Voltage amplification

Discussion in 'General Electronics Chat' started by PRS, Dec 8, 2008.

  1. PRS

    Thread Starter Well-Known Member

    Aug 24, 2008
    989
    35
    I have a question that has been confusing me lately and I'd sure be grateful for an answer to it. I've been playing with voltage amplifiers, especially common emitter circuits, and, just like the book says, if I try to amplify a signal larger than about 10 mV, the amplifier distorts the output.

    Here is my question. Starting with 1 microvolt at the antenae of a receiver, how do I turn that into, say, +/- 20 volts into an 8 ohm load? I can amplify up to 10mV, no problem, and then with a gain of 1000(!) I can turn that into 10 volts. But that amplifier must be a single stage and has all kinds of problems with stability and noise.

    What's the solution?
     
  2. bertus

    Administrator

    Apr 5, 2008
    15,645
    2,344
    Hello,

    What are you trying to recieve with the antenna?
    What kind of signal do you want to amplify?
    Is this a RF signal (that probably is modulated in some kind)?

    Greetings,
    Bertus
     
  3. hobbyist

    Distinguished Member

    Aug 10, 2008
    764
    56
    In my course material, they always taught that for a single stage amplifier the most voltage gain to design for is 20. That's without a emitter bypass cap. If you need more amplification you can cascade several amplifier stages. Also to increase gain per stage you can add a emitter bypass cap. being careful to not overdrive the stage into oscillation.

    Also I don't know if you know this or not, but you don't just simply add the gains of each stage, but rather you multiply the gains of each stage, for total gain of cascaded amp.

    ex. 2 stages with a volt. gain of 20 per stage will be calculated to be 20 X 20 = 400 for total gain.

    But this also has to take into acount load inpedance or input impedance of next stage in parrallel with the collector resistor. Which lowers the gain in that stage.

    Also staying with your question, your output stage would need to be a power amp. so as to produce the required voltage 20 volts across the small 8 ohm load.

    So to take a micro volt signal and amplify it to a 20 volt signal across an 8 ohm load requires small signal amps. as well as large signal power amp. for given aplication. And the power amp. will usually have a small volt gain, but will produce a large power gain, (volt X current) so see that's why you can not expect to get the results your looking for from a one stage small signal amp.

    Also from this, you can see why load impedance will dictate a lot of the design because if your load was a very high resistance 8000 ohms then you could probably use small signal amp. configuration, or if the output voltage was small then you could use a small sig. amp. so again it becomes pretty obvious that by looking at the required output with respect to voltages and currents you can assume what kind of stage would be rerquired to drive that load.

    As with the distorted output above 10 mv. sig. input, do you understand why that is, as you delve into transistor circuits, you could experiment and learn how you could restore that distortion above 10 mv. sig. in. one way would be by raising the supply voltage until distortion disappears. etc....it's a matter of taking one bit at a time and learning why the circuit behaves the way it does....

    hope this helps somehow...
     
    Last edited: Dec 8, 2008
  4. davebee

    Well-Known Member

    Oct 22, 2008
    539
    46
    One solution for better stability is to convert the frequency so that you don't apply much gain at either the original RF frequency or at audio frequencies.

    At the intermediate frequency where most of the gain occurs, since the frequency is fixed, the circuitry can be better protected against feedback oscillation.

    Signal leakage leading to oscillation is more a problem at the higher RF frequencies, and noise is more a problem at low signal levels, and microphonics (mechanical vibration) and signal distortion are more of a problem at audio frequencies, and heat is more of a problem at the output stage.

    So each stage of a receiver usually presents different problems that dominate the design of that stage.
     
  5. Audioguru

    New Member

    Dec 20, 2007
    9,411
    896
    You can't feed an amplified radio frequency signal to a speaker.
    First you need to filter it, amplify it, mix it with a different variable frequency oscillator to get a lower difference fixed frequency that is filtered and amplified a lot more. Then automatic-gain adjusts the level and the RF signal is demodulated into an audio signal that is amplified by an audio power amplifier that feeds a speaker.

    Look in Google for Super-heterodyne Radio Circuit.
     
  6. leftyretro

    Active Member

    Nov 25, 2008
    394
    2
    Or even a crystal radio detector followed by audio amplification if it's a AM transmitted signal. The point your are making for the OP, and that I wish to repeat for him, is that a radio station cannot send audio frequencies directly over the airwaves, they have to modulate a specific radio frequency (RF), called the carrier frequency, with the audio frequency information.

    A receiver will always have to demodulate this signal to extract the audio information. As well as super-het receivers, there are TRF, regenerative and other methods to demodulate an amplitude modulated RF signal.

