Help Needed with SCR and Optocoupler

[adamq]

Hi!

I have a project with a high voltage side (340VDC) and a low voltage side (5VDC). It's a high powered xenon flash with a low voltage trigger circuit.

I'm trying to use a 5VDC (currently just controlled by a momentary switch, ultimately will be from camera hotshoe adapter) to close the 340VDC circuit thus firing the the flash.

See the attached schematic.

Basically, this is what should happen (in my mind):
0) low voltage side is open circuit.
1) momentary "test" button is pressed, low voltage signal enters pin 1 of MOC3022 Optocoupler and leaves Pin 2.
2) Pins 4 and 6 of optocoupler close, allowing trigger current from high voltage side to gate pin of SCR.
3) SCR gate receives signal from optocoupler, thus SCR closes, allowing high voltage side to dump into trigger coil and fire flash.

My problem is that it simply isn't working. I thought maybe the SCR or the Optocoupler were fried somehow, so I replaced both, to no avail.

I'm sure it's something I'm missing or misunderstanding, but I can't figure out what it is.

Optocoupler is MOC3022. Previously MOC3020.

I've tried two different SCRs:
- NTE 54001
- 2N6403

Any help is much appreciated!

 

[adamq]

Also, I'm willing to try other ways of triggering the flash. If you can think of a better way, please suggest it.

 

[jcurrie]

don't know if i can help you but the first thing i would suggest is a fuse on the ac side , then the octo moc 3020 is rated for 115/240 vac.
jc

 

[KrisBlueNZ]

I don't see any definite problems with your circuit that would prevent it from working.

I have drawn up your design with a few suggested changes that you can try.

Normally when an SCR drives a trigger transformer, the SCR is connected with its cathode to the negative rail, for tidiness. The primary of the trigger transformer is connected between the bottom side of C2 and the 0V rail, with the opposite polarity from what you're currently using. When the SCR fires, its anode slams to 0V and the negative side of C2 generates a negative pulse into the trigger transformer. I've drawn it with the primary side connections reversed (tap to 0V) which produces a positive trigger voltage. I don't know whether the trigger voltage should be positive or negative (or whether it even matters).

The way you have it, the cathode of the SCR jumps around unnecessarily. I don't see a problem with it; it's just a bit untidy.

C2 seems unnecessarily large for this application. I doubt the trigger pulse needs to be very long at all. The Wikipedia article on flashtubes implies a minimum duration of 60 ns per centimetre distance between the end electrodes, so 1 µs should be plenty long enough. But you may know better. I've assumed that 47 nF will be plenty; this also means that R3 can be increased. To get 47 nF charged to 85% (2 RC periods) of the total supply voltage, and using two resistors in series for R3, each resistor needs to be about 4.7 megohms for a one second recharge time. If you need faster recharge, reduce them accordingly.

Regarding the trigger current into the optocoupler. The MOC3020 needs 30 mA to trigger, which you won't get from 6V through 300 ohms, but you said you changed to MOC3022, which only needs 10 mA. Even with the maximum LED forward voltage of 1.5V, leaving 4.5V across R2, you'll get 15 mA which is supposed to be enough. I would try with a higher current though.

I would suggest adding a resistor from gate to cathode on the SCR to prevent false triggering.

It's also possible that the triac in the optocoupler is being damaged by a very brief, extremely high current pulse that would occur between the time that it triggers and the time that the SCR has activated and clamped the voltage across the optocoupler's triac. This current should be limited to 1A so a series resistor of at least 360 ohms (I'd suggest two 180 ohm, 1W metal film resistors in series) should be added between the optocoupler and the SCR's gate.

The SCRs you've tried should be fine. I guess you could add a small resistor in series with the trigger transformer primary, to limit the peak current to 16A (for the 2N6403) or 55A (for the NTE54001) but I doubt it's worth it.

I also agree with jcurrie's suggestion that you should have a fuse on the AC side!

BTW it's nice to see the ISO date format! I wonder how long it will be before everyone uses it... One century? ten?

 

[jpanhalt]

I believe Vgt is referenced to the SCR cathode, as you (KrisBlueNZ) have shown it. The OP's design does not do that. His design might work if he moved the trigger connections from the anode and gate to the cathode and gate. There are other considerations, such as di/dt that are discussed in all references.

Here's a good resource: http://www.onsemi.com/pub_link/Collateral/HBD855-D.PDF

John

 

[KrisBlueNZ]

The OP's circuit will still cause gate-to-cathode current to flow when the optocoupler conducts. It uses the voltage that already exists across the SCR to feed current into the gate. There's no real difference between his circuit and mine; the fact that my circuit has the gate connected to the 0V rail doesn't affect the way the current flows in the gate drive circuit. You can't connect the optocoupler between gate and cathode because it doesn't generate voltage; it only allows current to flow (due to voltage present between anode and gate).

