Triggering multiple SCRs in sequence

Hello all,

I’m currently in the process of building a 3-stage gauss rifle. I designed and put together a protoboard circuit and it works with one issue I just cannot seem to solve.

I am using 3 high-voltage rated SCRs as my switches for each of the 3 stages, which each have their own respective capacitor banks. These SCRs are triggered via digital output pins from my arduino. I’ve had no problem with a single-stage using only 1 SCR connected to 1 digital pin. However, once I add additional stages, attempting to trigger each of the stages in sequence causes them to all trigger at the same instant when the first SCR is triggered rather than independently in a delayed sequence.

At first I thought maybe the separating diodes I placed between capacitor banks were not strong enough to keep one bank from discharging into the previous one, but I’ve confirmed that it is not a problem with them, as charging all banks and triggering only 1 SCR with the others not connected to their respective trigger pins will only discharge the 1st stage capacitor bank.

Strangely, charging all 3 capacitor banks with the 2nd and 3rd stage trigger pins disconnected will allow the first stage to fire without causing the 2nd and 3rd to fire, and I can then reconnect the 2nd stage and trigger it and that will fire the 2nd stage with the correct timing, as is the same if I connect the 3rd pin and fire that one after the 2nd. BUT leaving the pins connected while triggering the 1st stage will cause all three to trigger, even if I am not setting the other pins to HIGH.

I’ve confirmed there are no shorts between the pins and SCR gates with a multimeter continuity test.

What could possibly be causing this? I’ve attached my schematic and code below.

I would appreciate any feedback and suggestions, thank you everyone!

EDIT: Corrected schm with SCRs rather than MOSFETs:

int buttonPin = 11;
int scrPin = 7;
int scrPin2 = 8;
int scrPin3 = 9;

void setup() {
  Serial.begin(9600);
  
  pinMode(buttonPin, INPUT_PULLUP);

  pinMode(scrPin, OUTPUT);
  pinMode(scrPin2, OUTPUT);
  pinMode(scrPin3, OUTPUT);
}

void loop() {
  if(digitalRead(buttonPin) == HIGH){
    Serial.println(digitalRead(buttonPin));
    digitalWrite(scrPin, HIGH);
    delay(500);
    digitalWrite(scrPin, LOW);
    delay(500);
    digitalWrite(scrPin2, HIGH);
    delay(500);
    digitalWrite(scrPin2, LOW);
    delay(500);
    digitalWrite(scrPin3, HIGH);
    delay(500);
    digitalWrite(scrPin3, LOW);
    delay(500);
  }
  else {
  digitalWrite(scrPin, LOW);
  digitalWrite(scrPin2, LOW);
  digitalWrite(scrPin3, LOW);
  }
}

That schematic is completely wrong, you show the coil current coming out of the gate of a FET.

However, once I add additional stages, attempting to trigger each of the stages in sequence causes them to all trigger at the same instant when the first SCR is triggered rather than independently in a delayed sequence.

This is because you have no supply decoupling between each SCR circuit and the interference generated by one of them triggering, is setting off the other two.
You could also do with snubber circuits in the load of the SCR.

And as John says that schematic is all wrong, where is the gate trigger of the SCR? In fact whre is the SCR?

I switched to SCRs shortly after making the schematic and the actual circuit uses those. I have it wired correctly in my actual circuit, with the digital pins wired up to the SCR gates.

I will correct the schematic, apologies for the confusion

I have snubbers between the anode and cathode for each SCR consisting of a 330ohm resistor and a 10nF 50V ceramic capacitor in series. Is this an incorrect set up? I also shut off the boost converter for the capacitors before allowing the SCRs to be triggered, if that makes any sort of difference.

updated schematic here:

Hi,
Why instead feed the voltage to the coil do not ground the coil.?

Do you mean placing the coil between the capacitor bank anode and SCR anode rather than SCR cathode and capacitor bank cathode? I can modify my circuit to do this, what difference would this make?

Hi,
Why you do not use a mosfet? Also when you use an scr for DC voltage once it is ON you need to open the circuit. It will stay on until you open the circuit or short the scr.

have snubbers between the anode and cathode

A snubber goes in seriese with an SCR not across it. Its purpose is to reduce the rate of current rise in a circuit known as "di by dt "
Normally I would expect the load to be in the anode not the cathode. It is marginal safer that way and the gate voltage does not have to be so high. You also then do not need that 200R resistor that is just taking power for no reason except to supply a gate referance.

tauro0221:
Also when you use an scr for DC voltage once it is ON you need to open the circuit. It will stay on until you open the circuit or short the scr.

That happens automatically when the capacitor is discharged and an SCR turns on faster.

tauro0221:
Hi,
Why you do not use a mosfet? Also when you use an scr for DC voltage once it is ON you need to open the circuit. It will stay on until you open the circuit or short the scr.

I wasn't sure which would be better for a high-voltage/high surge current application. Will a mosfet be able to handling the 400V / ~100A surge? And will it prevent the interference mentioned earlier in the thread that causes the other SCRs to trigger?

