• Register
  • Login
  • looking for something?

    Single Transistor Amplifier Revisited Part 3, Common Base vs Common Emitter Configuration, Update

    One nagging question that I have long had is this: How does the common base compare with the common emitter configuration in voltage gain performance? The classic microphone circuit interfaces a loudspeaker (used as microphone) directly to the emitter of a common base amplifier. My initial guess was that the difference would be negligible. However, when I wired and compared the two circuits, I learned a few new things.

    Single Transistor Amplifier Schematic

    Single Transistor Amp Part 3 Schematic Update

    Immediately, I learned that the common base configuration gain was 6db lower than the common emitter. This blew my mind. Then I started checking input resistance and was shocked at how low it measured (8.5Ω). I had previously guessed that it would be about 100Ω. Then I plotted the common base input resistance beside the common emitter input resistance — very useful information. The gain was 6db lower in the common base configuration because the source resistance almost exactly equaled the input resistance.

    To obtain reliable input resistance data, I had to reduce the source resistance to 0.1Ω and connect via a Kelvin connection. Input resistance is surprisingly easy to determine experimentally by simply adding a pot between the low impedance voltage source and the amplifier input. Short the resistance and measure AC output voltage. Then increase resistance until the output voltage is exactly half. At this point, the pot resistance equals the input resistance and can be measured by a DMM.

    Preface to the update

    The initial data was in error due to a near resonant effect in the amplifier—this resonance caused unusually high voltage gain. After correcting the problem, I retook the data and updated the report.

    I wish to acknowledge the contribution of Mr. Colin Mitchell who flagged an error regarding coupling capacitor (C1) size. Since this circuit was tested at 1 to 2kHZ, I figured that the capacitor value was not an issue. However, when C1 was increased to 10 uF, the gain unexpectedly decreased. This indicated an additional problem which turned out to be an active filter effect that also involved C2. When C1 and C2 were 0.1uF, the circuit resonated at the 2kHZ test frequency — almost all the elements of a phase shift oscillator were present. This is clear evidence of the veracity of Murphy’s Law — how could all these conditions occur at random?

    As a result, Parts 1 and 2 had to be updated as well.

    Graph of Rin vs hFE

    Graph Rin vs hFE

    One curiosity is that in the common emitter configuration, input resistance tends to increase with hFE, but in the common base configuration, the input resistance tends to be constant. The one anomaly was the 2N5088. My guess is that it has a higher input resistance due to its small SO-23 package—all others were TO-92 devices.

    Graph of voltage gain vs hFE

    Graph Av vs hFE

    This graph clearly indicates that the gains of both configurations are identical within experimental accuracy—this assumes a very low impedance voltage source (10Ω for CE and 0.1Ω for CB).

    Data

    Part 3 Data

    Excel spreadsheet data

    Microphone amplifier experiment

    I then wired up a small computer loudspeaker (8Ω) as a microphone, and tried it on both circuits. This seems to be a practical circuit.

    Microphone Amplifier Schematic

    Microphone Amp CB v CE Schematic Update

    Whistle experiment

    Whistle Update 1

    I whistled directly into the microphone—this analysis is rather subjective because my lips are not calibrated for loudness and/or repeatability. However, it indicates a higher output signal with the common emitter amplifier.

    Coupled loudspeaker experiment

    Coupled Speakers

    I connected a loudspeaker to the signal generator via an output transformer and acoustically coupled it to a 2nd speaker used as a microphone. The mic signal was amplified by both the common base and common emitter configuration amplifiers and compared. The results are repeatable and clearly indicate that the common emitter amplifier has a 6db greater output signal (double).

    Conclusion

    The impedance of the 8Ω loudspeaker working into the low input resistance of the common base amplifier (8.5Ω) causes a 6db reduction of the signal at the input of the common base amplifier. At the input of the common emitter amplifier, there is no attenuation of the signal. As a result, the output signal of the common emitter amplifier is 6db greater—the common emitter amplifier wins in this regard… It is not an issue with the gain as much as it is with the low value of the input impedance. This may fly in the face of some, but here it is proven experimentally, and is repeatable.

    Of course, life is not always that simple—for best frequency response, the microphone may need to be loaded with a relatively low value resistor, and such a load will reduce the mic output voltage. If this is the case, then the common base amplifier is a good choice.

    Please understand that this is a fine point, and either configuration is an effective and acceptable method.

    Photos

    For the future

    Single Transistor Amplifier Revisited, Part 4 — effect of load resistance upon voltage gain

    Undocumented words and idioms (for our ESL friends)

    fly in the face –idiom – confrontation of accepted thought, truth or paradigm with new information or concepts — force someone to think outside the box

    Find more projects
    

    20 Responses to "Single Transistor Amplifier Revisited – Part 3"

    subscribe to newsletter
    1. Martins1 says: on December 14, 2012 at 4:05 pm
        member

      A sound energy input from a radio with the same loudness and pitch is most likely to produce the same result which would take care of the uncertainty from whistling.
      This is a good experimental study.

