SCR Based Pulse Switch Turn Off Problem? / Electromagnetic Radiation / Forums

Forums

4hv.org :: Forums :: Electromagnetic Radiation

« Previous topic | Next topic »

SCR Based Pulse Switch Turn Off Problem?

1 2 next

Move Thread

LAN_403

LutzH

Fri Nov 13 2009, 01:51AM

Registered Member #1721 Joined: Sat Sept 27 2008, 08:44PM
Location:
Posts: 136

I recently aquired a couple of nice big hockey puck SCR's on ebay which I want to use to switch the electrical energy pulse to my laser flashlamp. The rating of the SCR is fine, it will handle up to a 10,000 amp surge, at 1400V. My problem is as follows:

My power supply is a SMPS delivering up to 1200V at about 1-2A to the main storage capacitor. I am using a series trigger transformer to generate the HV pulse to trigger the lamp. My problem is as follows:

If I use the big SCR as a switch for tha main pulse, turning it on is easy, just supply the correct voltage and current to the gate of the SCR.

Now here is the problem: The SCR is now on after the gate is triggered and the flashlamp lights up, so far so good. Now the capacitor charge decreases as the lamps converts the energy. Now I fear the main SMPS power supply will simply start to discharge through the still ionized flashtube, like an arc lamp, because the SCR never recovered from its on state, due to the lack of any zero, or reverse voltage?

There should be a simple way to supply a momentary reverse voltage via a second capacitor, and maybe even with a second smaller SCR, like is done when SCR's are used to switch DC. The problem is that the only schematics which I can find use a manual switch etc. to do this. By then my flashlamp is probably toast from the CW discharge.

Does anyone have a schematic for a circuit which can turn off the main SCR without disconnecting the main power to the charging capacitor? I thought about using HV diodes to isolate it in some way, but I cannot seem to be able to figure out how to arrange such a switching circuit?

My hope is that someone reading this can figure out a way to modify say the SCR switching circuit from a coil gun etc. The only difference being the lack of the reverse voltage inductive kick back from the coil. Or on second thought; Is the inductive kickback from the series trigger transformer enough to re-set the SCR if I put a blocking diode between the capacitor and the SCR? Thanks in advance for any help with such a schematic .....Lutz :)

PS: I even tried the Perkin Elmer Flashlamp web site, and they have some example circuits using SCR's, but these do not address the turn off problem. Yes also an IGBT would be nice but the surge rating is to low.

Z28Fistergod

Fri Nov 13 2009, 03:12AM

Registered Member #2040 Joined: Fri Mar 20 2009, 10:13PM
Location: Fairfax VA
Posts: 180

Why do you want to switch the capacitor with the SCR?

LutzH

Fri Nov 13 2009, 09:25PM

Registered Member #1721 Joined: Sat Sept 27 2008, 08:44PM
Location:
Posts: 136

Hello:

So Sorry :) I forgot to mention one very, very important point, duuuhhh: The lamp is being run in simmer mode.

Without an SCR or IGBT switch, the capacitor will drain through the lamp which is always on in simmer mode. The simmer power supply keeps the lamp ionized at about 1KV with 50ma all the time. So the main charge in the capacitor needs to be held off by a SCR, or IGBT switch. Then to fire the laser the gate voltage is supplied to the hold off SCR, so that the capacitor can discharge through the lamp. Now comes the problem how to turn the SCR off? Thanks.....Lutz

Avalanche

Fri Nov 13 2009, 09:51PM

Registered Member #103 Joined: Thu Feb 09 2006, 08:16PM
Location: Derby, UK
Posts: 845

what you're referring to is called 'forced commutation', that might help with your searching.

There's a good few pages on it in this book preview

All you need is something to take the load current for long enough for the main thyristor to stop conducting, usually this is done with a capacitor arrangement, but in your case I cannot see why you can't just use something like a high voltage MOSFET connected across the thyristor - and then switch it on for ~1mS after the main pulse. Most important thing is that the voltage drop of whatever you use is less than the thyristor. It would in effect short circuit the main thyristor for long enough for it to switch off (look up the exact time required in the data sheet, non-commutation grade thyristors can be very slow!)

LutzH

Sat Nov 14 2009, 02:42AM

Registered Member #1721 Joined: Sat Sept 27 2008, 08:44PM
Location:
Posts: 136

Its amazing how much difference a word can make "Commutation" using this word I found everything ref. SCR DC switching that I needed. Thanks :)

Z28Fistergod

Sat Nov 14 2009, 02:22PM

Registered Member #2040 Joined: Fri Mar 20 2009, 10:13PM
Location: Fairfax VA
Posts: 180

I think I remember speaking to you about a pulsed YAG laser you were building. One which had a pfn of 4 capacitors and 4 inductors, and a pulse width of around 10ms. Is this the same project? What kind of current are you looking at anyway?