I cut off the boost converter power before I trigger the SCR to prevent it from staying latched after firing and the capacitors fully discharge so there is no longer any current to keep it latched.

Grumpy_Mike:
A snubber goes in seriese with an SCR not across it. Its purpose is to reduce the rate of current rise in a circuit known as "di by dt "
Normally I would expect the load to be in the anode not the cathode. It is marginal safer that way and the gate voltage does not have to be so high. You also then do not need that 200R resistor that is just taking power for no reason except to supply a gate referance.

I see, I will modify the placement of the load then.

Will limiting the current rise rate prevent the interference?

Thank you for your input

EDIT: Is this how I should be going about placing the load and snubber? Would the snubber not severely limit the output to the coils? Or did you perhaps mean have the snubber in series with the SCR while the coil is parallel to the snubber

Hi,
Sorry, but I overlooked the C2 capacitor. That will once the capacitors are discharged it will turn off the SCR. There it is a mosfet 650 volts 100 amp but the price it is out of the question. For the price of the mosfet the SCR it is the best alternative. It will do the job.

1/ I take it you're not really using 1N4148's to distribute the +400v? - they're only rated 75v / 150mA. They'll break down when the first coil fires and dump all the energy from all 3 capacitors into that coil . Suggest 1N4007's

2/ In photographic flashguns it's normal to use IGBT's for the current switching.

3/ Aren't you earthing the common -ve's of the big capacitors?

I'll think about the rest of the circuitry...

Allan.

allanhurst:
1/ I take it you're not really using 1N4148's to distribute the +400v? - they're only rated 75v / 150mA. They'll break down when the first coil fires and dump all the energy from all 3 capacitors into that coil . Suggest 1N4007's

2/ In photographic flashguns it's normal to use IGBT's for the current switching.

3/ Aren't you earthing the common -ve's of the big capacitors?

I'll think about the rest of the circuitry...

Allan.

Hi Allan,

You are correct, the applet I used to generate the schematic automatically set a placeholder part# for the diodes it seems, and I forgot to remove them. I am not sure of the part#, but the diodes I am actually using are rated for a good 1000V and 6A (I'm assuming far higher for surge currents). I already have tested these diodes and I can confirm they are powerful enough to prevent one capacitor bank from discharging through the diode back into another upon firing.

I'm using 600V 8A rated SCRs as that is what I was able to easily and quickly source for cheap from my local Fry's. Would switching to IGBTs prevent the apparent issue I'm having with the SCR interference? What IGBT would you recommend if so?

I'm not entirely sure what you mean by earthing the -ve's, I just had all of the capacitor banks wired up in parallel to the boost converter. Do you mean creating a path to ground it to the earth? I don't think it is necessary, as the capacitor banks are at at relatively (for coilguns) low 600uF each. The project is also intended to be portable so I don't know if having an earth connection would be easy to do without compromising that aspect.

Will limiting the current rise rate prevent the interference?

Given that is what causes interference then yes.

That snubber capacitor is wrong. All the current you are triggering has to go through that. Snubber - Wikipedia

Grumpy_Mike:
Given that is what causes interference then yes.

That snubber capacitor is wrong. All the current you are triggering has to go through that. Snubber - Wikipedia

Looking at the diode snubber, is this what I need?

It seems I originally used the wrong form of snubber where I had it across the SCR to suppress voltage rise rather than current rise, if I am understanding correctly. Are both necessary for my application? Though I have found having the snubber across the SCR acts as a safety for slowly draining the capacitor banks when they are not charging.

I am not convinced with your circuit. The grounding looks all wrong. The resistors R2, R4 and R6 seem to be ground lift resistors, why on earth ( pun ) would you want to do that? What do you think they do?

Rubbish comment deleted 'cause I'd misread your circuit

Grumpy_Mike:
I am not convinced with your circuit. The grounding looks all wrong. The resistors R2, R4 and R6 seem to be ground lift resistors, why on earth ( pun ) would you want to do that? What do you think they do?

Not sure what a lift resistor is, but those are there for the SCR gate return path back to the arduino. The arduino logic pins output 5V at a maximum of 40mA so I stuck 330ohm resistors to limit the gates to 15mA. Is this incorrect use of ground? I thought ground could be used to represent common return paths since I didn't want to clutter the schematic with more lines. Probably a noob mistake on my end but I suppose that's why I'm here to learn, haha.

Hi,
Your description at the beginning of your problem you said that by firing the others disconnected the system worked OKAY or from the micro. Means you experiencing noise interference that your already know. Then why not use a driver opto coupler like the MOC3071M or the MOC3072M. By using opto coupler your SCR gate would be isolated from the micro. Just a thought.
Attached it is the data sheet.

MOC3072M-1082602.pdf (495 KB)

It is very surprising that the gates of all the SCRs are at floating conditions having had connections at the cathodes of the isolated diodes. When the 1st SCR fires, the other two also undergo fire as their floating gate circuits work as antennas and pick up the firing noise of the 1st SCR. Therefore, the gate circuits of the SCRs should be tied to ground by 1k resistor. This is just one preventive measure.