That's a good reference. Thanks for the link.

 

[Arouse1973]

I know it might be a silly question but make sure you have the transformer connected correctly. Also it would be nice to see a bleed resistor across those capacitors in case the mains is unplugged at it's peak remember the fuse is there to protect the circuitry and wiring from overheating and not you from a shock. I don't know much about xenon tubes but you say it's a 200Joule version. Does that mean it needs 200J or is that the maximum. Because you only have 173.4 on a good day with all caps equal.

Thanks
Adam

 

[KrisBlueNZ]

Adam, yes, a bleed resistor is a good idea. For two 1W resistors in series, running at half their rated dissipation, each resistor should be about 68k. It will take about 20 minutes to discharge a 3000 µF capacitance to a safe voltage (30V DC).

Adam, do you know what polarity the trigger voltage needs to be? Or do you mean that he should check that he doesn't have the small and large windings exchanged?

 

[Arouse1973]

Adam, yes, a bleed resistor is a good idea. For two 1W resistors in series, running at half their rated dissipation, each resistor should be about 68k. It will take about 20 minutes to discharge a 3000 µF capacitance to a safe voltage (30V DC).

Adam, do you know what polarity the trigger voltage needs to be? Or do you mean that he should check that he doesn't have the small and large windings exchanged?

Hi Kris
I think the trigger needs to be negative. And yes I meant small and large windings exchanged.
Thanks
Adam

 

[adamq]

Hi guys,

It's been a couple days since your very helpful replies, and I wanted to give you an update.

Kris, I've gone ahead and implemented the changes in your schematic and replies.

Also, I have a 4A fuse protecting the circuit at the power supply that I neglected to include in my original schematic. You seem very knowledgeable and you've both been very helpful.

Having said that, I wasn't able to get it to work. At this point, I'm thinking it's user-error. I think the most likely situation now is that I've damaged some of the components from soldering, unsoldering, and resoldering, or possibly having it wired up incorrectly the first time.

Regardless of the cause, I'll be rebuilding it again. I'm working today, so tomorrow I will go back to the supplier and buy fresh components.

- MOC3022 Optocoupler
- 6-pin IC Socket
- 2N6403 SCR
- 180R 2W x 3
- 4.7MR 2W x 2
- 270R 2W x 1
- 47nF Capacitor

I'm determined to get this to work. The learning curve is part of the fun

PS: I'm surprised you noticed the ISO date format. That's the programmer in me. The only people who have ever noticed before are the one's complaining that I'm writing it wrong."

I don't see any definite problems with your circuit that would prevent it from working.

I have drawn up your design with a few suggested changes that you can try.

Normally when an SCR drives a trigger transformer, the SCR is connected with its cathode to the negative rail, for tidiness. The primary of the trigger transformer is connected between the bottom side of C2 and the 0V rail, with the opposite polarity from what you're currently using. When the SCR fires, its anode slams to 0V and the negative side of C2 generates a negative pulse into the trigger transformer. I've drawn it with the primary side connections reversed (tap to 0V) which produces a positive trigger voltage. I don't know whether the trigger voltage should be positive or negative (or whether it even matters).

The way you have it, the cathode of the SCR jumps around unnecessarily. I don't see a problem with it; it's just a bit untidy.

C2 seems unnecessarily large for this application. I doubt the trigger pulse needs to be very long at all. The Wikipedia article on flashtubes implies a minimum duration of 60 ns per centimetre distance between the end electrodes, so 1 µs should be plenty long enough. But you may know better. I've assumed that 47 nF will be plenty; this also means that R3 can be increased. To get 47 nF charged to 85% (2 RC periods) of the total supply voltage, and using two resistors in series for R3, each resistor needs to be about 4.7 megohms for a one second recharge time. If you need faster recharge, reduce them accordingly.

Regarding the trigger current into the optocoupler. The MOC3020 needs 30 mA to trigger, which you won't get from 6V through 300 ohms, but you said you changed to MOC3022, which only needs 10 mA. Even with the maximum LED forward voltage of 1.5V, leaving 4.5V across R2, you'll get 15 mA which is supposed to be enough. I would try with a higher current though.

I would suggest adding a resistor from gate to cathode on the SCR to prevent false triggering.

It's also possible that the triac in the optocoupler is being damaged by a very brief, extremely high current pulse that would occur between the time that it triggers and the time that the SCR has activated and clamped the voltage across the optocoupler's triac. This current should be limited to 1A so a series resistor of at least 360 ohms (I'd suggest two 180 ohm, 1W metal film resistors in series) should be added between the optocoupler and the SCR's gate.

The SCRs you've tried should be fine. I guess you could add a small resistor in series with the trigger transformer primary, to limit the peak current to 16A (for the 2N6403) or 55A (for the NTE54001) but I doubt it's worth it.

