[SOLVED] SCR triggering problem

[Sloth]

Hi,

I'm using a BGX15-1200CTT SCR and I'm triggering it with a op amp configuaration, giving 1,3 volts to the gate through a 10 ohm series resistor. Max current for the op amp is 30 mA. I don't have a clue about thyristors, but I thought that would trigger it. This is the only piece of data I found about the component:

https://www.americanmicrosemi.com/catalog/BGX15-1200CTT.html

I think I'm doing something wrong here. I also tried to trigger the thing with a 5 V lab power, using a 2k series resistor, but that only broke the SCR. I'm trying to have the SCR to short a DC voltage over it. Is there something I'm missing here?

 

[Prototyp_V1.0]

Hi.

It would make it easier to tell if you had a schematic of your circuit.

Are you sure about the pinouts on your scr?

 

[WimRFP]

The overview of characteristics says that it should trigger at 150mA maximum. This means it will trigger very likely with less current, but this may not be 30 mA.

To be sure you need a trigger current (at 25 degr) of at least 150 mA. According to NTE, they have the NTE57xx family that should be replacement part.

In general, make yourself familiar with thyristors before using them in serious applications. You can find good application info on manufacturers websites (NXP, ST, ONSEMI, etc).

 

[Sloth]

Here's something I drew to clarify the situation. The op amp current limit is 30 mA. The resulting output voltage from op amp output is 0,7V as it should be when thyristor should be on. For some reason this circuit works, when I only have ~12 volts in the cap (the voltage i'm trying to latch with the thyristor). I think I'm doing something very wrong. Often the control voltage in this kind of circuits is 10 volts or something. But according to the thyristor datasheet :

"I(GT) Max. (A): 150m
V(GT) Max.(V): 1.5"

So I guess I can't use that kind of a voltage?

 

[WimRFP]

You need more current.

As the opamp doesn't supply sufficient current, you need an emitter follower (NPN bipolar junction transistor), such as 2N2222 or equivalent transistor. Make sure you add a series resistor to limit the current to about 200mA. use a base resistor of about 200 Ohms to avoid current peaks from the opamp.

If your supply can't supply this current, you can get this current from an RC circuit. As you need to trigger the SCR until the anode current has risen sufficiently, an elco in the 100 uF range may be sufficient to store the required charge. Of course the actual value depends on your application. what is the supply voltage of your opamp?

Drive the output of the opamp to the supply so that the trigger current is determined by the emitter resistor.

 

[Sloth]

Thanks for the reply! I'm just afraid that I'll burn the thyristor again. The datasheet says:

I(GT) Max. (A): 150m

What does that mean? I thought it meant that the gate current should not exceed 150 mA. To get the 200 mA I could aswell forget the op amp and drive a mosfet with the microcontroller.

 

[WimRFP]

Thanks for the reply! I'm just afraid that I'll burn the thyristor again. The datasheet says:

I(GT) Max. (A): 150m

What does that mean? I thought it meant that the gate current should not exceed 150 mA. To get the 200 mA I could aswell forget the op amp and drive a mosfet with the microcontroller.

It means that the device will surely trigger at 150mA. It does not mean that the gate can handle maximum 150mA. The gate may accept significantly higher current for short periods (think of some A).

If your load results in high dI/dT at turn on, you may need higher trigger current (thing of 500mA for some 100us). In other words too low gate current can destroy the SCR.

If you need more detailled answers, you should provide us more info on what you are doing. Note that using a badly specified thyristor may leave you with some guessing.

Regarding the opamp, if you have a 5V output, you can drive the thyristor with an NPN transistor directly from the uc. I would not use mosfet as you lose the threshold voltage.

 

[Sloth]

I have a 5 V output in the microcontroller, but I also have 12 V on the board. I guess I should search a BJT, I'm just more familiar with FETs. I use the badly specified thyristor because I got a bunch of them for free .

I suspected that I(GT) Max could be the minimum triggering current, but it just sounds strange. Although I understand why the put it that way.

Thanks for clearing that out! I do have high dI/dT at latching. I didn't come to think that would be a problem, but ofcourse it is, as the "channel" isn't completely open at startup! Thanks alot! Instead of running too much current, I've broken those with too little current Shame on me.

