SCR firing troubles

Hi, I'm trying to get some help with my circuit. I've had people tell me that they won't help me because my project is dangerous. I would like to say that my safety is my own responsibility and any advice will not make this project more dangerous or put me in danger.

I've been having some problems firing my SCRs. I had two Teccor/Littelfuse S8040R SCRs (in parallel) hooked up to a capacitor charged to about 400v. The load (a coil) is connected to the + side of the capacitor, and the other end of the coil is connected to the anode of the SCR. The SCR cathode is connected to the capacitor minus. The gate is connected to a pushbutton switch in series with a 2.4v (two NimH AA's in series) supply, and the - of this is connected to the cathode. It would not fire when I pressed the button. I then disconnected the capacitor and measured the voltage between the gate and the cathode when the button was pressed, and it was only 1.44 volts! The SCR gates draw a lot of current which drags the battery voltage down a lot. When I hooked it up to my variable supply, the SCRs drew 3v, 3a! This makes sense to me because the SCR gate is supposed to be a diode, but then it is rated for 1.5v 30ma gate.

I then measured the resistance between the tab and anode pin of the SCR, and they're supposed to be connected, but overloaded my multimeter...

And the resistance between the SCR gates and cathodes is 82 ohms measured either way, and the diode voltage drop is around .1 volts, either way...

I'm very confused. Why won't the SCR fire? Why does the gate draw so much current? And why aren't the tab and anode pins connected? Thanks in advance, for any assistance at all.

Post a schematic. I get blurred vision reading a schematic in words.

Thanks for the reply. Here is a schematic, sorry about the quality. I think it is still readable, though.

Connecting a battery directly across the gate-cathode is the same as connecting a battery directly across a forward biased diode. You have to limit the current in such a connection. Look at the trigger current, and chose an appropriate resistor. You may have damaged your devices.

But when I connect a small (10 or 15 ohm resistor in series) the gate voltage drops down to about .8v, which is well below the triggering voltage, and I assume is the gate forward voltage.

fizzit said:

But when I connect a small (10 or 15 ohm resistor in series) the gate voltage drops down to about .8v, which is well below the triggering voltage, and I assume is the gate forward voltage.

Click to expand...

What/where is the spec for this? .8v is close to a diode drop you should see across the gate-cathode, but the trigger current will be in mA, so a larger resistor should be used.

Here is the datasheet.
https://www.electro-tech-online.com/custompdfs/2010/07/E6SCR.pdf
The trigger current is about 40ma, so since I have two in parallel that would mean 80ma total, and if the voltage drop is .8v then that's 1.6v over the 10ohm resistor, which means, 160ma... so shouldn't that be enough?

The gates should be connected in parallel, each with it's own resistor. (2.4 - 1.5)/.04 = 22.5 ohms.

OK! Thanks! I'll try that. Hopefully I didn't screw up the chips too much, but the tabs not showing connection would indicate otherwise. And I cut the anode pins to be really short because I assumed I could just connect to the tab. D'oh.

Oops, my calculation above was for Max Igt, try this:

(2.4-1.5)/20mA = 45Ohms ( or close standard resistor )

gate voltage is still .87v when triggered with no load...

Don't worry about that. The spec is for "MAX" trigger voltage. .87 is within spec. Hopefully you haven't damaged your devices when you slammed the voltage across the gate-cathode without a resistor. Try a different device if you have one.

Alright, sounds good. Unfortunately these are the only devices I have and it'll take a week for more to arrive... Is the tab supposed to show open circuit with the middle pin, even though they are supposed to be connected? Or did I somehow fry the connection between them by overloading the gate?

I wouldn't think the open circuit should exist. Try testing the devices with a voltage and load, say 12V and a 12V lamp. See if the lamp is switched on when the device is triggered. Or else use a resistor and measure the voltage across it before and after triggering.

I just did. It doesn't appear that either SCR fired or is ever going to fire :/ I just don't understand this. The first time I tried it, I hooked up the SCRs with a 2.4v source and 10 ohm resistor and triggered it. It didn't trigger. Now I've tried a lot of different things, and they still don't! I don't understand how I could have broken them from the very start, but I also don't understand how both could have arrived defective.

If you used a 10 ohm reisistor, that would give:

(2.4-1.5)/10 = 90mA

Or, using your measured Vt:

(2.4-.87)/10 = 153mA.

Either way, you violated Max It.

Ah. So if you go over, it won't fire? But the datasheet also lists Igm as 3.5a, which I know I never went over.

If you go over, you might damage the device. IDK if your devices were ever good, but you should never exceed maximum ratings.

OK! So I got new SCR's and installed them. Finally. The same SCR as before. I wired two them up with all pins connected together, and used the resistor size you suggested. I charged the capacitors, and fired the SCRs. It worked fine. Then, I began to charge to 400v, but at around 330v, the SCRs (or one of them) triggered on their own. I attempted to recharge the bank, but it now appears that one or both of the SCRs have become a short, with around 5 ohms resistance. What am I doing wrong?

Edit: I just desoldered both SCRs. It appears that both of them failed to some extent. One has no resistance between cathode and anode and the other has a resistance of around 200 ohms. The gate shows as being connected to the anode by a few hundred ohms, which it does not on the other SCRs that I have not used and ordered as extras this time around.

You need a pull down resistor to keep the gates biased to a off condition when the switch is open. A 1k should work just fine.