Connect with us

fast risetime hv circuit

Discussion in 'Electronic Design' started by amdx, Feb 26, 2008.

Scroll to continue with content
  1. amdx

    amdx Guest

  2. whit3rd

    whit3rd Guest

    Search on 'Blumlein pulse generator', is the only idea that occurs to
  3. John Larkin

    John Larkin Guest

    Spark gap! Or exotic avalanche semiconductors, much more work.

  4. Joerg

    Joerg Guest

    It's often done with a large stack of transistors that are avalanched. I
    can't find the link but there was a Dutch group that needed a pulse of
    over a kilovolt and they used a stash of about 20 transistors. AFAIR
    those were the old BC107 transistors.
  5. kevin93

    kevin93 Guest

  6. Joerg

    Joerg Guest

    Great link. Figure 13 shows how it's done. For sub-nsec pulse widths
    coax stubs are often used.
  7. amdx

    amdx Guest

    Have seen any information about sharpening the risetime of sparkgaps?
  8. John Larkin

    John Larkin Guest

    There's a huge amount of literature on generating very fast HV pulses
    from spark gaps, avalanche semiconductors, laser-triggered switches,
    and even mechanical contacts.

    What do you really need, in terms of voltage, risetime, pulse width,
    load, rep-rate? What's the application?

  9. Joerg

    Joerg Guest

    Seems a bit old but figure 13 shows how it's done. You can achieve
    sub-nsec transitions with proper layout and other transistors.
  10. John Larkin

    John Larkin Guest

    The Zetex parts are very repeatable, trigger beautifully, and slam on
    to nearly zero voltage drop when they fire. But they tend to be slow.
    I've not seen them get down to 1 ns risetime.

    They would be good for driving a shock line or a Grekhov avalanche
    diode to speed up the edges.

    Some interesting stuff and links:

    McEwan did a fair amount of stuff on using common diodes to make fast
    HV edges. Some of his papers are public.

  11. Joerg

    Joerg Guest

    Yes, the challenge in this case will be to find a transistor that
    avalanches fast enough and also has a sufficiently long lifetime in that
    mode. I guess the guys at Photonicstech have, and then probably
    carefully dremeled away the lettering on them ;-)

    Thanks! Great link.
    I've finally found the Dutch paper again. They used the BC107 which is
    sluggish but then again this was in 1978 when everone was tooling around
    in VW-buses with peace stickers and flower vases:
  12. John Larkin

    John Larkin Guest

    The problem with stacking transistors like this is inductance.

    The Zetex parts are made in Russia, possibly on an older fab line. My
    limited testing indicates that older diffused transistors avalanche
    usefully, but newer epitaxial ones don't.

  13. amdx

    amdx Guest

    10kv, Risetime? not sure yet but sub nanosecond. Working on drive for a
    TEA laser.
  14. John Larkin

    John Larkin Guest

    Hey, here's a spark gap!

    I've seen IGBTs driving saturating magnetics to get close to these
    speeds, in MOPA eximer lasers used in ic fab.

  15. Joerg

    Joerg Guest

    It sure ain't ancient. Digikey has over 20000 of the Fairchild PN2369 in
    stock and they wouldn't do that if there was no demand. Of course it
    can't take any voltage to write home about so you'd need a boatload.
    OTOH they cost less than $0.03 in large quantities.

    Then there are tons of STO23 versions such as ZUMT2369, PMBT2369,
    MMBT2369. Tens of thousands in stock at Digikey, and cheap.
  16. Tim Williams

    Tim Williams Guest

    I don't know about that. *Useful* range no, but as far as avalanche, they
    go at like, 80V.

    2N3904s avalanche, too -- I have a batch of Fairchilds that go off in the
    110V range, with less than a couple nanosecond rise time (probably limited
    by my scope's bandwidth).

  17. JosephKK

    JosephKK Guest

    I bet a gotcha that all the sot parts don't avalanche worth snot.
  18. John Larkin

    John Larkin Guest

    Whyzat? They're probably the same silicon.

  19. Ecnerwal

    Ecnerwal Guest

    Use a krytron or thyratron? Not sure about getting down to low hundreds
    of picosecond rise times, though. I worked with a lot of those, as well
    as funky SF6 Marx generator spark gaps several decades back, but I can't
    really recall specific risetimes - I was in the making the equipment
    work end of things, not the collecting and analyzing data end of things.

    Not a bad read for 11.5 years ago:

    ....just whip out your high intensity laser and shoot a pulse into the
    spark gap - you can probably cobble that up over your lunch hour, eh?
  20. Joerg

    Joerg Guest

    Wow, I didn't know they avalanched this high. Cool.

    You might beat that by using the SOT23 version MMBT3904 and RF-style
    trace layout.
Ask a Question
Want to reply to this thread or ask your own question?
You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.
Electronics Point Logo
Continue to site
Quote of the day