Exceeding the dV/dt for an SCR?

Discussion in 'General Electronics Chat' started by drodgers, May 17, 2008.

  1. drodgers

    Thread Starter Member

    May 11, 2008
    10
    0
    Hi I am triggering an coil-gun with an SCR, The circuit is below:
    [​IMG]

    SCR rated for:
    800 V, 16A, Peak I =167 A
    dV/dt = 500 V/us, dI/dt = 100 A/us

    Gate voltage Max = 12V

    Gate Current = 100mA

    The trigger works fine when the caps are charged to about 30V.
    Every time I trigger it when the caps are charged to more than 100V though the SCR becomes locked open - conducting between all pins, I have destroyed about 10 of them trying to solve this problem and I am getting desperate.

    Basically I want to know what I am doing wrong and how I can fix it.

    - Thanks
     
  2. The Electrician

    AAC Fanatic!

    Oct 9, 2007
    2,281
    326
    One thing that would help a lot would be to disconnect the 400 volt charging supply before you trigger the SCR. With the 400 volt supply connected, there is a continuous source of current to keep the SCR turned on. If it's disconnected, the only source of energy is the 9400 uF cap, and it will not be able to hurt the SCR on its own.
     
  3. drodgers

    Thread Starter Member

    May 11, 2008
    10
    0
    I do disconnect the supply, I just left the supply and switch out for simplicity.

    -Thanks
     
  4. The Electrician

    AAC Fanatic!

    Oct 9, 2007
    2,281
    326
    My next guess would be that you have exceeded the I^2*t rating for the SCR. What is the part number for the SCR? Can you provide a link to a spec sheet on the web?

    Do you have an oscilloscope? If so, then you should put a current shunt in series with the SCR so you can measure the current pulse in the SCR and see if you're exceeding the SCR ratings.
     
  5. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    Even though you have diodes in the circuit to take care of the reverse EMF pulse, they may not be turning on quickly enough to save the SCR.

    Try putting small caps across the two reverse EMF diodes, and the SCR as well; something in the 300pF - 3nF range; HV ceramic would be fine. They'll absorb the leading edge of the EMF spike until the diodes can begin conducting.
     
  6. drodgers

    Thread Starter Member

    May 11, 2008
    10
    0
    Thanks for the responses.

    The I^2t value is 128A^2s

    Unfortunately I don't have an oscilloscope.

    The spec sheet is here: http://www.dse.com.au/isroot/dse/support/Z4333.pdf

    I will try your capacitor idea, but I will wait for a few more suggestions first (I only have one SCR left on hand and so I don't want to blow it if the Cap suggestion doesn't work).

    Thanks
     
  7. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    OK.
    Look at the datasheets for the diodes you're using to suppress the reverse EMF pulse.

    The more current-capable they are, the longer their turn-on time. That's just the way it works.

    Small caps will help to "suck up" the big transients. Large caps have far too much resistance. You have a whole bunch of small inductors which are getting nailed with a big pulse, and they all have their own individual resonance with the rest of the circuit.

    You need a good number of small caps to soak up that big leading-edge pulse those inductors are going to put out.

    Fewer, larger caps just plain won't work. You need a number of small caps. The big caps just have too much in the way of parasitic resistance and inductance to be of much use.
     
  8. drodgers

    Thread Starter Member

    May 11, 2008
    10
    0
    My diodes are rated for 6A so they probably aren't turning on so fast.

    I have about 6 3kV 0.0047uF Caps lying around, would they do?

    How many should I place in parallel across the diodes?

    Should I place any across the SCR?

    -Thanks
     
  9. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    Your 4.7nF caps - what kind are they? Ceramic?

    I would put a pair across the SCR and both of the diodes.

    I would also get a few more smaller caps in the range of 220pF-470pF, and put them across the SCR and both diodes.

    This is to snub the very fast rising edge of the EMF pulse. You wouldn't see this pulse on a traditional O-scope - you would need a storage scope to see it.
     
  10. drodgers

    Thread Starter Member

    May 11, 2008
    10
    0
    Yes they are ceramic.

    Thanks - I will try your suggestion (but probably not for 3 or 4 days - Exams)
     
  11. drodgers

    Thread Starter Member

    May 11, 2008
    10
    0
    I have tried it with the caps - no luck, it blew with a firing at only 35V :(


    If I am exceeding the I^2*t value how do I know and how would I fix it?

    Should I just make a much bigger inductor (like a whole spool of wire)?

    Others seem to have had success in using SCRs as triggers for these ultra-fast pulse devices without killing their SCRs or having to restrict the current flow, should I just try to find some huge SCRs (on ebay)?

    Thanks again.
     
  12. The Electrician

    AAC Fanatic!

    Oct 9, 2007
    2,281
    326
    You could determine if you're exceeding the I^2*T by measuring the current pulse with a scope. Since you said you don't have a scope, you could calculate the theoretical current waveform and see if it exceeds the rating.

    I think if it were me, I'd get a bigger SCR.
     
  13. drodgers

    Thread Starter Member

    May 11, 2008
    10
    0
    Now I'm sure its either the current or the rise of current:

    I tested my last SCR at 300V (It was rated for 500V and I figured that if I was going to destroy another I may as well see how fast the gun can fire) and it exploded in a spectacular and scary way :eek:.

    400J got dumped through the SCR and now it is in pieces, I think I will need to find a bigger one... or maybe a few.

    Thanks for the help.
     
  14. The Electrician

    AAC Fanatic!

    Oct 9, 2007
    2,281
    326
    Go have a look at this little applet:
    http://www.coilgun.info/mark2/rlcsim.htm

    It will allow you to get an idea of what your current pulse probably looks like.

    In your first post you described your inductor as having 100 "coils" (I assume you mean turns) with inner diameter of 1 cm and outer diameter of 2 cm max, wound with .63 mm diameter wire (this is roughly equivalent to 22 gauge AWG).

    I scramble wound an inductor on an .5 inch plastic tube, closely packed to get an outer diameter of about 2 cm. and then measured the inductance and resistance of the inductor. I get these values:

    Inductance = 86 uH = .086 mH

    Resistance = .276 ohms.

    Using the applet, I put in 75 uH, 273 milliohms, 9500 uF (.95 mF) for the cap, charged to 400 volts.

    The applet says the first current peak will be 1217 amps. If I change the inductor to 125 uH, then the peak current is 1139 amps.

    If we add another 200 milliohms for the ESR (equivalent series resistance) of your 9400 uF capacitor (which is probably too high), and then we set the resistance slider to 472 milliohms we get a peak current of 744 amps.

    It seems clear that you need a bigger SCR, and you probably need to get the inductance up into the millihenry range instead of the neighborhood of 100 uH. You need to use more turns of bigger wire. The resistance of the inductor should be less than 1/2 ohm.
     
Loading...