I am not sure how “fashionable” discrete transistor design is nowadays, but anyone using discrete transistors, bipolar or otherwise, for high speed designs other than for switching probably knows about cascode designs. The purpose of a cascode amplifier (not to be confused with cascade which is a chain of two or more amplifiers) is to isolate the Miller capacitance. Miller capacitance is the apparent multiplication of the base-collector or drain-source capacitance in an inverting amplifier. This can be simply explained due to the wider voltage swing of the collector (or drain). The effect of this voltage swing increases the apparent capacitance compared to visualizing the drain as being an AC ground. This can be illustrated in a simple example:
The performance of this circuit is shown below:
Adding a cascode while keeping the collector current the same:
Gives the following results:
So, a more than three times improvement in 3dB bandwidth. The theoretical Miller effect would be to increase the base-collector capacitance by a factor of (1+ G) where -G is the gain of the stage. However, base-collector capacitance is not the only factor affecting the bandwidth so the bandwidth improvement is not the 28x you might hope for using that equation!
The improvement depends on the source impedance, load capacitance and gain as well as the characteristics of the device being used. The cascode transistor does not have to be the same part number as the amplifying transistor and in fact does not even have to be the same type. For example, you could use a bipolar cascode transistor with an HEMT (high electron mobility transistors) or GaAsFET:
The improvement in bandwidth is shown below:
Not that these circuits are simply to illustrate the effect rather than being exact designs to follow. One important factor to keep in mind when adding a cascode transistor is that it takes some voltage headroom. You need to allow for this when deciding where to bias the base of the cascode transistor. Also, the base of the cascode needs to be held fairly solidly at a the chosen bias voltage. If you do not bias it with a low impedance and allow it to be dynamically moved around it will have undesirable effects.
The cascode amplifier can be viewed in different ways depending on your point of view. One way of looking at it is that the cascode transistor is simply that it passes the current through the emitter to its collector (less some base current) while preventing the emitter from moving, which is the important point. Another way of looking at the cascode is as a common base amplifier. A common base amplifier has a low input impedance which is the characteristic required to prevent the Miller effect on the common emitter gain stage – if the collector of the common emitter amplifier cannot move it cannot amplify the base-collector capacitance.
Depending on the characteristics of the individual devices it could be beneficial to use something other than a bipolar transistor for the cascode such as an HEMT. The graph below illustrates the improvement in bandwidth from using an HEMT for the cascode as well as the inverting amplifier transistor.
There’s many reasons for cascoding, a common one is voltage ratings. I’m designing an audio amp right
now using an Intersil (formerly RCA) transistor array at the front end (to get matched transistors) but it cannot
handle the full 60 volt supply I’m using. So I cascode the collectors of the diff amp. Beauty is you can use crummy unmatched discretes with the voltage rating that you need.
But then I’m one of the few nutcases left designing with discrete transistors anymore. I even combine NPN and PNP back to back for homemade SCR’s and DIACS (add a zener) just to save the lenghty drive to a parts store.