Impedance Matching In Transistor Amplifier

Discussion in 'General Electronics Chat' started by sjgallagher2, Jun 6, 2014.

  1. sjgallagher2

    Thread Starter Member

    Feb 6, 2013
    111
    7
    I'm reading the book RF Circuit Design by Chris Bowick currently, and I've come across something really confusing. I was running through examples with small signal amplifier design. What happened was this circuit was designed:
    [​IMG]
    Which is a basic impedance matching network with source and load resistors. Now, that's all fine except one point- the inductor that goes up instead of down. I'm used to parallel meaning between the point and ground, but this seems like it's going to go between the point and something else, like the source voltage. Then it got worse when the bias circuitry was added:
    [​IMG]
    Now both the inductors are going up to the source instead of to ground. This isn't explained in the book and only this example is given with bias circuitry included.

    My theory is that he's making the inductors double as chokes, thus preventing the AC signal from going to the source. The capacitors next to them sort of enforces this (all caps going directly to ground are 0.1uF bypass caps) but I don't know if I'm right. This is important for if I design other circuits with different biasing schemes, so I know what to do.

    Thus the ultimate question is: why do the inductors go to the source?
    And another: what would happen if the schematic called for a series inductor and parallel capacitor? Would the capacitor go to ground instead of to the source? Would it go below or above R3 in the second picture, or in parallel with R3? Help is needed and appreciated people :)
     
  2. studiot

    AAC Fanatic!

    Nov 9, 2007
    5,005
    513
    If you are referring to the revised placement of L1, the power supply can be regarded as a short circuit at signal frequency, so to connect an inductior across the signal source one end goes to the sinal source, the other can go to either rail (perhaps through a series resistor).
     
  3. sjgallagher2

    Thread Starter Member

    Feb 6, 2013
    111
    7
    So my thinking was more or less correct! Now let's say C1 and L1 were switched (L1 in series then C1 in parallel), where would I put C1?
     
  4. ronv

    AAC Fanatic!

    Nov 12, 2008
    3,290
    1,255
    I'm no RF expert, but C1 is there to couple the high frequency to the base (low impedance at high frequency) and L1 is there to provide DC bias to the transistor while looking like a high impedance to the signal. So reversing where they are in the circuit wouldn't work.
     
  5. t_n_k

    AAC Fanatic!

    Mar 6, 2009
    5,448
    782
    Hi sjgallagher2,

    An issue with presenting the problem of the type you posted is that the design of matching networks must take the required impedance transformations into consideration. This impacts the matching topology that will achieve the desired matching.

    For instance, when considering input matching, is the source impedance higher or lower than the transistor input impedance? An RF power transistor in Class C may have a very low input impedance of only a few ohms. If the source impedance is greater than the transistor impedance - usually 50Ω for RF work - then one is restricted in the matching network choices. Another factor is the desired Q for the match, which may further limit network matching topology choice.

    For instance, the C1 / L1 topology shown in your posted examples can't be used to achieve a match with a source impedance of 50Ω and a transistor input impedance of 5Ω.
     
  6. Merlin3189

    New Member

    Oct 20, 2013
    7
    0
    I think what you need to remember is the difference between AC and DC ground.
    Your inductors may need to have one end at AC or signal ground, but not at DC ground (because of transistor bias requirements.) An inductor is a very low DC resistance, so you can't literally ground one end if the other must be at a DC bias potential. So you stick a big capacitor between it and the signal ground.

    That is another problem in looking at circuit diagrams: a 100pF capacitor looks the same as a 100nF capacitor, but at signal frequency the 100pF might resonate with an inductor and the 100nF be an effective short circuit.

    In your diagram C4 and C5 are much bigger than C1 and C2.
    C4 & C5 have low impedance at signal frequencies and ensure that the top end of L1 and L2 are at signal ground.
    C1 & C2 provide the frequency selection and have impedance comparable with the inductors at signal frequency.

    C3 is like C4 & C5, a low impedance at signal frequencies so the emitter is also at signal ground, but can have a DC resistance stabilise the bias point.

    "...why do the inductors go to the source? "
    Say L1 went to ground. You'd still need to put a big cap in series with it to block the DC. Then R4 would come directly to the top of R3 to give the DC bias. But now, for the signal, R3 and R4 are in parallel with each other and the B/E junction. That provides more damping on the tuned circuit than just R3 in parallel with B/E.
    For L2, you could make that go down to ground, again with a DC blocking capacitor, but also you now need a choke (as you said) to stop the signal output being wasted in R5 and damping of selectivity.
    Other effective configurations are possible, but this is simple enough.
     
    Last edited: Jun 7, 2014
  7. sjgallagher2

    Thread Starter Member

    Feb 6, 2013
    111
    7
    Great response, that's what I didn't understand at first, it really cleared it up for me.
    Now, here's my circuit I need to add bias circuitry to (already solved the impedance matching part, don't worry):
    [​IMG]
    And here are my ideas, based somewhat on the idea that capacitors act as shorts at AC, but generally they're total guesses:
    [​IMG]
    (Ignore the lack of R6 and C4)
    [​IMG]
    [​IMG]
    Now, the way I see it, a capacitor directly to ground would short circuit the signal which is bad. Maybe the resistors help with that, I'm not sure. Should I add a choke between R4 and Vcc? Wouldn't that have an effect on the impedance matching circuitry (L1 and C1)? What are some ways I can implement this circuitry together? There are a million options and I'm having a really hard time visualizing the circuit for AC and DC and what's required for each part. Help? Thanks a lot!
     
  8. sjgallagher2

    Thread Starter Member

    Feb 6, 2013
    111
    7
    I really need your help here people! I'm going to be immobile until I can figure this out.
     
  9. t_n_k

    AAC Fanatic!

    Mar 6, 2009
    5,448
    782
    You probably need to decouple the source from the base bias network. If C1 ise impedance match then it isn't going to short the signal to the base.
    Provided the parallel combination of R3 & R4 is much higher than the base input impedance the match should be preserved.
     
    Last edited: Jun 11, 2014
  10. sjgallagher2

    Thread Starter Member

    Feb 6, 2013
    111
    7
    Thanks a lot, I think I understand what you're saying. The help is much appreciated!!
     
  11. wakibaki

    New Member

    Jun 12, 2012
    7
    0
    Have a look at Wes Hayward's 'Solid State Design for the Radio Amateur.'

    The Bowick book is a classic, but you really need the practical examples in Hayward to flesh out your understanding, which I can tell is a bit patchy.

    The use of inductors to feed DC into RF circuitry is a commonplace, as you look at more examples you will become familiar with such techniques. Much of electronics, and particularly radio, makes use of topologies developed over years and decades, it takes some time to get experience. It's not about how smart you are, you can't get everything you need to know into your head immediately, it takes a little while.

    You'll get more feedback if you focus on a working example, and include component values, frequencies and source and load impedances, although these are frequently 50 ohms. Trying to make sense of a circuit in the abstract often leaves too many issues up in the air, focussing on a practical circuit eliminates a possible multiplicity of answers.

    w
     
  12. sjgallagher2

    Thread Starter Member

    Feb 6, 2013
    111
    7
    Helpful post, thanks for that. You're right, my understanding needs some work, some more experience would definitely help. I'll go look at that book.
     
Loading...