    2. This article is entirely incorrect and misleading.
      How can you get a gain of 460 from a transistor with a gain of 400.

      Secondly, a common-base amplifier works much better with a low-impedance device such as a dynamic microphone, than a common-emitter stage, that’s why you use a common-base amplifier for this type of application. So the author knows absolutely nothing about the subject.

      • Daniela says: on February 25, 2013 at 12:52 am
          member

        Mr Colin, I really think you are just trolling on the internet because your comment is useless. As you can see Mr Keith has tested almost all of his projects and presented the results in the articles. My guess is that you have never made any circuits in practice and you just talk from what you have read in books. Also Keith’s work is very appreciated as opposite of yours which btw has no option to comment on it (as you can do here). So basically no one can criticize your work on your website and maybe that is why you think you know so much.

      • Why do you think they have common-base stages?
        To interface low-impedance inputs.
        For a start, the 1mV from the speaker will not transfer 1mV via the coupling capacitor and that’s why the stage will not produce the same output as a common-base stage, where the coupling is direct.
        You should study electronics and see how false this article is.
        It gives the entirely wrong impression that a CB stage will produce the same as a CE stage for the 1mV from the dynamic microphone.

      • I get 6,000 visitors each day and dozens of emails. If there is a fault on my website, let me know.
        I have designed thousands of circuits and ran a magazine for many years, so I know what I am talking about.
        You will find 3 CB circuits on my website and it matches and amplifies the signal from a low impedance source much better than a CE stage.

      • You should go to my website where I clearly explain that 1mV from the microphone will becomes 0.3mV on the base of the transistor due to the impedance of the 100n capacitor, the 10k to deck and the input impedance of the transistor. This immediately gives the stage 30%; relative to the 100% of the CB stage. So there’s no comparison.

    3. Hello, Mr. Mitchell.
      I will explain what you are doing because you are not looking at yourself objectively–not an uncommon problem. You are using the logical fallacy referred to as “ad hominem” without realizing it.
      http://en.wikipedia.org/wiki/Ad_hominem

      Simply stated, if you do not like the message, shoot the messenger, or if you cannot logically handle or refute the argument, the character or appearance etc, of the author becomes the new issue–the response “the author knows nothing about the subject” or “you should study electronics…” has absolutely nothing to do with the truth at hand.

      Regarding, the voltage gain vs hFE issue–I agree to a point that voltage gain is a function of hFE, but my actual measurements indicate that aV > hFE up to about hFE = 400, and above that point it tends to decrease. Just wait for future articles in this series if you want to see really insane voltage gains.

      Regarding the application of the common base configuration for a mic amp being better than the common emitter, it is merely a matter of convention and preference. Historically speaking, the first microphones were generally carbon grain technology–while being quite noisy, they had a tremendous output signal level and could actually feedback on itself without amplification–this could be easily demonstrated with the early telephones where the ear peace could be placed against the mic. To function, the carbon mic needs DC bias and that is why there was a battery in the old phones.

      When radio came of age, high level modulation became a requirement. The easiest way to interface the carbon mic to a vacuum tube amp is via the common grid configuration in which it was simply inserted in series with the cathode–the DC bias was the cathode current–and it worked quite well. Yes, it could have been done via the common cathode amp, but that would require obtaining the DC bias from a relatively high 150V plate supply or the like–not really a good fit. The same convention carried over into transistor amplifiers and low impedance dynamic microphones, so today, most CB circuits are as you say common base configuration. This convention is based upon mainly practicality and is definitely not the only means available.

      Did I say that the CE config was better? No, I simply stated that it had the higher voltage gain. My original hypothesis stated that the voltage gain was essentially the same and I proved that it was true under certain conditions–the condition of zero source impedance. Unfortunately that is not the real world because microphones have a finite source impedance, and I demonstrated that a very low source impedance mic (8Ω loudspeaker(20Ω reactance)) fed into a 8.5Ω input impedance common base amp resulted in a 10db voltage attenuation–I do not think that you carefully read or understood my article.

      Regarding your statement:
      “This immediately gives the stage 30%; relative to the 100% of the CB stage”,
      are you in fact assuming that the CB input impedance is infinite? Is it not just the opposite?

      I have visited your web site and must agree that it is quite impressive, and that you have an axe to grind regarding shoddy engineering and circuits that has given rise to your classic “Spot the Mistakes” page.

      So let us base our arguments upon truth. My data and methods are clearly outlined and should be repeatable by others–I challenge you to actually do so.

