Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Quick & easy transistor question

  1. Mar 28, 2010 #1
    Thanks ahead for fielding my amateurish question - I'm a math guy and circuits are new to me.

    I'm trying to design a quick & dirty (aka need-not-be-linear) RF amplifier circuit and was wondering the following stupid thing: A lot of the circuits I've found are configured for NPN - if I replaced those with PNP, would I just have to reverse the supply voltage, assuming similar specs? Is that naive? I ask because when I ordered a bunch of parts, I got an assortment of (BJT) power transistors, but apparently they are all PNP (I have some 2n2222's, but I'm looking for higher power capability...)

  2. jcsd
  3. Mar 28, 2010 #2
    The quick & dirty answer is yes, it would work, provided you reversed all the other polarized components too such as diodes and electrolytic capacitors.
  4. Mar 28, 2010 #3
    Awesome! Thanks. Is there a not quick & dirty answer? I suppose I should just give it a shot first and keep reading my book if I want a more in-depth answer.
  5. Mar 28, 2010 #4
    It should be understood that the PNP transistors must have the bandwidth needed for the amplifier.
  6. Mar 28, 2010 #5
    The simplest explanation of differences between the two are as follows:

    NPN - voltage from Base to Emitter causes a proportional voltage (larger) from Collecter to Emitter

    PNP - voltage from Base to Emitter causes an inversely proportional voltage (smaller) from Collecter to Emitter

    In full saturation, NPN will pass the full voltage supply from Collector to Emitter. PNP will will block the full voltage from Collector to Emitter.

    There's a pretty good explanation of things here:

    The best explanation (though a bit more mindbending) can be found here:

    I have to admit that semiconductor physics make my brain hurt a little. Good luck!
  7. Mar 29, 2010 #6
    This can't be serious. Where did you learn about electronics?

    Whether a bjt is pnp or npn, the modus operandi is the same, except polarities are reversed.

    For npn, forward biasing the b-e junction results in collector current, but c-e voltage decreases as collector current increases. For a pnp, it is exactly the same. When configured as common emitter, an increasing input signal gets inverted at the output for an npn or a pnp.

    In saturation, turning an npn fully on results in the device behaving as a switch fully on. A small voltage drop is incurred from collector to emitter. With a pnp, the exact same thing happens, only the polarity is reversed.

    The "amasci" web site you recommend, is a contrarian source. The ideas advanced on that site go against the published info from semiconductor OEMs. The author makes statements that defy conservation of energy, fields, and semiconductor physics.

    A person beginning in electronics should start with peer-reviewed sources, and should consider semiconductor OEMs as the horse's mouth. After such a person has really mastered the basics, then they can be creative, adding their own innovations to advance the state of the art.

  8. Mar 29, 2010 #7
    Completely by myself. I'm quite used to getting questions like this.

    Yes. I'm trying to see where we don't agree. You have put it in a slightly larger number of words.

    I'd be interested in an explanation of how that site "makes statements that defy conservation of energy, fields, and semiconductor physics". I'd also be interested in knowing about any peer-reviewed sources that would cure me of my apparent lack of understanding. I'm not really looking to "advance the art" (there are enough people who have already dedicated their lives to that), just to have an acceptable understanding of how the device works (insofar as the interactions of P and N type substrates go).
  9. Mar 29, 2010 #8
    Please re-read your post, then mine. We did not say the same thing. As far as peer-reviewed sources go, good ones are (but not limited to) any text authored or co-authored by Horowitz, Hayt, Neudeck, Comer, Hill, Muller, Camins, & Sze.

    Regarding "amasci", he states that the bjt collector current is controlled by the width of the depletion layer in the b-c region, and that that layer is controlled by the field due to Vbe. I exchanged emails with him reminding him that the depletion zone does not impart energy to the carriers, but rather that the external source (bias or signal) provides said energy. The carriers attain a velocity limited by saturation, and when they transit through the b-c depletion zone, they are already kinetic.

