I asked this once before, but I'm going to ask it again, hoping someone who knows will enlighten me.

The output of an IF amp is perhaps 10 mV peak. If not, we must attenuate it, for the input of a common emitter amp is limited to 10 mV peak. Otherwise it distorts the output. Therefore cascading does no good. So how in the world do we get, say 10 volts peak, into a speaker? I mean consider it: the CE amp would have to have a gain of 1000 volts per volt. That creates an instability in itself.

Where am I going wrong?

 

Can you post up a specific circuit? Why must a common emitter amp be limited to a 10 mv input signal?

You realize that an intermediate frequency (IF) stage in a radio receiver is some processing distance from the audio amp stage?

 

the CE amp would have to have a gain of 1000 volts per volt. That creates an instability in itself.

Why this causes instability?

 

An audio power amplifier has an internal voltage gain of about 100,000 at low frequencies. Then negative feedback is used to reduce the gain to maybe 100 which reduces the distortion to maybe 0.003% at low frequencies.

A common-emitter transistor with a low resistance voltage input and no emitter resistor for negative feedback has a voltage gain of about 150 and a distortion of about 40%.
If its gain is reduced to 10 with negative feedback then its distortion is reduced to about 2.7% which is better but is still horrible. Many transistors are used in modern opamps so their distortion is as low as 0.00003% at full output.

 

Years ago I built an AM radio from fundamentals -- no ICs allowed. I did this to make sure I understood all of the principles involved. I wound my own toroids, avoided cans and even variable capacitors.

Recently I blew the dust off and hooked it up to see if it still worked. It does to the extent that I can see a 100 mV pk-pk audio signal on my oscilloscope. But the audio amp was used for something else and I've been trying to build one to take its place-- again with discreet components.

Be that as it may, if you have the time and material, breadboard any basic CE amp with bypassed emitter resistor and try to get an undistorted output using a sinewave greater than 10 mV peak. Be sure to allow for the needed swing at the collector. Here's a design if you want one. Rb1= 82K, Rb2=33K, RE=3.9K, RC=3.9K and bypass the emitter with, say, a 100 uF capacitor.

Now, up to 10mV pk, everything is fine. No distortion, after that you'll see one peak getting wider than its corresponding negetive peak. Slowly increase the input signal and watch the sinewave quickly become distorted.

The reason for this is the inherent nonlinear relationship between the input voltage, v(be), and the output current, I(c). It is an exponential. And yet we can still make a linear amplifier if we keep the change in v(be) small -- about 10mV. This is called the small signal approximation and is the basis of the hybrid-pi model.

 

I just posted the reason for distortion for inputs greater than 10 mV above. Instability in the single CE stage getting 1000 v/v gain is, in part, due to the need for a bypass capacitor, thus eliminating any stability due to RE.

I built a CE amp that gets 'only' 300 V/V gain and it wobbles up and down at 120 hertze -- the regulated supply voltage. Even though the ripple is small it shows up when we have a high gain amp unless some sort of feedback is used. But if we use feedback, we lose gain.

I tried decoupling the supply with a big cap and it didn't work. Got any ideas?

 

Thanks Audioguru. That is very interesting.
So using two stages of transistors, with loop feedback, would provide even less distortion?

 

Audio amps (and op amps) use a differential pair at the input. I'm going to breadboard one and experiment with the gain and most significantly any distortion at the output given a 10 mV and greater sinewave at the input.

I know for fact that the emitter follower does not have the same problem in this respect as does the CE amp. (See the reason for this problem in the CE amp in my post above). This is because the follower has an output in phase with the input and the Q point essentially rises and falls with the signal across the BE junction.

The problem with using feedback is that it comes at the price of gain. But I believe you're right about using an extremely high gain with feedback in order to acheive an undistorted output. I've noticed op amps and audio amps (ICs) don't seem to have any problem with large input voltages. And looking at their schematics it seems to me they get a huge voltage gain with the first stage and the rest of the circuit is either for feedback to the voltage amp or power gain with no further voltage amplification.