Can I use SCR as a diode? (in a RCD snubber, IGBT bridge)

I have a 350A SCR rectifier (3 contacts). I am thinking of using it as a diode in a particular application (snubber). I wonder if I could use it as a simple diode if I supply proper voltage to its gate continuously, so that the gate is always open.

I got it. Thanks John. Also, an SCR is not a "fast recovery" diode, is it?

I don't have any specific numbers, but, I doubt it, since there is more internal structure that might hold change carriers.

^ ^ ^ Freudian slip? ;-)

I don't think SCRs turn off much faster than 10us, even highspeed inverter grade devices. I suppose that's about the same as "reverse recovery time" in a rectifier. Easy enough to measure, if you have a several-amps squarewave handy (heh, which you will in a few weeks or whatever!).

Ya know, SCRs might not be that bad an idea for your project Iggy, since such an inverter is a shorting-commutation type setup...just wire it for DC coupling, and hell you can use convienient pulse transformers for drive! Rugged as hell, and it stays latched in whatever polarity you left it, making drive circuits easy.

Tim

-- Deep Fryer: a very philosophical monk. Website:

As a point of information, I looked at the limited schematic for my Lincoln TIG and it uses SCRs both for rectification and for switching to generate DC or AC.

Yes, I think that you are right. I looked at my SCRs last night, and realized that each is a half bridge rectifier, so two of them make a full bridge rectifier. I will save them for a future project or maybe will just sell them on ebay. 300 amp rectifiers ought to be valuable to some people...

I understand up to this point...

And this I just do not understand, sorry, but it sounds interesting.

I *think* it could work that way. Using a full ("H") bridge of SCRs, you should be able to wire it so that, at any given time, two SCRs are ON. When another pair (on opposite corners) are triggered, their terminal voltage drops, while simultaneously, the pair that were already ON are still ON; as a result, supply voltage drops (to say, 5V) and the PSU inductor starts charging due to the short-circuit condition. By some nice capacitor (and possibly transformer) trickery, you can make this surge turn OFF the pair of SCRs that were ON. After the dust settles, you now have the opposite pair of SCRs ON, with the result of reversed voltage on your load.

This is regularly used in inverter welders, but those use a (relatively high frequency) transformer to connect the torch to the bridge. Obviously this only works with AC, so the inverter has to be continuously flipping for it to simply *survive*. With an H-bridge and no inductive load (arc instead of transformer), you can leave it in one polarity forever, as long as the idle current is high enough (check the minimum holding current rating and put a suitable resistor across the output) and it doesn't outright melt (a beefy heatsink, white goo, and your power-limited DC welder supply will take care of that).

I mentioned a half-bridge, push-pull, transformer coupled topology (as opposed to the DC coupled output you want). This is arranged as a center-tapped transformer with CT to +V supply, and an SCR hanging off each end. When one is on, +V is delivered through that half of the transformer primary. The other end swings to 2*(+V), by turns ratio. When the half cycle is up, the opposing SCR is triggered, causing that end to drop to approx. 0V. This would cause the other end to swing to 2 * +V, but it's still anchored to 0V by the SCR. Shit. So what is done is a capacitor is added between the SCRs. When the opposing SCR pulls down to 0V, this pulse is coupled through the capacitor, pulling down the voltage on the "ON" SCR, turning it off. After a short period of time, leakage inductance (a property of the transformer, imagine it as an inductor in series with each wire on the transformer) overcomes capacitance and the voltages establish themselves at the 0V / 2*(+V) levels you expect.

The leakage inductance provides a "squishiness" between the ends of the primary winding, allowing each SCR to turn off the opposing one, as the case may be. This circuit is in fact impossible with ideal components. You'd have to add a series inductor to each end of the primary winding for that to work. And go figure, that (plus some resistance and capacitance) is how you represent a real transformer in idealized terms.

Tim

-- Deep Fryer: a very philosophical monk. Website:

I *think* it could work that way. Using a full ("H") bridge of SCRs, you should be able to wire it so that, at any given time, two SCRs are ON. When another pair (on opposite corners) are triggered, their terminal voltage drops, while simultaneously, the pair that were already ON are still ON; as a result, supply voltage drops (to say, 5V) and the PSU inductor starts charging due to the short-circuit condition. By some nice capacitor (and possibly transformer) trickery, you can make this surge turn OFF the pair of SCRs that were ON. After the dust settles, you now have the opposite pair of SCRs ON, with the result of reversed voltage on your load.

This is regularly used in inverter welders, but those use a (relatively high frequency) transformer to connect the torch to the bridge. Obviously this only works with AC, so the inverter has to be continuously flipping for it to simply *survive*. With an H-bridge and no inductive load (arc instead of transformer), you can leave it in one polarity forever, as long as the idle current is high enough (check the minimum holding current rating and put a suitable resistor across the output) and it doesn't outright melt (a beefy heatsink, white goo, and your power-limited DC welder supply will take care of that).

I mentioned a half-bridge, push-pull, transformer coupled topology (as opposed to the DC coupled output you want). This is arranged as a center-tapped transformer with CT to +V supply, and an SCR hanging off each end. When one is on, +V is delivered through that half of the transformer primary. The other end swings to 2*(+V), by turns ratio. When the half cycle is up, the opposing SCR is triggered, causing that end to drop to approx. 0V. This would cause the other end to swing to 2 * +V, but it's still anchored to 0V by the SCR. Shit. So what is done is a capacitor is added between the SCRs. When the opposing SCR pulls down to 0V, this pulse is coupled through the capacitor, pulling down the voltage on the "ON" SCR, turning it off. After a short period of time, leakage inductance (a property of the transformer, imagine it as an inductor in series with each wire on the transformer) overcomes capacitance and the voltages establish themselves at the 0V / 2*(+V) levels you expect.

The leakage inductance provides a "squishiness" between the ends of the primary winding, allowing each SCR to turn off the opposing one, as the case may be. This circuit is in fact impossible with ideal components. You'd have to add a series inductor to each end of the primary winding for that to work. And go figure, that (plus some resistance and capacitance) is how you represent a real transformer in idealized terms.

Tim

-- Deep Fryer: a very philosophical monk. Website: