Class B RF amplifiers

Discussion in 'Amateur Radio Amplifiers' started by IZ8JFD, Oct 10, 2019.

ad: L-HROutlet
ad: l-rl
ad: Subscribe
ad: L-MFJ
ad: Left-2
ad: K5AB-Elect-1
ad: MessiPaoloni-1
  1. IZ8JFD

    IZ8JFD Ham Member QRZ Page

    Hi Everyone,

    I'm Valerio IZ4JFD and I think I am missing something about Class B RF power amplifiers principles & design.

    I have attached a PNG image which I have extracted from page 2 of the Motorola Application Note 779, a well known project by Helge Granberg (SK).

    The schematic shows an HF SSB driver circuit made of 2 active stages: the first includes 2 MRF476 transistors, and the second features 2 MRF475 semiconductors. Both stages work as Class B (in OM Granberg's words, Class AB) amplifiers.

    Nothing new, so far.

    Now, the primary winding in T2 and T3 (interstage and output transformer, respectively) has a center tap for DC supply to each couple of transistors.

    - During TRANSMISSION with NO SIGNAL at the driver input, the stand-by currents generated by both transistors in each stage of the circuit are equal and have opposite direction, so they meet and annihilate themselves at the center tap.

    - During TRANSMISSION with SIGNAL at the driver input, the positive half-cycle is amplified by the transistor to which it is applied, while the negative half-cycle takes the other transistor to inactivity (i.e., it behaves like an open circuit). In this situation, current flows inside the amplifying transistor from emitter to collector, enters the primary winding of the transformer, and returns to the DC supply through the center tap.

    My question is: what's the path that RF follows during TRANSMISSION with SIGNAL? Is it the DC one (from the emitter of the amplifying transistor to the center tap of the transformer) or is it a different one?

    Thanks in advance.

    Best 73's,

    Valerio Passeri

    Attached Files:

  2. K7TRF

    K7TRF Premium Subscriber QRZ Page

    I'd try looking at that a bit differently, the entire circuit is really ground referenced and all current flows from the power supply to ground in the classic current flow sense:

    - When signal is applied to the unbalanced side of T1, there's a pair of balanced and opposite phase signals present at the balanced secondary with each of the transistors Q1 and Q2 seeing one of those phases.

    - Whichever signal is in the positive half cycle with respect to ground increases the Base to Emitter voltage on the associated transistor moving that transistor into Collector to Emitter conduction, the transistor seeing the negative phased signal with respect to ground sees increasing reverse bias during that particular half cycle and does not conduct.

    - The transistor seeing the positive half cycle at its base conducts Collector to Emitter current in proportion to the amplitude of the signal on its Base and in the conventional current flow sense (flowing from positive to negative) the power supply is sourced through the transformer primary center tap with the current sink to ground at the Emitter of the transistor currently conducting.

    - So RF current (the signal is RF) flows from the transformer primary center tap to one end of the transformer primary during one half of each cycle of RF and from the center tap to the other end of the transformer primary during the other half cycle.

    - Neither transistor sources current, it just allows current sourced from the power supply through the transformer center tap to flow to ground through the transformer primary.

    - The transformer secondary is thus driven by magnetic flux coupled from one half or the other half of the transformer primary during each half cycle. From that standpoint the transformer turns ratio is one half of the primary windings to the full secondary windings as during each half cycle only one half of the transformer primary is carrying current.

    That's how the driver stage is configured and the finals stage is basically the same topology with different active devices.

    IOW, in each stage of the two stage amplifier there are two balanced transistors but only one conducts for each half cycle of the input RF signal. The current from the power supply flows through one half of the primary winding to whichever transistor is currently conducting and no current flows in the transistor in cutoff. Those RF currents flowing through one half of the transformer primary flux couples a balanced signal into the secondary of the stage's output transformer. The exact same process repeats itself in the second stage and again the current flowing through one half of T3's primary flux couples to T3's secondary and the unbalanced output load.
    Last edited: Oct 11, 2019
  3. IZ8JFD

    IZ8JFD Ham Member QRZ Page

    Thank you very much, Dave, for your CLEAR and COMPLETE answer.

    I'm trying to build a driver circuit based on the AN779. The main differences between my project and Granberg's schematic are the following:

    - I will use 2 x 2SC2166 instead of the MRF476s and 2 x 2SC3133 in lieu of the MRF475;

    - I don't have 43 mix ferrites, but I can use some 61 mix cores.

    Using 61 mix ferrites, I have found out (through practical experimentation) that I got the best results winding 4 turns at the unbalanced (50 ohms) side of a BN61-202 binocular core, with a 220pF ceramic capacitor in parallel with the winding. I used different wire diameters, coated with enamel or PVC.

    Starting from here, I have already wound T1 on a BN61-202 core this way:

    - 4 turns in parallel with 39pF at the primary winding:

    - 2 turns in parallel with 150pF at the secondary.

    T2 and T3 both have a center tap, and I have wound T2 (again, on a BN61-202) as follows:

    - 6+6 turns at the primary winding;

    - 3 turns in parallel with 82pF at the secondary.

    As you can see, T1 and each half of T2 (half the primary winding and the secondary) have a 4:1 impedance ratio (corresponding to a 2:1 turns ratio). Now, I have a question.

    Everytime there is a transformer with a center tap at the primary winding, is it correct for me to consider (like you said) that the turns ratio is:

    half the turns at the primary/the turns at the secondary

    and that, consequently, the impedance ratio is:

    (half turns at primary/turns at secondary)^2

    with all this meaning that, in transformers like T2 and T3, each half of the primary (in conjunction with the secondary) acts as a transformer in its own?

    Thank you.

    Best 73's,

    Valerio Passeri
  4. K7TRF

    K7TRF Premium Subscriber QRZ Page

    In this particular circuit topology yes, but don't extend that to 'everytime there is a transformer with a center tap...'

    IOW, this applies because the circuit topology for both stages ensures that only one transistor in each pair will be active at a given moment in time so time varying RF current will only flow in one half of the primary or the other half of the primary at any given time and time varying RF current will never flow in both halves of the primary winding at the same time. But center tapped transformers are used in many other ways so don't extend this concept to all possible uses of center tapped transformers.

    For this circuit T2 and T3 behave as described above where the effective turns ratio is half of the primary turns to the full secondary turns.
  5. K7TRF

    K7TRF Premium Subscriber QRZ Page

    Without comparing the data sheets for those devices it's hard to say what if any circuit changes you might have to make when switching active devices in the finals stage.

    The most obvious change might be to the resistor feeding the biasing diode (D2). The combination of that series resistor from the positive power supply line and the forward biased diode establishes the Base voltage and resulting idle mode current to both of the output transistors. The bias voltage and resulting idle mode bias current should be adjusted to minimize crossover distortion. Set the bias voltage on the Bases of the final devices too low and crossover distortion will increase as there's a small zero crossing plateau in the output signal, set the bias voltage too high and both devices will be active for a small portion of each RF cycle where you really only want one device or the other active for each half cycle.
  6. N7EKU

    N7EKU Ham Member QRZ Page


    The amp design (also by Granberg) in AN762 is a much nicer design:

    This has a much better bias design and also the DC feed design is better, relieving the RF transformers of DC feed current requirements.


    K7TRF likes this.

Share This Page

ad: hrd-2