1
\$\begingroup\$

Is a short circuit that occurs for as long as an instance safe?

In the following schematic done from here, I have a silicon controlled rectifier that is used as a latch (the circuit simulation doesn't seem to be really accurate).

enter image description here

That white link thingy is a closed switch. In the diagram, the SCR is turned on and I have just closed the switch connected to ground. At this moment, a theoretical huge surge of current is going through the anode, out of the gate and through the switch to ground, which immediately turns off the SCR, so this stops and thus happens quicker than the blink of an eye.

In real life, I connected the gate to positive which turned the LED on and then connected the gate to negative, which turned the LED off with no problem.

However as shown in the previous image, it shows a large amount of current going through the anode of the SCR and out of the gate.

After looking around for a bit, I found an image (and added some edits) which gave me an idea as to where the simulator's result may have come from.

enter image description here

NOTE1: If gate is pulled to ground, the current would rather go through here than through the base of the NPN transistor out of the emitter to the cathode. This means no current is going out of the PNP transistor so the PNP transistor turns off and everything turns off. But is this safe?

Seems like a normal SCR can behave like a GTO thyristor.

The catch is, for a short moment when the gate is connected to ground before the SCR turns off, there would theoretically be a large surge of current going through the anode and out of the gate for a short period of time. Is this safe (in a way you can pass a safety regulation if selling as a commercial product, excluding the fact that the circuit allows you to enable both switches at the same time)? Can I do this to all thyristors? If not, what do I need to look for to ensure it's safe for certain types of thyristors to have their gates brought to ground?

I'm using a BT169D SCR, which is not labelled as a GTO thyristor but seemingly behaves like one.

\$\endgroup\$
  • \$\begingroup\$ What is white link thingy gate to ground - a switch ? Std SCRs do NOT act as a GTO - anode current MUST be reduced to below holding current level (v low) to turn off. You need to use std cct symbols and/or add enough labels / ext to allow what is being done to be obvious to others. \$\endgroup\$ – Russell McMahon Oct 26 '16 at 10:53
  • \$\begingroup\$ The funny thing here is when I put the gate to ground, the current stops going from A to K, but rather A to G. This puts the holding current to something very low, and thus the SCR turns off. \$\endgroup\$ – Bradman175 Oct 26 '16 at 11:24
  • \$\begingroup\$ You have not answered my question about the "link thingy". OF that is a switch then depending on switching timing you may be shorting the battery via the two switches. This will of course divert current from the SCR or TRIAC and allow it to turn off. If your device are all always turning off whenyou implement your circuit then EITHER your load current is under Ihold or you are doing something wrong and possibly damaging. Try answering the question above to get a better answer. Try using Ctrl-M in the question to open the circuit lab diagram drawer. \$\endgroup\$ – Russell McMahon Oct 26 '16 at 23:20
  • \$\begingroup\$ @RussellMcMahon That white link thingy is a closed switch. In the diagram, the SCR is turned on and I have just closed the switch connected to ground. At this moment, a theoretical huge surge of current is going through the anode, out of the gate and through the switch to ground, which immediately turns off the SCR, so this stops and thus happens quicker than the blink of an eye. \$\endgroup\$ – Bradman175 Oct 26 '16 at 23:55
2
\$\begingroup\$

To a degree... small-signal SCR potentially can be force commutated with a large negative gate current but that doesn't make them GTO's, neither is it advisable.

enter image description here

An SCR is basically an PNPN stack but there is a bit more to it, especially the specifics of doping.

enter image description here

The lightly doping and higher doping regions helps determine the effectiveness of "emitting minority carriers" & likewise to sweep them up.

In a GTO the gate is heavily doped to facilitate suppressing the injection from the cathode once a negative pulse is provided. This will cut off the flow of charge and turn the device off. There is also interdigitated finger structures in the CATHODE and GATE to help facilitate this.

The device you are looking at is a VERY VERY low power SCR on the grand scale of things 0.8A and it only needs a gate current of 50uA. If you provide a large enough negative gate pulse you are quite likely going to cause a very similar effect that is seen in GTO - cut off the flow of charge from the CATHODE to the intermediate N- layer.

However... this will not scale and once you enter power devices (not these small signal devices) you are more likely to burn out the gate before even experiencing any forced-commuted event. Likewise because of the lower doping concentration the equivelent turn-off speed will be alot slower and thus the switching speed will be drawn out.

\$\endgroup\$
  • \$\begingroup\$ Well looks like I will have to find a small GTO thyristor (which is virtually impossible). \$\endgroup\$ – Bradman175 Oct 28 '16 at 23:55
1
\$\begingroup\$

In DC circuits an SCR is not turned off by making the gate negative. You have to commute the load. This can be done by momentarily shorting anode to cathode or by opening the circuit with another device. Something is wrong if you can get a large reverse current through the gate because both the PN junctions it connects to will be reverse biased. In AC circuits SCRs are turned off during the negative half cycle. You need to get an actual GTO if you want to use a negative pulse to turn the device off.

\$\endgroup\$
  • \$\begingroup\$ Uh, well how come I could do it to my SCR? And my TRIACS? Basically all my thyristors? \$\endgroup\$ – Bradman175 Oct 26 '16 at 11:31
  • \$\begingroup\$ I double checked. Tried all my SCRs and TRIACS. Circuit works. \$\endgroup\$ – Bradman175 Oct 26 '16 at 12:01
  • \$\begingroup\$ I believe what you say is happening. Basically you are shorting the anode to cathode through the device which is not an accepted way to commute the SCR. I don't believe any device manufacturer would warrant your part for this type of operation. You could use your same turn off switch from the anode to the cathode. Doing it this way, prevents the large gate current that could potentially damage the device. \$\endgroup\$ – owg60 Oct 26 '16 at 12:22
  • 1
    \$\begingroup\$ Well you mentioning that the manufacture is not going to warrant the part should be added to your answer. And the fact that it may damage the thyristor but you probably need to explain why. \$\endgroup\$ – Bradman175 Oct 26 '16 at 20:59
  • \$\begingroup\$ Well you may get away with this in this circuit because you have such a small load. The first BT169D data sheet I come across says the peak gate current is 1A but doesn't give a time. I have references that say the time for this parameter is 20us. This specification has to do with things melting, either the bond wires or the semiconductor. How often you turn the LED on an off this way will determine the actual temperature reached. \$\endgroup\$ – owg60 Oct 28 '16 at 11:23

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.