    Lefty
     
  7. PRS

    Thread Starter Well-Known Member

    Aug 24, 2008
    989
    35
    I thank you all for the kind response, but it seems I did not make my question clear. It's about harmonic distortion. The base-emitter voltage is nonlinearly related to the output current, Ic. The second harmonic is directly proportional to the Vbe. Thus the need for a very small Vbe in order to stay within a fairly linear region of current gain. This is about 10 mV for a BJT operating as a common emitter amplifier with bypassed emitter.

    My problem is acheiving an overall voltage gain of 1000. Ignoring the power amp and the 8 ohm load it drives. By the way this happens to be an audio amp. I've had success using an op amp for the voltage gain, but I get stuck when trying to get this gain from CE amplifiers. To see the problem, consider 3 stages, each with a gain of about 32. By applying 10 mV at the input, I get 320 mV at the output which in turn drives the second stage, but, this creates harmonic distortion and the problem is further made bad when the 3rd stage is reached.

    As a solution, I've considered a step-up transformer right before the power amp. The first stage of the power amp would have high resistence -- a follower, for example. Even so, the load is reflected to the collector of last voltage amp by the inverse of the square of the turns ratio. Any thoughts? Does anyone know of a design having a gain of 1000 and an output of 20 volts peak to peak?
     
  8. KL7AJ

    AAC Fanatic!

    Nov 4, 2008
    2,039
    287
    Greetings:

    To keep things in perspective.....a typical communications reciever...such as might be used for shortwave communications, has to have an overall gain of over 100dB. (This assumes a relatively strong 10 microvolt signal at the antenna terminals, and a power output on the order of a watt.) This represents an overall power gain on the order of 10 BILLION to one!
    The good news is that you DON'T have to do it all in one stage (though regenerative receivers, very common in the 1930s and beyond, certainly did it this way).

    Now, assuming that you do have a "front end" R.F. amplifier with a voltage gain of 1000. If the input and output impedance of this are equal, this is a POWER gain of 1 million (60dB). This is certainly a trivial matter to do with a couple of small-signal pentodes....a little tricker to pull off with two transistors....actualy physically impossible. This leaves you with another 40 db of gain to pick up in the audio stages (assuming a very efficient detector stage). This is not hard to do with a couple of transistor stages...a class A driver and a class A power amp.

    Now, let me return to your original question....and perhaps depart a bit from many people on this forum in rendering this opinion. However, I have a few decades of R.F. design experience under my belt to back me up. It is this.

    There is seldom any advantage to attempting maximum theoretical gain out of ANY amplifying device. All the best designs, whether it's R.F., or audio, or instrumentation, use DISTRIBUTED gain, where each device works with a comfortable headroom. The only possible exception to this rule is R.F. power amplifiers, where efficiency is the one overruling consideration. Other than that, you are making like more difficult for yourself by "squeezing every drop" of gain out of any stage.

    Now, in the particular case you mention, what you are seeing is absolutely normal. In most modern receivers, the majority of the overall gain is achieved in the I.F. (Intermediate frequency) stage, which may have two or three active devices to achieve 60dB of gain. This leaves the R.F. stages and the Audio output stages a comfortable 20dB or so, each, in which to do their jobs.

    Hope this helps!

    Eric
     
    Last edited: Dec 9, 2008
  9. Audioguru

    New Member

    Dec 20, 2007
    9,411
    896
    Negative feedback is used to reduce distortion in transistor stages. Bootstrapping the collector resistor also reduces distortion.
    Negative feedback also reduces the gain.

    Here is a two transistors circuit with its gain reduced to 3 with plenty of negative feedback. Its distortion is 0.15% when its output level is 1V.

    The second transistor is an emitter-follower with a low output impedance and it drives the bootstrapping capacitor to reduce the distortion to 0.015%.

    An LM4562 opamp has a distortion of only 0.00003% when its gain is reduced with negative feedback to 1.
     
  10. hobbyist

    Distinguished Member

    Aug 10, 2008
    764
    56
    Sorry, my reply was way off base, (post #3)
    I thought you were just starting out in basic electronics, and were using the reciever question as an example of wanting to know how to amplify a signal.
     
  11. studiot

    AAC Fanatic!

    Nov 9, 2007
    5,005
    513
    Now that discussion has revealed what you really meant I would recommend one of these techniques, if you must have a single stage.

    Use a (relatively) high voltage supply (and of course high V transistors).
    You should be easily able to use a long tail pair plus perhaps an output follower to get your required gain and low distortion in a single stage. This can be done with either FETs or BJTs. An active load will raise the collector impedance and thus gain to high levels.
    An alternative would be to use the 'Anti Symmetric Pair' or ASP configuration, promoted by Linsley Hood some years ago.
     