I also agree with jcurrie's suggestion that you should have a fuse on the AC side!

BTW it's nice to see the ISO date format! I wonder how long it will be before everyone uses it... One century? ten?

 

[jcurrie]

Kris just a thought but would a phototransistor be better than the triac? we used the MOC3020 for a/c motor control.

jc

 

[adamq]

Kris just a thought but would a phototransistor be better than the triac? we used the MOC3020 for a/c motor control.

jc

I'd be interested in hearing the pros/cons of this

 

[KrisBlueNZ]

Well, I don't have a lot of experience with this kind of application, so I could be overlooking something here, but I think that once the SCR fires, its anode, gate, and cathode will all be within a volt or two of each other, and no significant current will flow in the output circuit of the optocoupler.

If it's an MOC3022 (triac output) optocoupler, the current will drop below the holding current and it will turn OFF (until the SCR turns OFF and the anode-gate voltage starts to rise, anyway). If it's a phototransistor, it will continue to conduct until the control pulse goes away.

So whether it's a transistor or a triac shouldn't make any difference. The triac output may give a more definite transition from OFF to ON and provide a cleaner trigger to the SCR. That's all I can suggest.

 

[adamq]

Well, I don't have a lot of experience with this kind of application, so I could be overlooking something here, but I think that once the SCR fires, its anode, gate, and cathode will all be within a volt or two of each other, and no significant current will flow in the output circuit of the optocoupler.

If it's an MOC3022 (triac output) optocoupler, the current will drop below the holding current and it will turn OFF (until the SCR turns OFF and the anode-gate voltage starts to rise, anyway). If it's a phototransistor, it will continue to conduct until the control pulse goes away.

So whether it's a transistor or a triac shouldn't make any difference. The triac output may give a more definite transition from OFF to ON and provide a cleaner trigger to the SCR. That's all I can suggest.

Sounds good to me. Thanks again!

I'll be picking the parts up in the morning and trying again. Should know by tomorrow evening (Toronto time).

 

[jcurrie]

thanks Kris my only use of triac opto,s were with AC all my experance with photo ones are DC, also adamq 3000uf at over 240v will defently wake you up if you bridge them .
jc

 

[KrisBlueNZ]

Yeah this circuit is VERY dangerous!

adamq, I guess you know that the electrolytics need to be specified for photo flash use? Standard electrolytics will fail in this application because of the extremely high currents and fast discharge.

 

[(*steve*)]

Of some interest might be the electronic strobe published in this month's silicon chip magazine which uses a 100W LED array.

It might be worth looking in to.

The peak light output will be lower, but you'll get a much faster rising/falling edge.

and the voltages involves are far less dangerous.

 

[adamq]

Yeah this circuit is VERY dangerous!

adamq, I guess you know that the electrolytics need to be specified for photo flash use? Standard electrolytics will fail in this application because of the extremely high currents and fast discharge.

Affirmative

Of some interest might be the electronic strobe published in this month's silicon chip magazine which uses a 100W LED array.

It might be worth looking in to.

The peak light output will be lower, but you'll get a much faster rising/falling edge.

and the voltages involves are far less dangerous.

Hey steve!

Great suggestion but I think I'll stick with xenon for the time being.

Don't get me wrong, LED arrays are okay for some types of photography, but they're more useful for videography. I'm not familiar with LED arrays at fast strobe rate, but an LED array is a fairly "large" light source, physically speaking, so it wraps around your subject, almost like using a softbox or light diffuser on a strobe.

The arc of a xenon strobe is much smaller physically, so it can produce very sharp light/shadows when used without a light modifier.

For examples of what I mean, look at these sets of photos of mine. Please excuse the large format, it's what I had handy.

These photos of Nichole are done with one single unmodified strobe at 400wattseconds in a very very dark room (very small light source, sharp distinct shadows, hard edge to the light).

LEDs would also have a very hard time illuminating to this extent.


These photos of Kay were done with 3 strobes: 1 umbrella, and 2 softboxes (both types are very softened with diffuser panels. Notice there are no hard shadows because the light source becomes much bigger. Edges are very smooth. LED arrays can only deliver this type of lighting (unless somehow focused/directed with lenses, snoots, whatever... not worth it IMO).


These photos of Hunter were done with primarily soft light (softboxes and spot grids) as well. LED arrays could possibly be used to accomplish this, but I don't know how, nor how well.


So, you see the need for the strobe vs LED arrays. Don't get me wrong, LEDs are great, they just serve a different purpose, are usually bulky, and not quite as versatile.

Of course, LED arrays do have their place. It'd be very difficult to shoot video with a xenon strobe

Strobes are still by far the predominant artificial light sources in photography.

Alright, off to bed.

Thanks again for everyone's input and suggestions. I'm really pumped to get this finished up

 

[(*steve*)]

Don't get me wrong...

Don't worry about the explanations, I'm a photographer