Gate protection diode for SCR?

Thread Starter

kenw232

Joined May 18, 2009
54
I made a small pulse circuit. More like a coil gun. I charge the cap bank (C1) up to about 1200V. I push a button (S1) and the charge in the cap bank is dumped via the SCR into the transformer and I get a nice single spark across the spark gap. This works fine, until it doesn't. Which is when the gate of the SCR gives out, I presume as shown in the scope view (channel 2, circled in purple) because it reaches like 250V. I don't know where this voltage comes from or how to stop it. It's suppose to only be 9V or so (note the value for R2 is probably wrong but irrelevant).
circuit.png gate1-1.png

Does anyone know what is happening here or how to stop it? I've tried a BY448 diode across the + and - lines to suppress backemf but they just fry for whatever reason. Is an opt-islolator the only option to protect the gate?
 

DickCappels

Joined Aug 21, 2008
10,153
It might be that when the SCR turns off the high rate of change on the anode (yes, I see your snubber) is enough to induce current into the gate that is sufficient to trigger it back on, which might not be good if a large current happens to be flowing at the time. If that is the case, then your snubber does not snub sufficiently.

To what use will you put this circuit?
 

MrAl

Joined Jun 17, 2014
11,388
Hi,

If your yellow trace shows the anode voltage then the only other way it can get a positive voltage on the gate is through the cathode or through EMR. I see that your scope ground is way off to the left, is that true in the real circuit too? Because if it is then you could be seeing the voltage across the wire inductance which could be significant for a short time. So you could really be measuring the gate voltage plus the line inductance voltage. The snubber has to be in tight next to the SCR with short leads too.
If that's not the case then it could just be radiated from the transformer or lines.
If it is possible to look at the primary you can see if the wave follows the primary wave.
 

Thread Starter

kenw232

Joined May 18, 2009
54
put a zener doide across the gate and cathode terminals, say 30V 1Watt, what is the thyristor model number?
Its a CMA30E1600PN-ND SCR. They get destroyed a lot. I can't go for long with it getting ruined because of this. I don't know if I'm even measuring properly.
 

Thread Starter

kenw232

Joined May 18, 2009
54
I had a BZX85C15 15V zener and put it across the gate and cathode and it did nothing unfortunately. Scope shows the same thingzener1.png .
 

Thread Starter

kenw232

Joined May 18, 2009
54
If your yellow trace shows the anode voltage then the only other way it can get a positive voltage on the gate is through the cathode or through EMR. I see that your scope ground is way off to the left, is that true in the real circuit too?
The yellow trace shows the charge/voltage on the cap bank, which is the same as the voltage against the anode. The scope ground is off to the left but that's the exact same thing as putting it directly on the cathode is it not? I even did it and its the same reading.

So you could really be measuring the gate voltage plus the line inductance voltage.
I don't know exactly what the line inductance voltage is your referring too? You mean going into and out of the transformer's primary? How does any additional voltage get to the gate at all? If the SCR dumps the charge from C1 into the primary of the transformer , the inductance stored in the primary collapses when that single pulse finishes - is that backemf coming into the gate then?
 

ian field

Joined Oct 27, 2012
6,536
It might be that when the SCR turns off the high rate of change on the anode (yes, I see your snubber) is enough to induce current into the gate that is sufficient to trigger it back on, which might not be good if a large current happens to be flowing at the time. If that is the case, then your snubber does not snub sufficiently.

To what use will you put this circuit?
ST make an scr specifically for florescent starters that can handle the sharp dv/dt involved - not sure they're rated for the voltages the OP mentioned, but they go fairly high.
 

Thread Starter

kenw232

Joined May 18, 2009
54
I see that now, the TYN1225 has a dv/dt of 1500v/uS. Thanks. That might stop them from triggering automatically which was happening once in a while too.
 

MrAl

Joined Jun 17, 2014
11,388
The yellow trace shows the charge/voltage on the cap bank, which is the same as the voltage against the anode. The scope ground is off to the left but that's the exact same thing as putting it directly on the cathode is it not? I even did it and its the same reading.