  12. PRS

    Thread Starter Well-Known Member

    Aug 24, 2008
    989
    35
    Thanks, Eric! In the post just above this one I explained that I was building an audio amplifier with a gain of 1000 and an output to the power amp of 20 volts peak to peak. Getting 100dB from the RF and IF stages is no problem since we are dealing with a small signal with, say, 10mV at the detector.

    The problem I'm having arises from the need for signals in excess of 10 mV at the input of the stages. Say I have 3 stages to get my voltage gain. The first input is 10mV which is ok, no distortion. But it's output, say with a 32V/V gain, is 320 mV. Now I input this to stage 2 and immediately I'm distorting the signal. The problem is furthered at the 3rd stage. Hence the need for one stage with 60 dB voltage gain feeding high impedence followers of some sort.

    By the way, let me introduce myself. I'm Paul. Long ago I got a BSEE with the goal of becoming an electronics engineer. But I became a lawn mower instead. Electronics became a hobby, then I dropped it for over 10 years. Now I'm retired and I'm trying to get it back.

    Perhaps you can help me with this: I've been matching transistors by setting up a CE amp on a protoboard and comparing the gain of transistors. In lieu of a curve tracer, is this a practical technique?
     
  13. PRS

    Thread Starter Well-Known Member

    Aug 24, 2008
    989
    35
    I'm desirous of a high fidelity (a fraction of 1 percent distortion) audio amp. Negetive feedback is a good idea. I'll look into it. I found another way of increasing the allowable input voltage to a CE amp, and that is to include a resistor at the emitter. Here again, we're trading gain for purity. Thanks for the circuit!

    Paul
     
  14. PRS

    Thread Starter Well-Known Member

    Aug 24, 2008
    989
    35
    No harm, no foul. Good to meet you, hobbyist!
     
  15. PRS

    Thread Starter Well-Known Member

    Aug 24, 2008
    989
    35
    A "long tail pair?" I've never heard the term. Would you be referring to a differential pair? Plus I've never heard of the ASP configuration. But as far as the high Vcc needed, that's a good point. I'll add that the current gain should be high -- around 3 or 400, I think. Question: Can I get a gain of 1000 V/V with one stage? I know Op Amps get it done, but they are multi-staged. I'm wondering why they don't distort while amplifying a 10 mV signal to a 10 volt signal. Any thoughts?
     
  16. KL7AJ

    AAC Fanatic!

    Nov 4, 2008
    2,039
    287
    Hi again:

    As suggested above, you need to move to a higher voltage transistor, and probably one with a lower load resistance. At this point POWER gain is more important than voltage gain. In fact, the complementary/symmetry pair, common in the output stages of most modern audio power amps, is actually a voltage follower (complementary emitter-followers).

    Your method of matching transistors is certainly valid, and probably even more educational than a curve tracer.

    By the way....matching transistors was FAR more crucial in the old germanium days....tolerances were abominable back then. Nowadays, most modern transistors from the same lot are remarkably well-matched right off the shelf. But don't take my word for it....measure them!

    Eric
     
  17. PRS

    Thread Starter Well-Known Member

    Aug 24, 2008
    989
    35
    Thanks for the advice. By the way I see you're into tube radios. I had scavanged over the years planning to get into it. I had shoe boxes full of tubes still in their packages, a collection of aluminum variable capacitors, and a variety of transformers. When I lost interest I just gave them away! WHAA!
     
  18. studiot

    AAC Fanatic!

    Nov 9, 2007
    5,005
    513
    Tubes? sounds like you are in North America?

    Let's get a few design parameters solidified.

    You wish to design an amplifier for audio frequencies with an input sensitivity of:

    10 millivolts, though the source impedance is currently unknown?

    And output of 10 volts RMS into an 8 ohm load.

    You wish to use discrete transistors

    This is a particularly sensitive power amplifier design! Most power amps are designed for at least an order of magnitude less sensitivity to overcome noise problems.

    The source impedance will have a huge influence on the noise/hum performance of this amplifier.

    The debate about distributing the voltage gain over one or several stages has been raging for at least 5 decades. There have been good and bad solutions from both camps.

    If you can obtain the book

    Valve and Transistor Audio Amplifiers, by J L Linsley Hood you will have a masterpiece in your hands for your purposes. He discusses the question of obtaining voltage gain from long tail pairs (yes we call diffrential pairs that) in the range 4000 to 600,000 from discrete transistors and/or FETS at great length pages 182 to 190.


    I am still mystified where the connection to the antenna came in.
     
  19. KL7AJ

    AAC Fanatic!

    Nov 4, 2008
    2,039
    287
    The antenna reference was in the first post...so I presumed he was talking about receivers.

    North America? Yes, last I checked, this is where Alaska is located. :)

    Eric
     
  20. PRS

    Thread Starter Well-Known Member

    Aug 24, 2008
    989
    35
     
Loading...