    He also states that the E fields are due to voltage not current. I reminded him that in order to change E, one must change the energy, requiring power times time. Power is current*times voltage, which he should realize, but won't acknowledge. He is clueless regarding fields, energy, and general physics. I have the emails, which I can forwards to you if you wish. Just let me know. BR.

  10. Mar 30, 2010 #9
    Yes, that's just the Ebers-Moll model of BJT operation: taken from basic diode physics. If you object to Ebers-Moll, or object to the conventional explanation of semiconductor rectifiers, you need to list their failings first. It's impossible to have any sort of discussion as long as you're attacking common textbook material.

    For anyone interested in understanding BJTs, I strongly suggest that you don't listen to transistor OEMs, but instead take a look at undergrad engineering texts such as Horowitz and Hill's AOE book. They start out with a simplified "little man inside the transistor" approach where Ic is controlled by Ib via the "Transistor Man." Then they go on to "Improved transistor model: transconductance amplifier;" introducing the Ebers-Moll equation to show how Ic varies with Vbe, and give a list of "Rules of thumb for transistor design" based upon it.

    My article "How transistors work" is just a visual/verbal explanation of the Ebers-Moll BJT model for beginners, with math removed (it's Babylonian rather than Euclidian: re-stating basic semiconductor concepts from undergrad textbooks so your children and grandparents can understand.)

    When experts state that "BJTs are voltage controlled," they're really just saying that if we stop treating transistors as black boxes, and instead look inside that box, we find that they are well described by the Ebers-Moll equation.
    Last edited: Mar 30, 2010
  11. Mar 30, 2010 #10
    The "transconductance" amp, is only valid for small signal action. For large signals, "gm" is not well defined, as it varies with Ic, the dc value of collector current.

    Experts regard bjt's as "charge controlled", and this is consistent with OEMs, and semiconductor physics. Regarding Ebers-Moll, the original 1954 paper, too big to post here, depicts the bjt as 2, current controlled current sources. One in the conventional mode, Ic = alpha*Ie, and the other upside down for reverse mode operation, useful in the saturated state. I'll send it to you if you wish. It gives an illustration showing the current controlled current sources.

    All reputable texts and references depict the bjt as current controlled when viewed as a black box, then charge controlled at the micro scale. Ebers-Moll, which you cite, lists the bjt as current controlled where Ic is alpha*Ie. No OEM says otherwise.

    The OEMs are not reliable? Get serious. They make the raw silicon crystal starting with a seed in the oven. They inject the dopants to form junctions using pretty sophisticated equipment, i.e. diffusion, ion implantation. Nobody knows the bjt like they do. Even the great Ebers & Moll got their info at some point from OEMs who produce bjt devices.

    When you say that OEMs don't know it like you do, I can only reply "get serious". Do you produce semiconductors?

    You simply cannot fathom any electrical device being current controlled. You insist that the whole electrical universe be defined in terms of voltage because you believe that fundamentally voltage is more important. It's merely a mental block. If the bjt was really voltage controlled, it would be universally known and stated by OEMs. Why would they call it CC when it's really VC? The FET (J- and MOS-) has been depicted as VC for as long as I can remember. Did they get the FET right? The fact that 1 is classified as CC, and the other as VC, indicates that there is a difference, and that they know from first hand knowledge producing both.

    Enough for now. Cheers.

  12. Mar 30, 2010 #11
    Glad I could bring up an exciting topic! haha.. I have another real question:

    I have some power transistors now, which I have chosen somewhat arbitrarily based on their max power dissipation (I need ~ 1watt for total RF power, so I got a couple transistors in the 5-10W max range). How do I go about establishing bias points based on the specs of the transistors? I.e. I'm pretty sure I can't just drop a power transistor into a circuit designed for a low-power 2n3906 or something (or can I?)

    Is there a simple way to calculate total power output? Perhaps in terms of a transfer function...? (that's an annoying question, I know, feel free to send me back to circuits 101...)
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook