Hi Cruts,
That's very interesting and would be also interesting to test in a real-life circuit.

I would think then that the distortion comes from the small change in base emitter diode forward voltage drop.

Yes - I think that this observation can be explained easily:

* When the base is driven with a current source the non-linearity of the exponential Ib=f(Vbe) characteristic creates a voltage Vbe with positive half-waves which are somewhat smaller than the negative half-waves. This effect is cancelled (corrected) because the collector current is varying due to a similar non-linearity of the Ic=f(Vbe) characteristic. This is because now the opposite effect can be observed: The smaller positive Vbe half-waves will cause a larger Ic variation than the smaller negative amplitudes.

* When the base is driven with a voltage, there will be the non-linear influence of the Ic=f(Vbe) only (no cancellation effect).

* Can we conclude that we, therefore, should prefer to use this current injection method? To answer this question for a certain application we should not forget that in most cases we want to amplify and process voltages. This requires to realize a (non-ideal) current source using a large resistor.
Therefore, for a fair comparison we have to consider the available/possible voltage gains for both cases.

 

I agree that the original circuit has no bass response because CE is too small and has horrible but normal distortion.
A transistor normally produces distortion when its output level is fairly close to the supply voltage and it has no negative feedback. It is impossible to measure the gain when the even-harmonics distortion is so high.

If I increase the value of CE it increases the gain by about 25% but also increases the THD a little.

 

If I increase the value of CE it increases the gain by about 25% but also increases the THD a little.

Of course. The amount if distortion goes up with the output swing.

 

If I increase the value of CE it increases the gain by about 25% but also increases the THD a little.

Increasing the capacitance of CE does not "boost" the lower frequencies. Instead it simply allows the lower frequencies to have the same higher gain as the higher frequencies.

 

If I increase the value of CE it increases the gain by about 25% but also increases the THD a little.

Hi,

Yes, the secondary effect is the input impedance decreases and so the effect of the nonlinear Vbe is it gets even more nonlinear and thus higher distortion.
To help keep it more linear a small value resistor is placed in series with that cap, with an associated small decrease in gain.

 

Hi,

Yes, the secondary effect is the input impedance decreases and so the effect of the nonlinear Vbe is it gets even more nonlinear and thus higher distortion.
To help keep it more linear a small value resistor is placed in series with that cap, with an associated small decrease in gain.

This is the classical trade-off between gain and THD.

 

Hi,

Yes, the secondary effect is the input impedance decreases and so the effect of the nonlinear Vbe is it gets even more nonlinear and thus higher distortion.
To help keep it more linear a small value resistor is placed in series with that cap, with an associated small decrease in gain.

thanks

 

This is the classical trade-off between gain and THD.

Hi,

Yes, and it is too bad we don't usually connect stages with controlled current sources. Using another transistor might help, but it brings in more noise. Hopefully the noise will not be noticeable.
I am not sure we can always use a high value resistor which of course kills some of the gain, but it's good to know if we have the choice, the higher value resistor would be better.

We can see the effect in the current of the diode by DC biasing a diode and then perturbing it with an AC sine in the same way the input to the transistor does. The current through the diode becomes distorted a little.