    4. I picked you up on this article because of the number of inaccuracies and incorrect conclusion.
      You are presenting this information to newcomers to electronics and they will get the totally wrong impression of a CB and CE amplifier, from your results.
      Nowhere have you mentioned the attenuation of the input capacitor on the common emitter circuit.
      This is where the attenuation takes place and accounts for the fact that the CE circuit produces a result of about half as compared to a CB circuit.

      ”are you in fact assuming that the CB input impedance is infinite? Is it not just the opposite?”
      The input impedance of the Common Base Circuit is very LOW and when a very low microphone is added in series, it changes the input impedance very little. That’s why I have neglected the extra 8-20 ohms.
      Because the microphone in a CB circuit is connected directly to the stage, the 1mV produced by the microphone will be passed into the stage without any attenuation and that’s the big difference between the two stages.
      This is the point that you missed and produced a totally inaccurate set of results.
      Yes, I am annoyed at the amount of inaccurate material on the web, and that’s why I have identified hundreds of faulty circuits on my website.
      But what annoys me more is the fact that the authors fail to remove this false information and fail to see their mistakes.

    5. The reactance of the input coupling capacitor is 169Ω @ 2kHZ. This is negligible compared to the 2K input impedance of the CE amp–accounts for less than 1db error. I can see that a 0.1uF cap would be poor for a wide audio bandwidth app, but works fine at my test freq.

      Regarding the source impedance of the speaker turned mic (20Ω) and the CB input impedance of the amp (8.5Ω), there is significant attenuation of the signal level at the input (-10db). There would be tremendous attenuation (-25db) if a typical low impedance mic (150Ω) were used.

      Note the absence of a coupling capacitor in the CB schem–a bias of 4mA through the speaker (turned mic) is a non-issue, but would definitely offset the diaphragm position in a 150Ω mic (equivalent to a whopping +56dbmV signal).

      So we disagree here. Thank you for the friendly response.

    6. The reactive capacitance of 100n at 2kHz is 800 ohms – not 170 ohms.

      Regarding the source impedance of the speaker turned mic (20Ω) and the CB input impedance of the amp (8.5Ω), there is significant attenuation of the signal level at the input (-10db). There would be tremendous attenuation (-25db) if a typical low impedance mic (150Ω) were used. Note the absence of a coupling capacitor in the CB schem–a bias of 4mA through the speaker (turned mic) is a non-issue, but would definitely offset the diaphragm position in a 150Ω mic (equivalent to a whopping +56dbmV signal).

      You are bringing in “red herrings.”
      Stick with the faults in your original assessment and your continuing mistakes. . .

    7. If you look at the audio frequency range for voice and instruments, you will find most of the signals are active in the range 500Hz to 1kHz and that is the value you should use for the capacitive reactance determination.
      This is where the gain of the CE stage will be 50% of the CB stage.
      You have skewed so many of your assessments, and this is one of them.

    8. “You are correct on that one–I was figuring it was a 0.47uF cap.”

      You don’t mention 470n ANYWHERE. Too many mistakes and too many excuses.

    9. Your diagrams clearly show an 8 ohm speaker.
      I have $200,000 worth of stock and not one 150 ohm dynamic mic.
      Where is the average person going to find a 150 ohm mic?????

    10. Let’s look at your diagrams.
      The common base circuit has 3.2mA flowing through the 8R speaker. This produces 25mV across the speaker. This allows the speaker to produce 40mV p-p without distortion.
      Suppose the speaker produces a 1mV waveform. This converts to 275mV across the 2k2. How did you get an output of 150mV? The 100n on the base should be at least 10u.

    11. There are so many mistakes, I don’t know where to start.
      In the Common Emitter circuit, you say 16uA flows into the base and the transistor has a gain of 400. This means 6.4mA will flow in the collector and the voltage-drop across the 2k2 will be 14v. This is 2v more than the supply!
      Thus the transistor does not have a gain of 400.
      It has a gain of 200.
      Taking a gain of 200 and taking 1mV from the speaker, we have 0.3mV into the base (after the attenuation of the 100n). The output will be 200 x 0.3 = 60mV.

      As I said in the beginning, the CB circuit is a much-better performer.
      You remind me of Fleischmann and his cold fusion.

    12. Yeah, the DC values were lifted from the top schematic that indicates a 2N3904–that was an oversight–the base current is self-adjusting–you should have known that. If you had checked the spreadsheet, you would see that the circuit is set up to be able to accommodate a very wide range of hFE’s without seriously affecting the DC operating point–remember your initial complaint where you did not like the base divider circuit? Well, that is the logic behind it.

    13. “remember your initial complaint where you did not like the base divider circuit?
      I NEVER mentioned the base divider circuit.

      “you would see that the circuit is set up to be able to accommodate a very wide range of hFE’s without seriously affecting the DC operating point”
      That is entirely WRONG. It only works for a BRIDGE CIRCUIT.

    What do you think about this article? Leave a comment!