 

[WimRFP]

Hello,

If you are familiar with mosfets, just use them. If your dI/dt is higher then in the spec, you very likely destroy the thyristor. If you have sufficient supply (or use the capacitor reservoir trick), you may drive the thyristor with about 0.5A for about 500 us.

You are correct, the problem with high dI/dt is that only a part of chip is conducting, hence the current density is too high. Is there any possibility that you can modify the circuit to change dI/dt, or that you use a saturable inductor?

with a saturable inductor you can keep dI/dt low for some time (until the core reaches saturation) and then the current rises based on your circuit parameters.

 

[Sloth]

Lowering the dI/dt might be hard, because of the characteristics of the whole circuit. I have used the thyristor successfully with some sort of a self built driver (I haven't got a clue why it worked), but the driver seized to work. So I think the thyristor should withstand the current when driven properly. And now that I understand a bit more about driving them, I might put in a bigger one.

 

[Sloth]

Would this work? 12V comes from car battery. Microcontroller can source 20 mA according to spec.

 

[WimRFP]

Hello,

Your circuit is an emitter follower, the output is about 0.7V below the input. So your uc goes to 5V, so the emitter goes to 5-0.7 = 4.3. The thyristor's drop is about 1.5V, so you have 4.3-1.5 = 2.8V across the resistor R1. To get about 400mA you need a resistor of about 2.8/0.4 = 7 Ohms (6.8 Ohms). We need to correct for the base resistor. When using hfe = 40 and R2 = 100, R1 must be reduced with 100/40 = 4 Ohms, so R1 must be 2.8 Ohms.

This solucion is not recommend as it depends on hfe.

Other option (you must feed from 12V in this case), is to set the current via the collector resistor. R1 = 0 and put a resistor in the collector of (12-2)/0.4 = 25 Ohms (22 Ohms).

Leave R2 = 270 Ohms. This provides a base current of about: (5-0.7-1.5)/270 = 10 mA. This results in a forced HFE = 0.4/0.01 = 40.

This last option gives the most stable drive current with low hfe dependency. Make sure that the resistor has sufficient pulsed power rating, as the pulsed power is about 10V*0.4A = 4W. Also make sure you limit the pulse duration as this consumes power.

 

[Sloth]

Hi,

Yea, I was a bit hasty with it. Thanks for the corrections. How about this one?

 

[WimRFP]

Hello,

I think this will work. I know you are not well into BJT, but I hope you understand how it works (to modify it if necessary).

 

[Sloth]

Hi,

Yeah, I do understand it. Your calculations made complete sense to me and I know how BJTs work too. I was just in a rush when I drew the picture.

I actually already tested the circuit, and it works!

Just a small BUT It sometimes seems to trigger at startup. I think I need to add some RC-delays somewhere in the HW. The uC startup sequence seems unstable. I have a relay that should cut the circuit to prevent problems like that. However the the capacitor (C1) charging circuit that I haven't drawn in the picture seems to short when this problem occurs I still need to figure this out. It might only need a small tweak.

 

[WimRFP]

Hello,

Do you have a real time sampling scope? Maybe you can trigger on the unintended event and look back to see whether the trigger pulses come from the uc, or from another source.

Some people add a resistor across gate and anode (for example 50 Ohms) to improve immunity against transients. Of course this doesn't help to avoid unintentional pulses that come from the uc.

You are sure that your supply (for the thyristor power circuit) is free from high dV/dt? Switching may result in fast transients.

 

[KerimF]

Is your triggering pulse long enough (or repeated enough) to reach the holding current at the output since there is an inductive load?
When SCR is turned on, is the output current limited to a proper value?

 

[Sloth]

The SCR seems to be working now. The current limit comes from the coil resistance. Coil inductance is pretty low, so the initial current is high, resulting in high dI/dT problems as WimRFP suggested. The problem now is my relay. I had a wearing feature in my program, but it should be fixed now. Also my battery charger seems to have broken, so I need to buy a new one before more testing. The relay might not be pulling well with low voltage. I hope that's the case.

 

[Sloth]

I bought a new battery charger, the old one had a "charging" LED on, because when it was connected to a battery, the battery would power the LED. However, the actual charger was broken and provided no current/voltage.

With the new charger everything is running smooth. Thank you for the help!

 

[Ogu Reginald]

Before making any design you should look at the datasheet of all the component you are to use. In this case I want you to check for the value of voltage that will trigger the gate of the SCR so that it will be in conduction. Thank you