BJT negative resistance experiments

Discussion in 'General Electronics Chat' started by t_n_k, Jun 30, 2014.

  1. t_n_k

    Thread Starter AAC Fanatic!

    Mar 6, 2009
    5,448
    782
    MrChips likes this.
  2. RichardO

    Well-Known Member

    May 4, 2013
    1,238
    385
    Thanks. The link to Dick Cappels' project pages answered a question I had literally a couple of days ago -- will a transistor avalanche with the collector and emitter swapped? And obviously the answer is "yes".

    On a related note, I tried to avalanche a PNP transistor (a 2N3640) today without any luck. I don't know enough of the physics of transistors to know if PNP's won't avalanche or if it is just the one I tried won't. More work needed here.
     
  3. GopherT

    AAC Fanatic!

    Nov 23, 2012
    6,067
    3,837
    Pathetic youtube video claims it goes negative resistance in only on a specific voltage and resistance combination.
     
  4. t_n_k

    Thread Starter AAC Fanatic!

    Mar 6, 2009
    5,448
    782
    I can certainly get NPN types such as BC107, BC337 and 2N2222 to run as relaxation oscillators. None of the PNP types I had at hand showed the necessary negative resistance characteristic. I initially checked transistors by driving the E-C terminals [base open] with a variable current source adjusted from from a few uA up to about 6mA and watched for decreasing VEC as IEC increased.
     
  5. Jony130

    AAC Fanatic!

    Feb 17, 2009
    3,962
    1,097
  6. RichardO

    Well-Known Member

    May 4, 2013
    1,238
    385
    Thanks. This confirms what I saw with the single transistor I tried.
     
  7. RichardO

    Well-Known Member

    May 4, 2013
    1,238
    385
    Interesting. Thanks a lot for the link. I will have to study it more later.

    There is one question your observations bring to my mind. That is, how does the collector/emitter versus collector/base breakdown relate to when the transistor avalanches with the collector and emitter swapped? In this case, it seems that what is normally the 6 volts (or so) base/emitter breakdown is now the collector/emitter breakdown voltage.

    Am I missing something here?
     
  8. Jony130

    AAC Fanatic!

    Feb 17, 2009
    3,962
    1,097
    As you probably know the emitter-base junction of a bipolar transistor behaves as a Zener diode.

    [​IMG]

    With breakdown voltage greater than 6V.

    http://cr4.globalspec.com/thread/72501

    But when we swapped base with collector we expect that the breakdown voltage will increase his value up to 6V + Vbc = 6.6V

    [​IMG]

    But as we can see from our measurement is not true for NPN transistor.
    Because now BJT behaviors just like a poor's man tunnel diode.
    And tunnel diodes will have "negative resistor" region.
     
    • 1a.PNG
      1a.PNG
      File size:
      5.8 KB
      Views:
      226
    • aa.PNG
      aa.PNG
      File size:
      5.4 KB
      Views:
      202
  9. RichardO

    Well-Known Member

    May 4, 2013
    1,238
    385
    This may not always be true. I found that the spec for the 2SA2018 shows a breakdown voltage greater for the collector/base than the collector/emitter.
    Some FMMT717 may also have this characteristic.

    Do you think they would avalanche? The 2SA2018 is only $0.41 from Digi-Key. :D
     
  10. MrChips

    Moderator

    Oct 2, 2009
    12,449
    3,365
  11. t_n_k

    Thread Starter AAC Fanatic!

    Mar 6, 2009
    5,448
    782
    Thanks for the link.

    I replaced the LED & resistor with series LC and obtained the tuned oscillator version. Also worked with a 455kHz ceramic filter. Only the 2N2222A seemed to produce oscillation in either case. Not sure if there is any real practical use other than the simple flasher.
     
  12. t_n_k

    Thread Starter AAC Fanatic!

    Mar 6, 2009
    5,448
    782
    Thanks for the link.

    I replaced the LED & resistor with series LC and obtained the tuned oscillator version. Requires a slight change to flasher topology plus no big electrolytic. Also worked with a 455kHz ceramic filter. Only the 2N2222A seemed to produce oscillation in either case. Not sure if there is any real practical use other than the simple flasher.
     
  13. RichardO

    Well-Known Member

    May 4, 2013
    1,238
    385
    The one place where an avalanche circuit is useful is in generating a very fast risetime pulse. Jim Williams wrote a couple of application notes for Linear Technology about this. He was able to get pulses of tens of volts in amplitude with sub-nanosecond risetimes. He had to carefully select transistors to get this performance.

    I have been building an avalanche pulser of the Williams type. I built one circuit that had a risetime of about 1.5 ns on a solderless breadboard. When I tried to reproduce the circuit permanently soldered together, I couldn't get a risetime of less than about 6 ns. This has driven me crazy off and on for the last couple of weeks.

    Well, I finally found what the difference is. The transistor I used for the prototype was from a different manufacturer than the ones I was soldering into the circuit. I have tried several transistors from each manufacturer. Manufacturer "A" is always faster than manufacturer "B".

    Because of the above (and other rasons), I suspect that the avalanche characteristics for the 2N2222 make the flasher more a curiosity than a practical circuit where you need to build more than one.

    As a side note, I built an avalanche circuit using a 2N3904 with the collector and emitter swapped. It needed about 15 volts to oscillate and the risetime was very slow at about 1/2 microsecond.
     
  14. takao21203

    Distinguished Member

    Apr 28, 2012
    3,577
    463
    this circuits needs high voltage to work correctly, right?

    so it could not replace a joule thief circuit.

    How is the efficiency (to increase voltage)?
     
  15. The Electrician

    AAC Fanatic!

    Oct 9, 2007
    2,284
    328
    t_n_k likes this.
  16. MrChips

    Moderator

    Oct 2, 2009
    12,449
    3,365
    The emitter-collector avalanche effect occurs around 9V. Hence you need a supply voltage greater than this for the oscillator to work. 12V seems to work fine.

    The avalanche stops after a voltage drop of just a few volts. Hence this would not make an effective joule thief circuit.
     
  17. RichardO

    Well-Known Member

    May 4, 2013
    1,238
    385
    Yes, very good information.

    Zetex seems to have phased out the ZTX413 and is recommending the ZTX415 in its place. It is interesting that neither the ZTX413 or ZTX415 data sheet gives a risetime specification. The spec sheets do imply 20 or 25 ns wide pulses (at 50 or 60 amps!), however, indicating that the parts are fast. We just don't know how fast.

    Motorola used to make an avalanche transistor, the 2N5271. The risetime for these parts is given as less than 1 ns for a 100 volt minimum pulse amplitude! Unfortunately, these transistors don't seem to be readily available anymore.
     
Loading...