and anode current. If dv/dt for the thyristor is too large, the device will begin to conduct without a gate signal present. If di/dt is too large during turn-on, localized heating will result from the high current density in the region of the gate connection as the current spreads out over the whole junction.

This circuit can be applied to a lot of things -> ie. a buck converter

I know that I can follow the loop from Vcc to ground and determine what (dv/dt)max is using Laplace Transform. To when I'm determining C and L, do have have to pick a value for one (C or L) and then determine the other one ? and then check to ensure di/dt is ok ?

 

[turin]

OK, so we're talking about across and through the thyristor, I take it. I am considering the circuit now.

I know that I can follow the loop from Vcc to ground and determine what (dv/dt)max is using Laplace Transform.

Why (and how) would you use a Laplace transform to determine a specification that is already given? Can you explain what you mean in more detail?



Can the thyristor be modeled as having a threshold voltage and on resistance? Does is have parasitic capacitance (I'm assuming so according to your previous description) and inline inductance? How complicated of a model do you want to use? These issues seem like they would be important. For a first order approximation, I would just replace the thyristor by some typical on resistance value (and perhaps the parasitic cap) and then ensure the ratings are not breached for this resistor.

 

[Johnie]

snubber

I thought I'd look at the dv/dt ratio when the thyristor is off (immediately after). In this case I can follow the loop from Vcc to ground and use Laplace to determine V(output) -> dV(output)/dt. From this I was thinking that I would pick a value for either L or C and calculate the other one that I didn't choose. Then I'd do the same thing for the current (di/dt), when the thyristor is turning on and make sure that these values of L and C satisfy this parameter.

Another thing that I was thinking was that when the thyristor turns off...the inductor and the capacitor might go into resonance ?

 

[turin]

OK, that sounds pretty good. But it seems like you still need some sort of a model for your thyristor, because you're going to need to know what the state of the circuit is immediately before you "open" the path through the thyristor.

Consider the on-steady-state and a strictly on-resistance model for the thyristor:
There will be an initial current through the inductor given by I_L,0 = 120V/(R_thyristor + 1Ω).
There will be an initial voltage across the capacitor given by V_C,0 = I_L,0R_thyristor.
When the thyristor is switched off, this current I_L,0 must go into the cap.

I would start my analysis there, and then increase the order of approximation by including a parasitic cap on the thyristor if you get bored. I imagine that should be good enough, since these values don't really seem to be critical operational values. In fact, if your #1 concern is that the max specs must absolutely not be breached, then just get the largest L and largest C that you can.

About the resonance, I imagine that's why there's a res in parallel with the cap.