I don't know exactly what the line inductance voltage is your referring too? You mean going into and out of the transformer's primary? How does any additional voltage get to the gate at all? If the SCR dumps the charge from C1 into the primary of the transformer , the inductance stored in the primary collapses when that single pulse finishes - is that backemf coming into the gate then?
Hi,

By line inductance i mean ANY wire in the circuit. In this case the wire between the cathode and where you have that ground lead of the scope connected to in your drawing. This drew my attention right away because when you measure a signal in a high speed circuit (even if it is only a one time event) you need to place the probes directly onto the points being measured. So the scope probe would go directly to the gate and directly to the cathode, not to another place in the circuit where the cathode happens to be connected to. That's the only way to get a reliable reading.
If you look at your drawing again, you'll see that it looks like the scope lead on the left is FAR from the cathode physically. If that's the case then you are not just measuring the gate voltage you are also measuring any drop in voltage caused by the inductance of the lead that goes to the cathode. So the ground lead must be placed directly on the cathode.

You are dealing with a high speed high energy circuit here so there has to be more precautions then usual.

First, the dv/dt is usually responsible for the undesired turn on of the SCR. That means when the cap bank is not charged yet and the +1400v supply is first connected, a high dv/dt can occur across the SCR and thus turn it on right then and there, before the switch S1 is even closed. That's the possible dv/dt problem.
With L1 in the circuit that should help with that problem, along with the 1uf and transformer primary. If you suspect this is a problem, check it with the scope with leads right on the anode and cathode (within 1 inch). This doesnt seem to be the problem you are seeing though, unless the 1uf cap resonates with the transformer primary, which i am sure it does to at least some extent, but that doesnt seem to be the main problem right now.

Second, the switch S1 may bounce, which would give the SCR a mixed signal rather than a single turn on command. That could make the SCR behave erratically. A small cap might help with this, but a positive debounce circuit would definitely ensure this doesnt happen either. A lower value of R1 could help too.

Third, this circuit will necessarily have a huge di/dt when the SCR is first turned on. This means that not all of the SCR die area will conduct immediately, but will take some time to fully conduct. If the area closest to the gate area does not conduct right away, there could be a voltage divider action between the anode and cathode which would show up at the gate. In other words, for a short time period the SCR looks like a resistive voltage divider with the gate being the center node. With 250v on the gate and 1000v at the anode that would mean a 4:1 divider ratio.
An attempt to solve this could be as simple as lowing the impedance of the gate to ground, but that could also cause a higher current during that time so that may blow out the gate die area. A better idea would be to limit the di/dt itself with some extra resistance or inductance in the anode circuit. Unfortunately this has the adverse affect of lowering the dv/dt at the transformer primary, but that's the way it goes. Alternately a higher power SCR might work with what resistance and inductance already exists without having to add any more.

So the first thing to do is place the scope probes on the SCR itself and take new readings. If that still shows similar readings, then try limiting the di/dt. It's most likely not really the dv/dt that is a problem because that usually shows up when the anode voltage shoots up high very fast, and that only happens when the power is first turned on, and you could check that easily by measuring the anode to cathode voltage when connecting power for the first time.
Also a small cap from the left side of R2 to ground might be enough to clean up any bounce.
The ground lead of the battery should also be connected as close as possible to the SCR cathode.
 
Last edited:

Thread Starter

kenw232

Joined May 18, 2009
54
Thank for the response. The dv/dt rising too fast is not an issue. I see it sometimes, it triggers the SCR by itself but its rare and not my problem. My problem is my SCR's keep getting fried.

I measured directly against the leads and its mostly the same result. I'll try the small cap to stop any switch bounce but its a whole circuit design flaw that I don't understand.
 

Thread Starter

kenw232

Joined May 18, 2009
54
Can't see any mention of it, but do you remove the 1400V supply before you manually trigger the SCR?
yes, I charge the cap bank, and the 1400V source is cut off so all I have is the charge in the cap bank when I push the trigger button for the gate.
 

Thread Starter

kenw232

Joined May 18, 2009
54
looks like capacitance & inductance ringing to me, you need a snubber filter to flatten that out.
I thought I had a snubber, hence the .22uF and 100Ohm resistor. I'll have to try something else, I don't think those values are proper then.
 

t_n_k

Joined Mar 6, 2009
5,455
I would place a high PIV freewheeling diode across the transformer primary. I suspect that you may be seeing the result of the arc 'quenching' with the primary still energized. At that condition the transformer magnetizing inductance will come into play as the secondary current rapidly drops to zero. This presumably gives rise to the ringing or oscillatory effect you see on your scope traces.
With the diode across the primary you may need to add some current limiting resistance to the input charging side rather than just having the 20uH inductor.
 
Top