EDIT:
What I'm getting at is this: If it were me, I would arrange four capacitor-inductor series circuits in parallel, with each controlled by an IGBT. That would give you lots of control over the pulse shape and profile by varying the timing of each individual discharge. IE when one leg discharges relative to the others and how long the individual pulse is.

LutzH

Mon Nov 16 2009, 08:25PM

Registered Member #1721 Joined: Sat Sept 27 2008, 08:44PM
Location:
Posts: 136

Hello:

Yes basicaly you are correct, except now I am up to 6 capacitors, each one is 1500V at 220mfd. I scored this new capacitor on ebay recently, and it tested fine. It is a Genaral Atomics unit with 6 capacitors in one big can with a single common ground. It is roughly a 25cm cube. The best part is that the PC board with the 6 protective diodes for each cap, the drain resistors, and also the 4 input HV diodes, are still on the cap, and all test good! This unit came out of a pulse laser of some kind.

Since I plan on operating it at about 1KV, now I have a few more options on pulse control. I still would like to go with a multi-mesh pulse forming network for a nice flat top pulse, now by using 6 inductor coils. My reasoning is by doing it this way, I have some nice options: If I use all 6 caps then the pulse length will be the full 10mS, if I switch to fewer capacitors from the main unit, then the pulse will shorten of course, but it will still be in acceptable limits for the lamp, due in part to the now decreased per pulse energy. Also by doing it this way I want to avoid the extra complexity of using a semiconductor circuit to control the pulse length, and energy.

I very much like your idea of switching each capacitor with an IGBT! This would be sooo slick, but then I would need at least 3 big IGBT's. The peak current would then be about 100-200A regardless of how many of the 6 capacitors are in the lamp circuit. One thing that I have learned the hard way reference pulse power is; that I have to rate components for about 10x the actual pulse energy for reasonable component life.

I happend to already have 2 big SCR's which are rated at 1400V / 900A, with a 10KAmp surge rating. Yes I agree IGBT's would be much nicer, but I have the SCR's already, so they will cost me nothing, so it would be nice to be able to use them. One advantage also with the SCR's would be, if in the future I want to Q-switch the laser and I reduce the pulse length to 250uS, with say a single 140mfd cap., then the SCR's could still handle the high peak current.

I did find a nice article on SCR forced commutation by ON semiconductor, but they did not feature the math to calculate the value of the second SCR, and that of the commutation capacitor, and the resistors. Also all of the articles that I can find on the subject of forced commutation deal with examples of inductive loads like motors.

One thought which went through my mind was: What if I could somehow I could use a property of an existing component in the circuit, to accomplish the forced commutation of the SCR? There is a series pulse HV trigger transformer in the circuit already for example, and I was thinking what happens in this circuit when the main pulse goes through this transformer? Maybe I could tap into the primary of this pulse transformer, and use the pulse generated to somehow commutate the SCR? at the time of the laser pulse, when the series trigger transformer is not in use, and the lamp is ionized in simmer mode by a seperate simmer power supply. Again here is another possibility, maybe the simmer power supply could be used is some way to commutate the SCR? What if the lamp was simmered at the reverse polarity of the capacitor discharge? Or would the lamp even trigger then?

Right now I have more questions than answers, I am sure there is a way to utilize something already in the circuit. The big question is "what and how" :) Thanks for your thoughts.

Z28Fistergod

Tue Nov 17 2009, 03:39AM

Registered Member #2040 Joined: Fri Mar 20 2009, 10:13PM
Location: Fairfax VA
Posts: 180

Ok I can understand keeping it simple. For what it's worth, 200 amps for 10ms is pretty easy for many IGBTs.

LutzH

Tue Nov 17 2009, 10:44PM

Registered Member #1721 Joined: Sat Sept 27 2008, 08:44PM
Location:
Posts: 136

Maybe you are on to something, if I cannot figure out how to use the SCR's, then I could always sell them since I even have the spec sheets now, and just buy a couple big IGBT's on ebay. For now I want to see if I can work it out with the SCR's, and give my brain a little workout, if it becomes a project in itself then I will take your advice and just buy the IGBT's :) Thanks

klugesmith

Wed Nov 18 2009, 01:34AM

Registered Member #2099 Joined: Wed Apr 29 2009, 12:22AM
Location: Los Altos, California
Posts: 1716

LutzH wrote ...
For now I want to see if I can work it out with the SCR's, and give my brain a little workout, if it becomes a project in itself then I will take your advice and just buy the IGBT's :) Thanks

Maybe there's a solution with a small inductor in series with SCR and capacitor in the pulse loop,
that's not part of the simmer current loop. Just enough inductance for a slightly underdamped pulse discharge, that would reverse the current so SCR turns off naturally?
EDIT: On more careful reading of this thread, I see you're already plenty into inductors for pulse shaping.

1 2 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob