We know the properties of an ideal Op Amp are 0 input current, low output impedance from negative feedback using excess gain (f), from the impedance ratio, Av-= - Feedback/ Input . Although the transistor collector current source is high impedance (Z) with voltage gain and the emitter is the source reduced by the current gain for each stage.

If you want a simple linear amplifier with 1 transistor you have a lot of tradeoffs with impedance reduction or current gain and/or voltage gain. But by using negative feedback for DC and AC, you can AC couple with higher gain and lower sensitivity to the typical 4:1 ratio of max/min hFE DC current gain specs in datasheets.

With H bias you cannot get full swing output without tuning for this hFE variation. Even when you tune h bias for maximum swing, the +ve distortion from lower Vbe causes lower DC current and thus lower gain causes asymmetry and the -ve output peaks < 2V approaching saturation where hFE reduces rapidly.

This asymmetric difference in +/- peak gain to Vavg is almost equal to the Total Harmonic Distortion THD (%)

Falstad Sim. with slider variables RF amp 50 ohms load, gain = 23 dB

1. I created a linear RF amp using the above. I applied an RF input sweep with a gain of 20 and 5.4Vpp output from a 3.3V supply into 50 Ohms.

2. Although the NPN was biased to 1/4 W that makes the intrinsic base resistance much lower than 50 Ohms thus adding Re= 750m Ohms times hFE raises the input dynamic impedance for linearity somewhat.

3. The collector draws from 10uH to Vcc the current that creates high voltage gain, attenuated by the negative impedance such that Vbe variation is reduced and also THD.

4. The 4k feedback to base 50R input achieves this gain with a 50R load but needs additional DC current provided by the 10k pullup to the base.

5. Both inputs and outputs are AC coupled to minimize settling time and minimize AC voltage loss for the frequency range used.

6. Unlike an Op-Amp there is far less DC open loop gain to work with but enough to utilize the gain bandwidth of the single NPN transistor so that an internal integrator is not needed.

7. More exotic variations of this exist for higher frequencies but I thought this simple version would make it interesting and more useful than the H bias versions of Class A single transistor amps for a larger linear swing.